DEPARTMENT OF SOCIOLOGY AND WORK SCIENCE Robin Jonsson is affiliated with the Department of Sociology and Work Science at the University of Gothenburg, Sweden. DEPARTMENT OF SOCIOLOGY AND WORK SCIENCE Robin Jonsso is affiliated with th Department of Sociology and Work Science at the University of Gothenburg, Sweden. i Retaining the Aging Workforce Studies of the interplay between individual and organizational capability in the context of prolonged working lives Robin Jonsson Department of Sociology and Work Science University of Gothenburg Gothenburg 2021 Retaining the Aging Workforce: Studies of the interplay between individual and organizational ca- pability in the context of prolonged working lives © Robin Jonsson 2021 robin.jonsson@gu.se ISBN 978- 91- 8787 6 - 39 - 4 (PRINT) ISBN 978- 91- 8787 6 - 40 - 0 (PDF) C over illustration: Sixten Haage , 1980, Drypoint illustration “ factory in the Ruhr,” no. 26/265. Used with permission. Font in the thesis: “Silian Rail” Online: http://hdl.handle.net/2077/69889 Gothenburg Studies in Work Science No . 20 Department of Sociology and Work Science, University of Gothenburg Printed in Gothenburg, Sweden 2021 Printed by Stema Specialtryck AB, Gothenburg, Sweden Abstract Amidst the realities of workforce aging and policies to prolong the working life across Europe, policymakers and researchers have emphasized the employer’ s role in providing suitable working conditions for older workers under the human resource term “ age man- agement.” Although th e definition of age management refers to all age groups, previous research has primarily used the concept in the context of older workers’ employment par- ticipation. The age - management imperative stresses employers’ active role in adapting work to older workers’ abilities and preferences by developing, modifying, sustaining, and providing flexible work arrangements to strengthen these workers’ capabilities and en- courage them to remain in work longer. However, previous research has repeatedly show n that employers often take a passive or reactive approach to implementing measures to retain older employees. Also, knowledge of effective workplace interventions to promote older workers’ employability, work abil ity, and health is sparse. Against this background, this thesis seeks insights into enabling and hindering conditions at the individual and or- ganizational levels using the capability concept as a theoretical approach. Capability refers to individuals’ and organizations’ ability to convert existing resources and perform actions to realize selected goals, such as continuing to work or retaining older workers. This thesis intends to answer the following questions: What personal and work-related factors promote or hinder older workers’ ability and willingness to continue working? What barriers and opportunities influ- ence public welfare organizations’ capability to retain older workers? How do individuals’ capabilities interact with organizational capabilities in the context of prolonging the working life? The empirical studies of this thesis are based on three data sources: (a) a population - based representative panel and registered data covering 55 – 64 - year- olds ( from Panel Survey of Aging and the Elderly , PSAE), (b) a cross - sectional study of employees aged 55 years or older in the city of Gothenburg (Hearts – Lexliv study), and (c) an interview study with 19 line managers and their HR part ners in the healthcare sector (ArbetsKraftsprojektet). The first arti cle compared personal and work- related determinants of retirement preferences and subse- quent behavior. Findings indicated that people tended to retire later than initially pre- ferred. Personal h ealth, work ability, physical work exposure, and job satisfacti on were important determinants of preferences and actual retirement behavior. The second article explored the role of work accommodations among older workers with poor health meas- ured in terms of disease, illness, and sickness. Poor health cond itions in combination with few opportunities to accommodate work increased the risks of early retirement. However, these associations were not equally consistent with illness . The third article examined the impact of negotiated individualized work arrangements, so - called idiosyncratic deals, on retirement preferences. It found negotiated individualized work arrangements matching employees’ competence, experience, and growth opportunities to be important for public - sector employees’ retirement preferences. The fourth ar ticle examined organizational bar- riers to retaining older workers in Swedish public- welfare organizations. The findings sug- gested that public welfare employers’ ability to retain older workers is hindered by line managers’ high workload, inflexible HR strategies, and the absence of elaborated age- management strategies and coordination within the organization. Empirical data also in- dicate that preferences for earlier retirement and earlier actual behavior are more frequent among females and among individuals in lower socioeconomic positions with lower edu- cation. This thesis demonstrate s that older workers’ preconditions for prolonging the working life are largely shaped by the interaction between personal factors, such as health, SV AN ENMÄRKET Trycksak 3041 0234 Retaining the Aging Workforce: Studies of the interplay between individual and organizational ca- pability in the context of prolonged working lives © Robin Jonsson 2021 robin.jonsson@gu.se ISBN 978 - 91 - 8787 6 - 39 - 4 (PRINT) ISBN 978 - 91 - 8787 6 - 40 - 0 (PDF) C over illustration: Sixten Haage , 1980, Drypoint illustration “ factory in the Ruhr,” no. 26/265. Used with permission. Font in the thesis: “Silian Rail” Online: http://hdl.handle.net/2077/69889 Gothenburg Studies in Work Science No . 20 Department of Sociology and Work Science, University of Gothenburg Printed in Gothenburg, Sweden 2021 Printed by Stema Specialtryck AB, Gothenburg, Sweden Abstract Amidst the realities of workforce aging and policies to prolong the working life across Europe, policymakers and researchers have emphasized the employer’ s role in providing suitable working conditions for older workers under the human resource term “ age man- agement.” Although the definition of age management refers to all age groups, previous research has primarily used the concept in the context of older workers’ employment par- ticipation. The age -management imperative stresses employers ’ active role in adapting work to older workers’ abilities and preferences by developing, modifying, sustaining, and providing flexible work arrangements to strengthen these workers’ capabilities and en- courage them to remain in work longer. However, previous research has repeatedly show n that employers often take a passive or reactive approach to implementing measures to retain older employees. Also, knowledge of effective workplace interventions to promote older workers’ employability, work abil ity, and health is sparse. Against this background, this thesis seeks insights into enabling and hindering conditions at the individual and or- ganizational levels using the capability concept as a theoretical approach. Capability refers to individuals’ and organizations’ ability to convert existing resources and perform actions to realize selected goals, such as continuing to work or retaining older workers. This thesis intends to answer the following questions: What personal and work-related factors promote or hinder older workers’ ability and willingness to continue working? What barriers and opportunities influ- ence public welfare organizations’ capability to retain older workers? How do individuals’ capabilities interact with organizational capabilities in the context of prolonging the working life? The empirical studies of this thesis are based on three data sources: (a) a population - based representative panel and registered data covering 55– 64 -year -olds ( from Panel Survey of Aging and the Elderly , PSAE), (b) a cross -sectional study of employees aged 55 years or older in the city of Gothenburg (Hearts – Lexliv study), and (c) an interview study with 19 line managers and their HR part ners in the healthcare sector (ArbetsKraftsprojektet). The first arti cle compared personal and work-related determinants of retirement preferences and subse- quent behavior. Findings indicated that people tended to retire later than initially pre- ferred. Personal h ealth, work ability, physical work exposure, and job satisfacti on were important determinants of preferences and actual retirement behavior. The second article explored the role of work accommodations among older workers with poor health meas- ured in terms of disease, illness, and sickness. Poor health cond itions in combination with few opportunities to accommodate work increased the risks of early retirement. However, these associations were not equally consistent with illness . The third article examined the impact of negotiated individualized work arrangements, so -cal led idiosyncratic deals, on retirement preferences. It found negotiated individualized work arrangements matching employees’ competence, experience, and growth opportunities to be important for public - sector employees’ retirement preferences. The fourth ar ticle examined organizational bar- riers to retaining older workers in Swedish public- welfare organizations. The findings sug- gested that public welfare employers’ ability to retain older workers is hindered by line managers’ high workload, inflexible HR strategies, and the absence of elaborated age- management strategies and coordination within the organization. Empirical data also in- dicate that preferences for earlier retirement and earlier actual behavior are more frequent among females and among individuals in lower socioeconomic positions with lower edu- cation. This thesis demonstrate s that older workers’ preconditions for prolonging the working life are largely shaped by the interaction between personal factors, such as health, perceived work ability, and physical and psychosocial working conditions, and organiza- tions’ capability to adapt and provide flexible solutions concerning, for example, negoti- ated individualized work arrangements or work accommodations. The model presented in this thesis provides guidance in identifying vulnerable groups and conditions in the labor market. Keywords: Retirement, retirement preferences, active aging, working conditions, idiosyn- cratic deals, age management, senior policies, age-conscious leadership, healthcare, public welfare organizations, job design, capability, human resources management, sustainable working life, job quality , idiosyncratic deals, ISBN 978 - 91- 8787 6 - 39 - 4 (PRINT) ISBN 978 - 91- 8787 6 - 40 - 0 (PDF) Contents L IST OF STUDIES ................................ ................................ ................................ ............................... I SUMMARY OF STUDIES ................................ ................................ ................................ .................. II L IST OF FIGURES ................................ ................................ ................................ ........................... III A BBREVIATI ONS ................................ ................................ ................................ ............................ IV F ÖRORD — I NTRODUCTORY REMARKS IN SWEDISH ................................ ................................ . V 1 A CALL FOR WORKFORCE M OBI LI ZATI ON ................................ ................................ .......... 1 Aims and objectives ................................ ................................ ................................ ................. 4 Organization o f the thesis................................ ................................ ................................ ....... 5 2 O N WORKING LONGER IN LIFE ................................ ................................ ............................. 6 Toward the mobilization of the Swedish workforce: a policy shift ................................ 6 The Swedish pension system: historical background ................................ ......................... 8 Structural developments and challenges in the Swedish labor market ......................... 11 The case of Swedish public welfare organizations ................................ ........................... 12 3 I NDIVIDUAL AND ORGANI ZATION AL CAP ABI LI TY ................................ ........................... 15 Capability in the context of work ................................ ................................ ........................ 16 Individual capability ................................ ................................ ................................ ............ 18 Organizational capability ................................ ................................ ................................ ... 19 A model of individual – organizational capability ................................ ........................... 21 4 UNDERSTANDING THE EMP LOYM EN T PARTICIP ATI ON OF OLD ER WORKERS ............ 22 Societal- level determinants ................................ ................................ ................................ ... 23 Workplace- level determinants ................................ ................................ .............................. 25 Prejudices toward older workers ................................ ................................ ...................... 25 Working conditions ................................ ................................ ................................ ............ 26 Opportunities for individualized work arrangements ................................ .................. 27 Occupational health and safety management ................................ ................................ 28 Individual - level determinants ................................ ................................ ............................... 29 5 DA TA AND METHODS ................................ ................................ ................................ ............ 31 The quantitative studies ................................ ................................ ................................ ........ 32 Panel Survey of Ageing and the Elderly ................................ ................................ ......... 32 Study sample ................................ ................................ ................................ ................... 33 Data collection ................................ ................................ ................................ ............... 33 Hearts – Lexliv dataset ................................ ................................ ................................ ......... 34 Study sample ................................ ................................ ................................ ................... 34 Data collection ................................ ................................ ................................ ............... 34 perceived work ability, and physical and psychosocial working conditions, and organiza- tions’ capability to adapt and provide flexible solutions concerning, for example, negoti- ated individualized work arrangements or work accommodations. The model presented in this thesis provides guidance in identifying vulnerable groups and conditions in the labor market. Keywords: Retirement, retirement preferences, active aging, working conditions, idiosyn- cratic deals, age management, senior policies, age- conscious leadership, healthcare, public welfare organizations, job design, capability, human resources management, sustainable working life, job quality , idiosyncratic deals, ISBN 978 - 91 - 8787 6 - 39 - 4 (PRINT) ISBN 978 - 91 - 8787 6 - 40 - 0 (PDF) Contents L IST OF STUDIES ................................ ................................ ................................ ............................... I SUMMARY OF STUDIES ................................ ................................................................ .................. II L IST OF FIGURES ................................ ................................ ................................ ........................... III A BBREVIATI ONS ................................ ................................ ................................ ............................ IV F ÖRORD —I NTRODUCTORY REMARKS IN SWEDISH ................................ ................................ . V 1 A CALL FOR WORKFORCE M OBI LI ZATI ON ................................ .......................................... 1 Aims and objectives ................................ ................................................................ ................. 4 Organization o f the thesis................................ ................................ ....................................... 5 2 O N WORKING LONGER IN LIFE ................................ ................................ ............................. 6 Toward the mobilization of the Swedish workforce: a policy shift ................................ 6 The Swedish pension system: historical background ................................ ......................... 8 Structural developments and challenges in the Swedish labor market ......................... 11 The case of Swedish public welfare organizations ................................ ........................... 12 3 I NDIVIDUAL AND ORGANI ZATION AL CAP ABI LI TY ................................ ........................... 15 Capability in the context of work ................................ ................................ ........................ 16 Individual capability ................................ ................................................................ ............ 18 Organizational capability ................................ ................................ ................................... 19 A model of individual – organizational capability ................................ ........................... 21 4 UNDERSTANDING THE EMP LOYM EN T PARTICIP ATI ON OF OLD ER WORKERS ............ 22 Societal-level determinants ................................ ................................ ................................... 23 Workplace-level determinants ................................ ................................ .............................. 25 Prejudices toward older workers ................................ ................................ ...................... 25 Working conditions ................................ ................................................................ ............ 26 Opportunities for individualized work arrangements ................................ .................. 27 Occupational health and safety management ................................ ................................ 28 Individual -level determinants ................................ ................................ ............................... 29 5 DA TA AND METHODS ................................ ................................................................ ............ 31 The quantitative studies ................................ ................................ ........................................ 32 Panel Survey of Ageing and the Elderly ................................ ......................................... 32 Study sample ................................................................ ................................ ................... 33 Data collection ................................ ................................................................ ............... 33 Hearts – Lexliv dataset ................................ ................................ ......................................... 34 Study sample ................................................................ ................................ ................... 34 Data collection ................................ ................................................................ ............... 34 Operationalization of the dependent variable ................................ ............................... 35 Operationalization of the in dependent variables ................................ .......................... 37 Personal health indicators ................................ ................................ ............................ 39 Decision latitude................................ ................................ ............................................ 39 Idiosyncratic de als and face validity ................................ ........................................... 40 Operationalization of covariate variables ................................ ................................ ....... 42 Statistical methods ................................ ................................ .............................................. 42 The qualitative study ................................ ................................ ............................................. 43 From voice a nd transcripts to themes ................................ ............................................ 45 Ethical considerations ................................ ................................ ........................................... 45 6 SUMMARY OF RESULTS ................................ ................................ .......................................... 47 Article I: Personal and work -related determinants of retirement behavior ............. 47 Article II: Work accommodation to support older workers with poor health ........ 4 8 Article III : The importance of skill matching for retirement preferences ............... 49 Article IV: An organizational perspective on the retention of older work ers ......... 50 7 DISCUSSION ................................................................ ................................ ............................ 53 Reflections on working life prolongation: for whom and for what purpose? ............ 58 Methodological considerations ................................ ................................ ............................ 59 8 C ON C LUSION ................................................................ ................................ .......................... 63 Practical implications, contributions, and relevance to society ................................ ..... 63 Future perspectives ................................ ................................ ............................................... 64 SVENSK SAMMANF ATTN ING [ SUMMARY IN SWEDISH ] ................................ ........................... 67 A P P ENDIX A ................................ ................................ ................................ ................................ .. 75 Interview guide (translated from Swedish) ................................ ................................ ........ 75 REFERENCES ................................ ................................ ................................ ................................ . 77 i List of studies This thesis is based on the following studies, referred to in the text by their Roman nu- merals. I. Jonsson, R. (202 1) Preferred versus actual retirement and their per- sonal and work- related predictors: A comparison using representative Swedish survey and register data. Submitted m anuscript . II. Jonsson, R. , Dellve, L. , and Halleröd, B. (2019) Work despite poor health? A 14 - year follow- up of how individual work accommoda- tions are extending the time to retirement for workers with poor health conditions. S S M Population Health , Vol. 9 , Article No. 100514 1. III. Jonsson, R. , Hasselgren, C. , Dellve, L. , Seldén, D. , Larsson, D. , and Stattin, M. (2021) Matching the pieces: The presence of idiosyncratic deals and their impact on retirement preferences among older workers. Work, Aging and Retirement , Vol. 7 , No. 3 , pp. 240 – 255 . IV. Jonsson, R. , Lindegård, A. , Björk, L. , and Nilsson, K. (2020) Organi- zational hindrances to the retention of older healthcare workers. Nordic Journal of Working Life S tudies, Vol. 10, No. 1, pp. 41 – 58 . Operationalization of the dependent variable ................................ ............................... 35 Operationalization of the in dependent variables ................................ .......................... 37 Personal health indicators ................................ ................................ ............................ 39 Decision latitude................................ ................................ ................................ ............ 39 Idiosyncratic de als and face validity ................................ ................................ ........... 40 Operationalization of covariate variables ................................ ................................ ....... 42 Statistical methods ................................ ................................ ................................ .............. 42 The qualitative study ................................ ................................ ................................ ............. 43 From voice a nd transcripts to themes ................................ ................................ ............ 45 Ethical considerations ................................ ................................ ................................ ........... 45 6 SUMMARY OF RESULTS ................................ ................................ ................................ .......... 47 Article I: Personal and work - related determinants of retirement behavior ............. 47 Article II: Work accommodation to support older workers with poor health ........ 48 Article III : The importance of skill matching for retirement preferences ............... 49 Article IV: An organizational perspective on the retention of older work ers ......... 50 7 DISCUSSION ................................ ................................ ................................ ............................ 53 Reflections on working life prolongation: for whom and for what purpose? ............ 58 Methodological considerations ................................ ................................ ............................ 59 8 C ON C LUSION ................................ ................................ ................................ .......................... 63 Practical implications, contributions, and relevance to society ................................ ..... 63 Future perspectives ................................ ................................ ................................ ............... 64 SVENSK SAMMANF ATTN ING [ SUMMARY IN SWEDISH ] ................................ ........................... 67 A P P ENDIX A ................................ ................................ ................................ ................................ .. 75 Interview guide (translated from Swedish) ................................ ................................ ........ 75 REFERENCES ................................ ................................ ................................ ................................ . 77 i List of s tudies This thesis is based on the following studies, referred to in the text by their Roman nu- merals. I. Jonsson, R. (202 1) Preferred versus actual retirement and their per- sonal and work-related predictors: A comparison using representative Swedish survey and register data. Submitted m anuscript . II. Jonsson, R. , Dellve, L. , and Halleröd, B. (2019) Work despite poor health? A 14 -year follow -up of how individual work accommoda- tions are extending the time to retirement for workers with poor health conditions. S S M Population Health , Vol. 9 , Article No. 100514 1. III. Jonsson, R. , Hasselgren, C. , Dellve, L. , Seldén, D. , Larsson, D. , and Stattin, M. (2021) Matching the pieces: The presence of idiosyncratic deals and their impact on retirement preferences among older workers. Work, Aging and Retirement , Vol. 7 , No. 3 , pp. 240– 255 . IV. Jonsson, R. , Lindegård, A. , Björk, L. , and Nilsson, K. (2020) Organi- zational hindrances to the retention of older healthcare workers. Nordic Journal of Working Life S tudies, Vol. 10, No. 1, pp. 41– 58 . ii Summary of studies Article Aim Design Conclusion I To examine the as- sociation between personal and work- related factors and two outcome measures, i.e., retire- ment preferences, and actual retirement Prospective longitudinal study Findings indicated that people tended to retire later than initially preferred. H ealth, physical work ability, physical work expo- sure, and job satisfaction are important de- terminants of earlier retirement behavior. However, social support and recognition were only relevant to retirement prefer- ences. By using retirement preference in- stead of actual exit as the outcome meas- ure, researchers risk overestimating the im- pact of specific predictors on actual retire- ment. I I What poor health conditions are asso- ciated with earlier re- tirement timing? Can opportunities to ac- commodate and modify working con- ditions moderate these relationships? Prospective longitudinal study Disease, illness, and sickness are important predictors of earlier retirement. This effect is stronger when combined with few op- portunities for work accommodation, indi- cating an interaction. However, the associa- tion is not as strong for illness. Thus, c om- bating poor health and increasing opportu- nities to accommodate older workers through decision latitude appear important for delaying retirement. I I I Examine the role of individualized work arrangements—so - called idiosyncratic deals—on intentions to retire among older Swedish public- wel- fare employees Cross - sec- tional survey study Individual ized work arrangements regard- ing tasks and work responsibilities were found to be associated with employees’ re- tirement preferences. This stresses the im- portance of matching competence and ex- perience, offering growth opportunities, and offering tasks outside the ordinary area of responsibility. I V Gain knowledge of the organizational preconditions for re- taining older em- ployees in the Swe- dish healthcare sec- tor Qualitative interview study The ret ention of older healthcare workers is hindered by line managers’ high work- load, inflexible HR strategies, absence of elaborated age management strategies, and poor coordination within the organization. T he healthcare sector and especially HR strategies must be more proactive in ad- dressing these issues to retain older healthcare workers. iii List of figures Figure 1. Typology of individual willingness and abilities to continue working. ............... 18 Figure 2. Typology of organizational capability to implement age management based on the resources and preferences identified by Kadefors et al. (2020). ................................ .... 20 Figure 3. Preconditions for a prolonged working life: organizational capability to implement age- management practices, combined with a typology of individual abilities and willingness. ................................ ................................ ................................ ............................. 21 Figure 4. Flow chart of the PSAE dataset. ................................ ................................ ............... 33 Figure 5. Flow chart of the Hearts – Lexliv dataset. ................................ ................................ . 35 Figure 6. Organizational hindrances to the retention of older healthcare workers: a model developed in Article IV. ................................ ................................ ................................ .. 51 Figure 7. Preconditions for a prolonged working life: organizational capability to implement age- management practices, combined with a typology of individual abilities and willingness. ................................ ................................ ................................ ............................. 57 ii Summary of studies Article Aim Design Conclusion I To examine the as- sociation between personal and work- related factors and two outcome measures, i.e., retire- ment preferences, and actual retirement Prospective longitudinal study Findings indicated that people tended to retire later than initially preferred. H ealth, physical work ability, physical work expo- sure, and job satisfaction are important de- terminants of earlier retirement behavior. However, social support and recognition were only relevant to retirement prefer- ences. By using retirement preference in- stead of actual exit as the outcome meas- ure, researchers risk overestimating the im- pact of specific predictors on actual retire- ment. I I What poor health conditions are asso- ciated with earlier re- tirement timing? Can opportunities to ac- commodate and modify working con- ditions moderate these relationships? Prospective longitudinal study Disease, illness, and sickness are important predictors of earlier retirement. This effect is stronger when combined with few op- portunities for work accommodation, indi- cating an interaction. However, the associa- tion is not as strong for illness. Thus, c om- bating poor health and increasing opportu- nities to accommodate older workers through decision latitude appear important for delaying retirement. I I I Examine the role of individualized work arrangements— so- called idiosyncratic deals— on intentions to retire among older Swedish public- wel- fare employees Cross - sec- tional survey study Individual ized work arrangements regard- ing tasks and work responsibilities were found to be associated with employees’ re- tirement preferences. This stresses the im- portance of matching competence and ex- perience, offering growth opportunities, and offering tasks outside the ordinary area of responsibility. I V Gain knowledge of the organizational preconditions for re- taining older em- ployees in the Swe- dish healthcare sec- tor Qualitative interview study The ret ention of older healthcare workers is hindered by line managers’ high work- load, inflexible HR strategies, absence of elaborated age management strategies, and poor coordination within the organization. T he healthcare sector and especially HR strategies must be more proactive in ad- dressing these issues to retain older healthcare workers. iii List of figures Figure 1. Typology of individual willingness and abilities to continue working. ............... 18 Figure 2. Typology of organizational capability to implement age management based on the resources and preferences identified by Kadefors et al. (2020). .................................... 20 Figure 3. Preconditions for a prolonged working life: organizational capability to implement age-management practices, combined with a typology of individual abilities and willingness. ................................ ................................ ................................ ............................. 21 Figure 4. Flow chart of the PSAE dataset. ................................................................ ............... 33 Figure 5. Flow chart of the Hearts – Lexliv dataset. ................................ ................................ . 35 Figure 6. Organizational hindrances to the retention of older healthcare workers: a model developed in Article IV. ................................ ................................ ................................ .. 51 Figure 7. Preconditions for a prolonged working life: organizational capability to implement age-management practices, combined with a typology of individual abilities and willingness. ................................ ................................ ................................ ............................. 57 iv Abbreviations ATP Allmän Tilläggspension (English: general supplementary pension) CFA Confirmatory factor analysis EU European Union HEART S– LEXL IV Hearts – Lexliv dataset HRM Human resource management I - DEALS Idiosyncratic deals NDC Notional defined contribution system NPM New public management OECD Organization for Economic Co - operation and Development OHS Occupational health services OHSM Occupational health and safety management PAYG Pay as you go PSAE Panel Survey of Aging and the Elderly PWL Prolonged working life SALAR Swedish Association of Local Authorities and Regions (Swedis h: Kommuner och Regioner , SKR ) SCB Statistics Sweden (Swedish : Statistiska centralbyrån ) SEM Structural equation modeling SET Social- exchange theory SSIA Swedish Social Insurance Agency (Swedish : Försäkringskassan ) SWEA Swedish Work Environment Authority (Swedish : Arbetsmiljöverket ) v Förord — Introductory remarks in Swedish Det är slutet som gillas. Så sa de byns original, Ernst, till mig och en vän när vi var på cykelsemester på Gotland för ett decennium sedan. Ernst ägde ett slitet vitt eternithus i Slite, som var överfullt med prylar och trädgården likaså. Ernst menade att han snart skulle få tid att färdigställa allt och då skulle missnöjda grannar glömma den stora oredan – för det är ju ändå slutet som gillas. Det tog flera år innan jag förstod det djupa i Everts resonemang, allt från budskap i världsreligioner till ett misstag i människors vardag. Som doktorand har också en tonvikt på slutet en attraktiv lockelse, nämligen att misslyckanden under doktorandtiden kan re- pareras vid disputationen… men så närmar sig slutet och vad gör man då? Mycket har hänt sedan 2016. Att vara doktorand blev mer intensivt än jag föreställde mig, men det har också varit synnerligen utvecklande. Till min hjälp under dessa år har jag haft en handledartrio i Göteborg bestående av Agneta Lindegård Andersson, Lisa Björk och Lotta Dellve. Mina handledare har gett ett socialt och akademisk stöd som varit betydel- sefullt under denna tid. Jag, men främst Agneta , genomförde ArbetsKraftsprojektet i Gö- teborg. Agneta har alltid uppmuntrat mig i avhandlingsarbetet, gett värdefulla inspel i ar- betet med avhandlingen samt haft en sund och pragmatisk inställning till akademin i all- mänhet. Lisa har med enastående sk ärpa, engagemang och idéer uppmanat mig att göra det bästa för att förbättra språk och texter. Min huvudhandledare Lottas erfarenhet och träffsäkra framtidsspaningar har varit betydelsefull a för helhetsbilden i avhandlingen. Lotta har också alltid uppmuntr at nya idéer och perspektiv, en inställning jag tar med mig i framtiden. I ArbetsKraftsprojektet södra del blev det många turer ner till Skåne. Tack till Mikael Widell Blomé, Jonas Borell men särskilt min bihandledare Kerstin Nilsson som genomförde ArbetsK raftsprojektet i Skåne. Denna avhandling är byggd på ett sopberg av text och jag har följande opponenter att tacka för värdefulla inspel vid de olika gransk- ningsseminarierna: Gunnar Gillberg och Caroline Hasselgren vid PM - seminariet. Daniel Seldén och Jonas Bertilsson vid mittseminariet. Roland Kadefors och Stefanie König vid slutseminariet. I slutskedet av avhandlingen fick jag också bra kommentarer och korrek- turläsning av Sarah Philipson Isaac, Hannes Lagerlöf, Jonas, Hanna Uddbäck, Helena Hå- kansson, och Caroline. Tack för detta! Under min tid som doktorand har jag träffat ovanligt många personer. Om vi börjar med att jag haft två arbetsplatser : Institutet för stressmedicin (ISM) och institutionen för soci- ologi och arbetsvetenskap. Fördelen med att ha två arbetsplatser är många, exempelvis att man kan ha samma kläder två dagar i rad. Men ibland har det också varit svårt att räcka till. P å ISM har jag ändå alltid känt mig välkommen och diskussioner i personalrummet har alltid varit givande med högt i ta k. Det får kanske bli ett nytt pizzaförsök igen vid något tillfälle. Tack till alla på ISM och särskilt Ingibjörg Jonsdottir. Mest tid har jag emel- lertid kommit att tillbringa på I nstitutionen för sociologi och arbetsvetenskap . Här verkar den administrativa funktionen vanligtvis utan att synas för oss doktorander, förutom på institutionens olika tillställningar då de oftast håller lågan brinnande längst (en mycket bra iv Abbreviations ATP Allmän Tilläggspension (English: general supplementary pension) CFA Confirmatory factor analysis EU European Union HEART S– LEXL IV Hearts – Lexliv dataset HRM Human resource management I - DEALS Idiosyncratic deals NDC Notional defined contribution system NPM New public management OECD Organization for Economic Co - operation and Development OHS Occupational health services OHSM Occupational health and safety management PAYG Pay as you go PSAE Panel Survey of Aging and the Elderly PWL Prolonged working life SALAR Swedish Association of Local Authorities and Regions (Swedis h: Kommuner och Regioner , SKR ) SCB Statistics Sweden (Swedish : Statistiska centralbyrån ) SEM Structural equation modeling SET Social- exchange theory SSIA Swedish Social Insurance Agency (Swedish : Försäkringskassan ) SWEA Swedish Work Environment Authority (Swedish : Arbetsmiljöverket ) v Förord — Introductory remarks in Swedish Det är slutet som gillas. Så sa de byns original, Ernst, till mig och en vän när vi var på cykelsemester på Gotland för ett decennium sedan. Ernst ägde ett slitet vitt eternithus i Slite, som var överfullt med prylar och trädgården likaså. Ernst menade att han snart skulle få tid att färdigställa allt och då skulle missnöjda grannar glömma den stora oredan – för det är ju ändå slutet som gillas. Det tog flera år innan jag förstod det djupa i Everts resonemang, allt från budskap i världsreligioner till ett misstag i människors vardag. Som doktorand har också en tonvikt på slutet en attraktiv lockelse, nämligen att misslyckanden under doktorandtiden kan re- pareras vid disputationen… men så närmar sig slutet och vad gör man då? Mycket har hänt sedan 2016. Att vara doktorand blev mer intensivt än jag föreställde mig, men det har också varit synnerligen utvecklande. Till min hjälp under dessa år har jag haft en handledartrio i Göteborg bestående av Agneta Lindegård Andersson, Lisa Björk och Lotta Dellve. Mina handledare har gett ett socialt och akademisk stöd som varit betydel- sefullt under denna tid. Jag, men främst Agneta , genomförde ArbetsKraftsprojektet i Gö- teborg. Agneta har alltid uppmuntrat mig i avhandlingsarbetet, gett värdefulla inspel i ar- betet med avhandlingen samt haft en sund och pragmatisk inställning till akademin i all- mänhet. Lisa har med enastående sk ärpa, engagemang och idéer uppmanat mig att göra det bästa för att förbättra språk och texter. Min huvudhandledare Lottas erfarenhet och träffsäkra framtidsspaningar har varit betydelsefull a för helhetsbilden i avhandlingen. Lotta har också alltid uppmuntr at nya idéer och perspektiv, en inställning jag tar med mig i framtiden. I ArbetsKraftsprojektet södra del blev det många turer ner till Skåne. Tack till Mikael Widell Blomé, Jonas Borell men särskilt min bihandledare Kerstin Nilsson som genomförde ArbetsK raftsprojektet i Skåne. Denna avhandling är byggd på ett sopberg av text och jag har följande opponenter att tacka för värdefulla inspel vid de olika gransk- ningsseminarierna: Gunnar Gillberg och Caroline Hasselgren vid PM -seminariet. Daniel Seldén och Jonas Bertilsson vid mittseminariet. Roland Kadefors och Stefanie König vid slutseminariet. I slutskedet av avhandlingen fick jag också bra kommentarer och korrek- turläsning av Sarah Philipson Isaac, Hannes Lagerlöf, Jonas, Hanna Uddbäck, Helena Hå- kansson, och Caroline. Tack för detta! Under min tid som doktorand har jag träffat ovanligt många personer. Om vi börjar med att jag haft två arbetsplatser : Institutet för stressmedicin (ISM) och institutionen för soci- ologi och arbetsvetenskap. Fördelen med att ha två arbetsplatser är många, exempelvis att man kan ha samma kläder två dagar i rad. Men ibland har det också varit svårt att räcka till. P å ISM har jag ändå alltid känt mig välkommen och diskussioner i personalrummet har alltid varit givande med högt i ta k. Det får kanske bli ett nytt pizzaförsök igen vid något tillfälle. Tack till alla på ISM och särskilt Ingibjörg Jonsdottir. Mest tid har jag emel- lertid kommit att tillbringa på I nstitutionen för sociologi och arbetsvetenskap . Här verkar den administrativa funktionen vanligtvis utan att synas för oss doktorander, förutom på institutionens olika tillställningar då de oftast håller lågan brinnande längst (en mycket bra vi egenskap) . Anna - Karin Wiberg, Pia Jacobsen och Anna Larsson är några som jag haft återkomm ande kontakt med. Jag har också haft mycket att göra med de alltid så vettiga Göran Sundqvist och Cecilia Hansen Löfstrand som väglett doktorandgruppen under denna tid. Och inte att förglömma institutionens eminenta prefekt, Anna Peixoto. Doktorandgruppen under denna tid är som barnen från Frostmofjället i ständig rörelse. Nya doktorander har tillkommit samtidigt som andra försvunnit. Jag har delat resor, kon- ferenser och AW men också vardagar i korridorer, lunchrum och möten med några nu- varande och tidigare doktorander: Jon Sunnerfjell, Sarah, Hanna, Peter O’reilly, My Hyltegren, Helena, Jane Pettersson, Hannes, Jonas, Linda Weichselbraun, Nadine -Kram- winkel-Jha, Kristin Jesnes och Caroline. Med Caroline har det kommit att bli många öl, friterad mat och diskussioner med stor spännvidd på den eleganta krogen Muchacho vid Vasagatan. Den sista tiden som doktorand har blivit mycket mer givande och roligare med dig Caroline. Jag får inte heller glömma doktorandernas feedbackgrupp som har varit ett trevligt inslag under ett antal eftermiddagar. Gruppen bestod av Sarah, Linda, Hanna, Jon, Helena och mig själv. Som doktorand har man också axlat rollen som lärare. Undervisning och handledning av uppsatser blev ett stort och givande inslag i doktorandutbildningen. Hann a, Monica Andersson Bäck, Karin Allard, Carl -Christian Trönnberg och Dietmar Rauch är några som varit kursledare och kurskollegor på bland annat personal- och ar- betsvetarprogrammet, men också andra kurser. Under min tid som doktorand har jag också varit d el av den F ortefinansierade nationella forskarskolan för åldrande och hälsa (SWEAH) samt centrum för åldrande och hälsa (AGECAP). Med AGECAP och SWEAH har det blivit flera träffar runt om i landet. 2017 erhöll jag Svensk Nationell datatjänst stipendium för metodkurser vid sommarskolan i ICPSR i Ann Arbor, Michigan. Det övergripande forskningsprojektet har genomförts inom ramen för det Fortefinansierade projektet, ” Möjliggöra deltagande, hälsa och välbe- finnande hos äldre anställda - för ett hållbart och inkl uderande arbetsliv” (Forte Dnr. 2012-1696). Tack till alla som deltog i ArbetsKraftsprojektet och Göteborgsenkäten. Tack till Peter och Sixten Haage som hjälpte till med torrnålsgravyren för bokens omslag, en tavla som jag ibland betraktade när jag var lit en. E tt tack också till Ronnie Pingel som via forumstrådar 2006 motiverade mig att börja läsa och ta examen i statistik, något som hjälpt på många sätt inom och utanför akademin. I mitt privata liv har avhandlingsarbetet varit föremål för maximalt två timmars diskussion de senaste åren, trots stor nyfikenhet från mina föräldrar Conny och Lena , samt systrar Sandra och Felicia. Det finns mycket annat som är mer intressant att prata om, men nu hoppas jag denna bok ger en lite bättre bild. Mina föräldrar ha r gett en trygg uppväxt, sunda värderingar och uppmuntran, som varit helt avgörande i mina grund - och dokto- randstudier. Några som har ännu mindre koll på avhandlingsarbetet är vännerna i Team SG . Med SG höjs dagligen humöret på distans eller tillsammans , Tack till Shahab A, Daniel H, Patrik S, Per T, Daniel A, Martin O, Eric E, Alexander K och Christian L ! En hälsning skickas också till min vän Marcus L ! Bäst under dessa år som doktorand är såklart bildandet av en egen familj och mina under- bara barn Alva och Cornelis. Alva kom till världen precis i början av min doktorandtid vii medan Cornelis kom några år senare. När detta skrivs är de två respektive fem och ett halvt år år. De både tar och ger energi samtidigt. Det är först den senaste månaden som Alva har blivit varse om att sin pappa skriver en bok. Alva väntar med spänning på en bildbok. Slutligen, inte minst , min älskade Klara som gett stöd och mening i framgång och motgång, från glädje och sorg, till hydda och borg. Det är slutet som gillas . Men än är det inte slut. Robin Jonsson Karlstad, november 2021 vi egenskap) . Anna - Karin Wiberg, Pia Jacobsen och Anna Larsson är några som jag haft återkomm ande kontakt med. Jag har också haft mycket att göra med de alltid så vettiga Göran Sundqvist och Cecilia Hansen Löfstrand som väglett doktorandgruppen under denna tid. Och inte att förglömma institutionens eminenta prefekt, Anna Peixoto. Doktorandgruppen under denna tid är som barnen från Frostmofjället i ständig rörelse. Nya doktorander har tillkommit samtidigt som andra försvunnit. Jag har delat resor, kon- ferenser och AW men också vardagar i korridorer, lunchrum och möten med några nu- varande och tidigare doktorander: Jon Sunnerfjell, Sarah, Hanna, Peter O’reilly, My Hyltegren, Helena, Jane Pettersson, Hannes, Jonas, Linda Weichselbraun, Nadine - Kram- winkel- Jha, Kristin Jesnes och Caroline. Med Caroline har det kommit att bli många öl, friterad mat och diskussioner med stor spännvidd på den eleganta krogen Muchacho vid Vasagatan. Den sista tiden som doktorand har blivit mycket mer givande och roligare med dig Caroline. Jag får inte heller glömma doktorandernas feedbackgrupp som har varit ett trevligt inslag under ett antal eftermiddagar. Gruppen bestod av Sarah, Linda, Hanna, Jon, Helena och mig själv. Som doktorand har man också axlat rollen som lärare. Undervisning och handledning av uppsatser blev ett stort och givande inslag i doktorandutbildningen. Hann a, Monica Andersson Bäck, Karin Allard, Carl - Christian Trönnberg och Dietmar Rauch är några som varit kursledare och kurskollegor på bland annat personal- och ar- betsvetarprogrammet, men också andra kurser. Under min tid som doktorand har jag också varit d el av den F ortefinansierade nationella forskarskolan för åldrande och hälsa (SWEAH) samt centrum för åldrande och hälsa (AGECAP). Med AGECAP och SWEAH har det blivit flera träffar runt om i landet. 2017 erhöll jag Svensk Nationell datatjänst stipendium för metodkurser vid sommarskolan i ICPSR i Ann Arbor, Michigan. Det övergripande forskningsprojektet har genomförts inom ramen för det Fortefinansierade projektet, ” Möjliggöra deltagande, hälsa och välbe- finnande hos äldre anställda - för ett hållbart och inkl uderande arbetsliv” (Forte Dnr. 2012- 1696). Tack till alla som deltog i ArbetsKraftsprojektet och Göteborgsenkäten. Tack till Peter och Sixten Haage som hjälpte till med torrnålsgravyren för bokens omslag, en tavla som jag ibland betraktade när jag var lit en. E tt tack också till Ronnie Pingel som via forumstrådar 2006 motiverade mig att börja läsa och ta examen i statistik, något som hjälpt på många sätt inom och utanför akademin. I mitt privata liv har avhandlingsarbetet varit föremål för maximalt två timmars diskussion de senaste åren, trots stor nyfikenhet från mina föräldrar Conny och Lena , samt systrar Sandra och Felicia. Det finns mycket annat som är mer intressant att prata om, men nu hoppas jag denna bok ger en lite bättre bild. Mina föräldrar ha r gett en trygg uppväxt, sunda värderingar och uppmuntran, som varit helt avgörande i mina grund - och dokto- randstudier. Några som har ännu mindre koll på avhandlingsarbetet är vännerna i Team SG . Med SG höjs dagligen humöret på distans eller tillsammans , Tack till Shahab A, Daniel H, Patrik S, Per T, Daniel A, Martin O, Eric E, Alexander K och Christian L ! En hälsning skickas också till min vän Marcus L ! Bäst under dessa år som doktorand är såklart bildandet av en egen familj och mina under- bara barn Alva och Cornelis. Alva kom till världen precis i början av min doktorandtid vii medan Cornelis kom några år senare. När detta skrivs är de två respektive fem och ett halvt år år. De både tar och ger energi samtidigt. Det är först den senaste månaden som Alva har blivit varse om att sin pappa skriver en bok. Alva väntar med spänning på en bildbok. Slutligen, inte minst , min älskade Klara som gett stöd och mening i framgång och motgång, från glädje och sorg, till hydda och borg. Det är slutet som gillas . Men än är det inte slut. Robin Jonsson Karlstad, november 2021 1 1 A call for Workforce M obilization Being born, beginning school, getting one’s first job, starting to work, and continuing to work until retirement— that’ s a familiar life course for many of us. With this life - course perspective in mind, participating in the labor market is central to western societies and adult life, shaping our lives and bodies until we migrate to a life outside work as a retiree. In the last two deca des, older people ’ s health and employment participation have in- creased substantially in Sweden and other European countries (Eurostat, 2018 ) . In 2019, the Swedish average expected working life duration was the highest in Europe at 42.0 years, versus the EU average of 35.9 years (Eurostat, 2021a) . Although the 65 - year thresh- old still constitutes a strong normative reference point and retirement age for many people in most countries, certain groups face unequal opportunities to continue working or retire based on, for example, previous work histories, sex, health, social relations , and financial resources (Halleröd et al., 2013; Hasselhorn, 2020; Kadefors et al., 2018b; König et al., 2020) . There are no simple explanations for the variation in retirement age, with financial resources, social relations, longevity, and general health all playing roles, but important determinants are to be found in the labor market: the work tasks we pursue, the work environments we are exposed to, and how the work is organized and managed where we work. This section entitled “ A call for workforce mobilization” reflects the broad and prioritized goals of extending the working life and increasing the employment participation of older workers in Sweden (see, for example, Ds, 2013:19; SOU, 2012; SOU, 2013) , the Euro- pean Union (European Council, 2010) , and the OECD (2006) . Efforts to mobilize the workforce for a prolonged working life are being made in response to the financial risks associated with a situation in which a smaller share of the economically active population in the labor market must support a larger share of the inactive. Moreover, recent forecasts point to significant skill needs and labor shortages, not least in the public welfare sector in Europe (Brunello and Wruuck, 2019) and Sweden (Ahlberg et al., 2019; SALAR, 2020) . After years of investigations and negotiations, the Pension Group, consisting of repre- sentatives from most elected parties in the Swedish parliament, reached an agreement in February 2019. In October 2019, parliament voted to reform the pension system, includ- ing a gradual increase in the threshold for obtaining state pension. The earliest withdrawal from the labor market to the state pension will be changed from 62 years in 2020 to 64 years in 2026. This threshold is expected to increase, tracking the average life expectancy, 1 1 A call for Workforce M obilization Being born, beginning school, getting one’s first job, starting to work, and continuing to work until retirement—that ’ s a familiar life course for many of us. With this life -course perspective in mind, participating in the labor market is central to western societies and adult life, shaping our lives and bodies until we migrate to a life outside work as a retiree. In the last two deca des, older people ’ s health and employment participation have in- creased substantially in Sweden and other European countries (Eurostat, 2018 ) . In 2019, the Swedish average expected working life duration was the highest in Europe at 42.0 years, versus the EU average of 35.9 years (Eurostat, 2021a) . Although the 65 -year thresh- old still constitutes a strong normative reference point and retirement age for many people in most countries, certain groups face unequal opportunities to continue working or retire based on, for example, previous work histories, sex, health, social relations , and financial resources (Halleröd et al., 2013; Hasselhorn, 2020; Kadefors et al., 2018b; König et al., 2020) . There are no simple explanations for the variation in retirement age, with financial resources, social relations, longevity, and general health all playing roles, but important determinants are to be found in the labor market: the work tasks we pursue, the work environments we are exposed to, and how the work is organized and managed where we work. This section entitled “ A call for workforce mobilization” reflects the broad and prioritized goals of extending the working life and increasing the employment participation of older workers in Sweden (see, for example, Ds, 2013:19; SOU, 2012; SOU, 2013) , the Euro- pean Union (European Council, 2010) , and the OECD (2006) . Efforts to mobilize the workforce for a prolonged working life are being made in response to the financial risks associated with a situation in which a smaller share of the economically active population in the labor market must support a larger share of the inactive. Moreover, recent forecasts point to significant skill needs and labor shortages, not least in the public welfare sector in Europe (Brunello and Wruuck, 2019) and Sweden (Ahlberg et al., 2019; SALAR, 2020). After years of investigations and negotiations, the Pension Group, consisting of repre- sentatives from most elected parties in the Swedish parliament, reached an agreement in February 2019. In October 2019, parliament voted to reform the pension system, includ- ing a gradual increase in the threshold for obtaining state pension. The earliest withdrawal from the labor market to the state pension will be changed from 62 years in 2020 to 64 years in 2026. This threshold is expected to increase, tracking the average life expectancy, 2 currently at a rate of 0.9, meaning that people born in 1986, like myself, are expected to be eligible for public pension at the age of 70 (Swedish Parliament, 2019) . The legislative actions to prolong working lives (PWL) —sometimes referred to as active aging policies— outlined in Sweden and many European countries require older workers to remain in the workforce and maintain their motivation and abilities as long as possible. Given the PWL policies and the rapidly aging workforce, researchers have advocated work-specifi c interventions and tailored human resource (HR) practices to accommodate a heterogeneous group of older workers (Bal et al., 2015; Hasselhorn and Apt, 2015; Ilmarinen, 2006; Kooij et al., 2014; Truxillo et al., 2015) . These actions are often sub- sumed under the umbrella term “ age management,” although other related concepts exist in the research literature, for example, “ age-awareness policies ,” “ senior policies, ” or “ age- friendly policies.” Age -management strateg y refers to an organization’ s explicit focus on age in human resource management (HRM) practices. Ilmarinen (2006) defined age man- agement as “ taking the employee’ s age and age-related factors into account in daily work management, work planning, and work for the organization; thus everyone —regar dless of age—can achieve personal and organizational targets healthily and sa fely” (Ilmarinen, 2006, p. 120). Although this definition of age management refers to all age groups, previous research has primarily used the concept in the context of older workers’ employment participation. There are two basic premises in the age -management literature : (1) older workers are unique and have specific needs and preferences; and (2) employers can respond by adapt- ing, modifying, developing, and enhancing flexibility in accordance with older workers’ abilities and preferences (A variety of different conceptualizations on age management, see Eppler -Hattab et al., 2019; Kooij et al., 2014; Wilckens et al., 2020; Zacher and Yang, 2016) . Previous research shows that early retirement is often caused by the working envi- ronment not being sufficiently adapted to older workers’ needs. Accordingly , age -man- agement strategies are assumed to improve the person– job fit and strengthen an individ- ual’ s motivation and ability to continue working. Even though age-management strategies have been part of the Eur opean Union’ s policy agenda for over two decades (Foster and Walker, 2015; Walker, 2005) , evidence suggests that organizations often rely on universal HR practices and take a relatively passive or reactive approach to implementing measures to retain older employees (Conen et al., 2012; Fuertes et al., 2013; Principi et al., 2020; Vickerstaff et al., 2003) . At the same time, contemporary knowledge of effective work- place interventions to promote older workers’ employability, work ability, and health is sparse (Cloostermans et al., 2015; Hilsen and Midtsundstad, 2015; Söderbacka et al., 2020; Truxillo et al., 2015) . T he increasing share of older workers in the European labor market in the last two dec- ades leaves a bitter aftertaste (Eurostat, 2018) . Despite improvements in population health 3 and increasing longevity, a considerable proportion of the workforce is expected to ac- quire work - related disorders and other chronic diseases. To illustrate, 46.7 % of 55 – 64 - year- olds in the EU state d that they had a long- standing illness or health problem, and 38.3 % reported that their self- perceived health was fair, bad, or very bad in 2019 (Eurostat, 2021c; Eurostat, 2021d) . At the same time, stricter eligibility rules have been imposed for sickness compensation (formerly referred to as “ disability pension”) in the last few decades in Sweden (Johansson et al., 2018) , and eligibility has been similarly tight- ened in other European countries (Hinrichs, 2021) . It is well established that people in lower occupational strata and with poor health exit the workforce prematurely (Kadefors et al., 2017; Rad et al., 2017; Topa et al., 2018) . Recent studies consolidate these associ- ations, revealing inequalities in retirement behavior based on, for example, health, occu- pation, education, and gender (Kadefors et al., 2018b; König et al., 2018; McAllister et al., 2020) . In relation to universal PWL policies, scholars acknowledge that factors such as poor health, skill deficiency, and lack of work ability among older workers pose a sig- nificant challenge to realizing a shift toward the sustainable prolongation of working life (Hasselhorn, 2020; Ilmarinen, 2006; Krekula et al., 2017; Nilsson et al., 2017) . In con- trast, studies also demonstrate that the voluntary prolongation of working life is not nec- essarily detrimental to older people’ s health, physical functioning , or wellbeing (Eyjólfsdóttir et al., 2019a; Prakash et al., 2021; Stenholm et al., 2014) , and most people express positive reasons for their decision to continue working or retire from the labor market (König et al., 20 20) . Still, the sustainable extension of working life requires that individuals maintain their work ability and motivation to continue working, while organizations must recognize the aging workforce as a resource and be able to manage older employees’ heterogeneous needs and desires. Here, the Swedish welfare organizations, in this thesis exemplified by healthcare, eldercare, and other municipal services, can serve as an interesting case. These services are vital for society, are predominantly publicly driven, employ a large proportion of the total workforce, and can act as a role model for other sectors and branches . However, the Swedish public welfare organizations face multiple challenges related to recruitment dif- ficulties and increasing demand for welfare services. The Swedish Association of Local Authorities and Regions ( SAL AR ) has outlined several strategies to mitigate the recruit- ment needs, including a call for public and private employers to prioritize recruiting and retaining older healthcare workers (SALAR, 2018b; SALAR, 2020) . For labor market sectors experiencing labor shortages, such as Swedish welfare organizati ons, the ability to recruit and retain older workers might depend on their capability to provide stimulating, attractive, and decent work environments. However, despite overarching ideas and strat- egies concerning extended working lives, the Swedish welfare organizations’ capability to manage and retain older workers remains uncharted. Against this background, th is thesis seeks insights into enabling and hindering conditions at the individual and organizational 2 currently at a rate of 0.9, meaning that people born in 1986, like myself, are expected to be eligible for public pension at the age of 70 (Swedish Parliament, 2019) . The legislative actions to prolong working lives (PWL) — sometimes referred to as active aging policies— outlined in Sweden and many European countries require older workers to remain in the workforce and maintain their motivation and abilities as long as possible. Given the PWL policies and the rapidly aging workforce, researchers have advocated work- specific interventions and tailored human resource (HR) practices to accommodate a heterogeneous group of older workers (Bal et al., 2015; Hasselhorn and Apt, 2015; Ilmarinen, 2006; Kooij et al., 2014; Truxillo et al., 2015) . These actions are often sub- sumed under the umbrella term “ age management,” although other related concepts exist in the research literature, for example, “ age- awareness policies,” “ senior policies, ” or “ age- friendly policies.” Age - management strategy refers to an organization’ s explicit focus on age in human resource management (HRM) practices. Ilmarinen (2006) defined age man- agement as “ taking the employee’ s age and age- related factors into account in daily work management, work planning, and work for the organization; thus everyone — regardless of age— can achieve personal and organizational targets healthily and safely” (Ilmarinen, 2006, p. 120). Although this definition of age management refers to all age groups, previous research has primarily used the concept in the context of older workers’ employment participation. There are two basic premises in the age - management literature: (1) older workers are unique and have specific needs and preferences; and (2) employers can respond by adapt- ing, modifying, developing, and enhancing flexibility in accordance with older workers’ abilities and preferences (A variety of different conceptualizations on age management, see Eppler - Hattab et al., 2019; Kooij et al., 2014; Wilckens et al., 2020; Zacher and Yang, 2016) . Previous research shows that early retirement is often caused by the working envi- ronment not being sufficiently adapted to older workers’ needs. Accordingly , age - man- agement strategies are assumed to improve the person– job fit and strengthen an individ- ual’ s motivation and ability to continue working. Even though age- management strategies have been part of the Eur opean Union’ s policy agenda for over two decades (Foster and Walker, 2015; Walker, 2005) , evidence suggests that organizations often rely on universal HR practices and take a relatively passive or reactive approach to implementing measures to retain older employees (Conen et al., 2012; Fuertes et al., 2013; Principi et al., 2020; Vickerstaff et al., 2003) . At the same time, contemporary knowledge of effective work- place interventions to promote older workers’ employability, work ability, and health is sparse (Cloostermans et al., 2015; Hilsen and Midtsundstad, 2015; Söderbacka et al., 2020; Truxillo et al., 2015) . T he increasing share of older workers in the European labor market in the last two dec- ades leaves a bitter aftertaste (Eurostat, 2018) . Despite improvements in population health 3 and increasing longevity, a considerable proportion of the workforce is expected to ac- quire work -related disorders and other chronic diseases . To illustrate, 46.7% of 55 – 64- year-olds in the EU state d that they had a long-standing illness or health problem, and 38.3% reported that their self-perceived health wa s fair, bad, or very bad in 2019 (Eurostat, 2021c; Eurostat, 2021d) . At the same time, stricter eligibility rules have been imposed for sickness compensation (formerly referred to as “ disability pension”) in the last few decades in Sweden (Johansson et al., 2018) , and eligibility has been similarly tight- ened in other European countries (Hinrichs, 2021) . It is well established that people in lower occupational strata and with poor health exit the workforce prematurely (Kadefors et al., 2017; Rad et al., 2017; Topa et al., 2018) . Recent studies consolidate these associ- ations, revealing inequalities in retirement behavior based on, for example, health, occu- pation, education, and gender (Kadefors et al., 2018b; König et al., 2018; McAllister et al., 2020) . In relation to universal PWL policies, scholars acknowledge that factors such as poor health, skill deficiency, and lack of work ability among older workers pose a sig- nificant challenge to realizing a shift toward the sustainable prolongation of working life (Hasselhorn, 2020; Ilmarinen, 2006; Krekula et al., 2017; Nilsson et al., 2017) . In con- trast, studies also demonstrate that the voluntary prolongation of working life is not nec- essarily detrimental to older people’s health, physical functioning , or wellbeing (Eyjólfsdóttir et al., 2019a; Prakash et al., 2021; Stenholm et al., 2014) , and most people express positive reasons for their decision to continue working or retire from the labor market (König et al., 20 20) . Still, the sustainable extension of working life requires that individuals maintain their work ability and motivation to continue working, while organizations must recognize the aging workforce as a resource and be able to manage older employees’ heterogeneous needs and desires. Here, the Swedish welfare organizations, in this thesis exemplified by healthcare, eldercare, and other municipal services, can serve as an interesting case. These services are vital for society, are predominantly publicly driven, employ a large proportion of the total workforce, and can act as a role model for other sectors and branches . However, the Swedish public welfare organizations face multiple challenges related to recruitment dif- ficulties and increasing demand for welfare services. The Swedish Association of Local Authorities and Regions ( SAL AR) has outlined several strategies to mitigate the recruit- ment needs, including a call for public and private employers to prioritize recruiting and retaining older healthcare workers (SALAR, 2018b; SALAR, 2020) . For labor market sectors experiencing labor shortages, such as Swedish welfare organizati ons, the ability to recruit and retain older workers might depend on their capability to provide stimulating, attractive, and decent work environments. However, despite overarching ideas and strat- egies concerning extended working lives, the Swedish welfare organizations’ capability to manage and retain older workers remains uncharted. Against this background, th is thesis seeks insights into enabling and hindering conditions at the individual and organizational 4 levels, using capability as a theoretical approach. Capability refers to individuals ’ and or- ganizations’ ability to convert existing resources and perform actions to realize selected goals, such as continuing working or retaining older workers, respectively. Aim s and objectives This thesis seeks to investigate how older workers’ decisions to prolong or end their work- ing lives are formed through interaction between, on one hand, personal and work -related conditions and, on the other, organizations’ capabilities to implement age-management practices. Using a capability approach as a theoretical framework, conditions that influ- ence an individual’ s capabilities to work and the organization’ s capabilities to implement age-management measures are studied first separately and then in combination. Conse- quently, by combining studies of personal and work -related determinants of retirement with studies of organizational opportunities and constraints, the thesis aims to contextu- alize individuals’ retirement decisions and understand how these conditions could influ- ence one another. The specific research questions address (a) individual capability , (b) organizational capability, and their interaction (c) in the context of the prolonged working life: A. What personal and work-related factors promote or hinder older workers’ ability and willingness to continue working? (Article s I – III ) B. What barriers and opportunities influence public welfare organizations’ capability to retain older workers? (Article s III and IV) C. How do individuals’ capabilities interact with organizational capabilities in the con- text of prolonging the working life? (Article s I – IV ) The se research questions are examined using a combination of quantitative and qualitative methods. Three data sources draw on a representative sample of working Swedes aged 55– 6 4 years, Swedish public welfare employees aged 55 years or older, and transcribed interviews with line managers and their HR partners in the Swedish healthcare sector. In addition, I integrate prior literature from the interdisciplinary field of work sciences and HRM to discuss the underlying mechanisms that enable or hinder prolonged working life at the individual, organizational, and societal levels, which is crucial to mobiliz ing the workforce. 5 Organization of the thesis T his thesis is organized as follows. The second section briefly presents the historical back- ground to prolonged working life policies and this case study: public welfare organizations in Sweden. In the third section, the theoretical perspective— individual and organizational capability— is outlined. T he section ends with a discussion of the interaction between in- dividual and organizational capabilities. In section 4 , retirement determinants at the soci- etal, workplace, and individual level s are presented. Finally, i n section 5 , the methods and data used in the qualitative and quantitative studies are presented in detail, followed by the “Summary of results” ( section 6 ), “D iscussion” ( section 7 ), and “C onclusion” ( section 8 ). 4 levels, using capability as a theoretical approach. Capability refers to individuals ’ and or- ganizations’ ability to convert existing resources and perform actions to realize selected goals, such as continuing working or retaining older workers, respectively. Aim s and objectives This thesis seeks to investigate how older workers’ decisions to prolong or end their work- ing lives are formed through interaction between, on one hand, personal and work - related conditions and, on the other, organizations’ capabilities to implement age- management practices. Using a capability approach as a theoretical framework, conditions that influ- ence an individual’ s capabilities to work and the organization’ s capabilities to implement age- management measures are studied first separately and then in combination. Conse- quently, by combining studies of personal and work - related determinants of retirement with studies of organizational opportunities and constraints, the thesis aims to contextu- alize individuals’ retirement decisions and understand how these conditions could influ- ence one another. The specific research questions address (a) individual capability , (b) organizational capability, and their interaction (c) in the context of the prolonged working life: A. What personal and work-related factors promote or hinder older workers’ ability and willingness to continue working? (Article s I – III ) B. What barriers and opportunities influence public welfare organizations’ capability to retain older workers? (Article s III and IV) C. How do individuals’ capabilities interact with organizational capabilities in the con- text of prolonging the working life? (Article s I – IV ) The se research questions are examined using a combination of quantitative and qualitative methods. Three data sources draw on a representative sample of working Swedes aged 55 – 6 4 years, Swedish public welfare employees aged 55 years or older, and transcribed interviews with line managers and their HR partners in the Swedish healthcare sector. In addition, I integrate prior literature from the interdisciplinary field of work sciences and HRM to discuss the underlying mechanisms that enable or hinder prolonged working life at the individual, organizational, and societal levels, which is crucial to mobiliz ing the workforce. 5 Organization of the thesis T his thesis is organized as follows. The second section briefly presents the historical back- ground to prolonged working life policies and this case study: public welfare organizations in Sweden. In the third section, the theoretical perspective—individual and organizational capability—is outlined. T he section ends with a discussion of the interaction between in- dividual and organizational capabilities. In section 4 , retirement determinants at the soci- etal, workplace, and individual level s are presented. Finally, i n section 5 , the methods and data used in the qualitative and quantitative studies are presented in detail, followed by the “Summary of results” ( section 6 ), “D iscussion” ( section 7 ), and “C onclusion” ( section 8 ). 6 2 On working longer in life The world ’ s population is aging. Decreasing fertility rates, increasing longevity, adva nces in medicine, better health care, and improved nutrition are some factors driving these de- velopments (United Nations, 2017) . In many countries with universal public pension sys- tems, workforce aging has become a concern because the ratio between the economically active contributors and the non-active beneficiaries in the labor market —often referred to as the old- age dependency ratio—threatens fina ncial sustainability (Bengtsson and Scott, 2011; Hofäcker et al., 2016) . The European Centre for the Development of Voca- tional Training has identified labor shortages and skill needs that jeopardize Europ ean companies’ ability to sustain knowledge and to remain productive and competitive (Brunello and Wruuck, 2019) . The Organization for Economic Co -operation and Devel- opment (OECD) and the European Union (EU) both point to the same remedy: the workforce should work more and work longer (Axelrad and Mahoney, 2017; Eurostat, 2018) . Since the introduction of PWL policies, employment participat ion among older workers aged 55 – 64 years has increased substantially in the EU, from 38.1 % in 2002 to 57.7% in 2018 (Eurostat, 2018) . For the European country in focus here, Sweden, the 2019 employment rate of 55– 6 4 -year -olds is the highest in the EU at 77.7 % (Eurostat, 2021b) . Toward the mobilization of the Swedish workforce: a policy shift During the 20th century, Sweden experience d a remarkable transition from being an agri- culturally based economy to an industrialized modern welfare state with an unbroken thread of improvements in people’ s working and living conditions. The emergence of an industrialized economy and a universal and tax-financed welfare state change d not only how people made their livings through employment in the labor market, but also how people supported themselves after the end of employment participation. After being the first country to introduce and institutionalize a universal public pension in 1913 to reduce old-age poverty, the Swedish pension system became a natural component of the social security system. Thus, while the system helped to reduce old-age poverty, it simultane- ously strengthened people’ s dependence on the welfare state. However, by the end of the 20th century, increasing longevity, falling fertility, and concerns about the “ baby boom” generation born in the 1940s leaving the labor market raised questions about the financial 7 stability of the pension system and shifted the political orientation toward workforce pro- longation in Sweden (SOU, 2012) . A similar political shift occurred in many other Euro- pean countries (Hofäcker and Radl, 2016) . The Swedish debate on retirement age and pension system reform has been ongoing for several decades, engaging various political parties and labor market interest groups. In many ways, the PWL policies can be said to be integrated into the longstanding Swedish work strategy (Swedish: arbetslinjen ), according to which individuals are expected to support themselves by means of work and high employment rates are at the center of labor market policies. From a national economic perspective, policymaker s’ focus on sustaining high employment rates is motivated by the fact that about 60 % of Swedish tax revenues origi- nate from payroll and income taxes (ESV, 2021; appendix 1) . Although most actors acknowledge the necessity of adjusting the pension system to secure financial stability, opinions vary as to how this should be achieved. Among labor unions, the general re- sistance to raising the retirement age relates to workers’ abilities and willingness to work longer. In contrast, the resistance of the employers ’ federation primarily concerns main- taining employment flexibility and concerns about additional expenses for companies. In relation to the relatively moderate debate between labor unions, the employers’ federa- tions, and political parties , several interest groups and debaters have accused the current pension system of not recognizing structural gaps in the labor market and of contributing to societal tensions. These interest groups and debaters, for instance, all point to inade- quate pensions, past income inequalities and differences in employment participation be- tween men and women, socioeconomic differences in working conditions, and the con- sequences for migrants with a shorter registered working period after their relocation to Sweden (see, for example, Dahlberg, 2013; Mårtensson, 2018; Mårtensson, 2020; PRO, 2015; Rundström, 2017; Sangregorio, 201 8; Scherman, 2014; SVT, 2016) . Moreover, Swedish researchers have urged policymakers to improve working conditions and job quality in the labor market and consider people ’ s health and gender when designing poli- cies to extend working lives (Krekula et al., 2017; Nilsson et al., 2017) . Similar concerns have been raised in relation to universal PWL policies outside of Sweden (for an overview, see, Hasselhorn, 2020) . Besides the legislative adjustment s of retirement age and employment protection dis- cussed above, discrimination based on age became prohibited in 2009. Moreover, several regulations concerning the work environment have been strengthened in recent years. For example, the Swedish Work Environment Authority upda ted and added 57 regulations concerning work adaptation, medical checkups, chemical exposure, and workplace design between 2015 and 2020 (SWEA, 2020b) . In 2015, the regulation concerning the organiza- tional and social work environment was of particular importance, as it stipulates the employer’ s responsibility for preventing and managing unhealthy workloads, working times, and vic- timization at Swedish workplaces (SWEA, 2016) . To enhance knowledge and identify problems related to prolonged working life at the individual, organizational , and societal 6 2 On working longer in life The world ’ s population is aging. Decreasing fertility rates, increasing longevity, adva nces in medicine, better health care, and improved nutrition are some factors driving these de- velopments (United Nations, 2017) . In many countries with universal public pension sys- tems, workforce aging has become a concern because the ratio between the economically active contributors and the non- active beneficiaries in the labor market— often referred to as the old- age dependency ratio— threatens financial sustainability (Bengtsson and Scott, 2011; Hofäcker et al., 2016) . The European Centre for the Development of Voca- tional Training has identified labor shortages and skill needs that jeopardize Europ ean companies’ ability to sustain knowledge and to remain productive and competitive (Brunello and Wruuck, 2019) . The Organization for Economic Co - operation and Devel- opment (OECD) and the European Union (EU) both point to the same remedy: the workforce should work more and work longer (Axelrad and Mahoney, 2017; Eurostat, 2018) . Since the introduction of PWL policies, employment participat ion among older workers aged 55 – 64 years has increased substantially in the EU, from 38.1 % in 2002 to 57.7 % in 2018 (Eurostat, 2018) . For the European country in focus here, Sweden, the 2019 employment rate of 55 – 6 4 - year- olds is the highest in the EU at 77.7 % (Eurostat, 2021b) . Toward the mobilization of the Swedish workforce: a policy shift During the 20th century, Sweden experience d a remarkable transition from being an agri- culturally based economy to an industrialized modern welfare state with an unbroken thread of improvements in people’ s working and living conditions. The emergence of an industrialized economy and a universal and tax- financed welfare state changed not only how people made their livings through employment in the labor market, but also how people supported themselves after the end of employment participation. After being the first country to introduce and institutionalize a universal public pension in 1913 to reduce old- age poverty, the Swedish pension system became a natural component of the social security system. Thus, while the system helped to reduce old- age poverty, it simultane- ously strengthened people’ s dependence on the welfare state. However, by the end of the 20th century, increasing longevity, falling fertility, and concerns about the “ baby boom” generation born in the 1940s leaving the labor market raised questions about the financial 7 stability of the pension system and shifted the political orientation toward workforce pro- longation in Sweden (SOU, 2012) . A similar political shift occurred in many other Euro- pean countries (Hofäcker and Radl, 2016) . The Swedish debate on retirement age and pension system reform has been ongoing for several decades, engaging various political parties and labor market interest groups. In many ways, the PWL policies can be said to be integrated into the longstanding Swedish work strategy (Swedish: arbetslinjen ), according to which individuals are expected to support themselves by means of work and high employment rates are at the center of labor market policies. From a national economic perspective, policymaker s’ focus on sustaining high employment rates is motivated by the fact that about 60 % of Swedish tax revenues origi- nate from payroll and income taxes (ESV, 2021; appendix 1) . Although most actors acknowledge the necessity of adjusting the pension system to secure financial stability, opinions vary as to how this should be achieved. Among labor unions, the general re- sistance to raising the retirement age relates to workers’ abilities and willingness to work longer. In contrast, the resistance of the employers ’ federation primarily concerns main- taining employment flexibility and concerns about additional expenses for companies. In relation to the relatively moderate debate between labor unions, the employers’ federa- tions, and political parties , several interest groups and debaters have accused the current pension system of not recognizing structural gaps in the labor market and of contributing to societal tensions. These interest groups and debaters, for instance, all point to inade- quate pensions, past income inequalities and differences in employment participation be- tween men and women, socioeconomic differences in working conditions, and the con- sequences for migrants with a shorter registered working period after their relocation to Sweden (see, for example, Dahlberg, 2013; Mårtensson, 2018; Mårtensson, 2020; PRO, 2015; Rundström, 2017; Sangregorio, 201 8; Scherman, 2014; SVT, 2016) . Moreover, Swedish researchers have urged policymakers to improve working conditions and job quality in the labor market and consider people ’ s health and gender when designing poli- cies to extend working lives (Krekula et al., 2017; Nilsson et al., 2017) . Similar concerns have been raised in relation to universal PWL policies outside of Sweden (for an overview, see, Hasselhorn, 2020) . Besides the legislative adjustment s of retirement age and employment protection dis- cussed above, discrimination based on age became prohibited in 2009. Moreover, several regulations concerning the work environment have been strengthened in recent years. For example, the Swedish Work Environment Authority upda ted and added 57 regulations concerning work adaptation, medical checkups, chemical exposure, and workplace design between 2015 and 2020 (SWEA, 2020b) . In 2015, the regulation concerning the organiza- tional and social work environment was of particular importance, as it stipulates the employer’ s responsibility for preventing and managing unhealthy workloads, working times, and vic- timization at Swedish workplaces (SWEA, 2016) . To enhance knowledge and identify problems related to prolonged working life at the individual, organizational , and societal 8 levels, in the last decade, the Swedish government conducted two investigations (SOU, 2012; SOU, 20 13; SOU, 2020) and produced several thematic reports (e.g., SWEA, 2012b; SWEA, 2014) . As discussed above, the Swedish government’ s response to work- force aging has emphasized public policy adjustments of the overall social security system and work environment regulations. In practice, several private - and public-financed work- place interventions have been carried out, often research driven and seemingly inde- pendently of one another. Although no existing review cover s all such interventions in Sweden, several interventions can be traced in previous research articles and reports. For example, several Swedish municipalities have implemented reduced working time and fi- nancial benefits for older workers ( “ the 80 – 90– 100 model” and “ Attraktiva seniorer ” ) and age-management education (“BestAgers”) for managers in public dental care (Domschke et al., 2014) . In the latest official report from the Swedish government concerning pro- longed working life entitled Older peopl e have never been younger (SOU, 2020) , the delegation for senior labor noted several favorable conditions for sustainably prolonging the working life. For instance, the workforce is healthier, more educated, and less exposed to hazard- ous working conditions, and extend ed employment participation is economically benefi- cial for the individual. Still, the delegation for senior labor identified several remaining obstacles related to age discrimination and ageism, namely, employers’ negative attitudes to older workers and a lack of age awareness in personnel policies, norms, and procedures that nudge people to retire at 65. The Swedish pension system: historical background Before the implementation of the public pension system in 1913, there was a mix ture of various pension arrangements and solutions in Sweden. 1 The earliest forms of pension originated in the Swedish Navy in the 17 th century and provided veterans with some in- come security after service. However, for most of the population living in the countryside, such as farmers, pension security relied on family and private assets up to and beyond the introduction of the public pension system. The rapid industrialization, urbanization, and expansion of the private and public sectors starting in the 19 th century paved the way for new private pension solutions. For instance, profession - specific pensions were financed through voluntary and mandatory contributions and attracted an increasing number of teachers, bakers, railway workers, civil servants, and other occupational groups. Many of these occupational pension arrangements survived and supplemented the public pension in the decades following the 1913 pension reform (Hagen, 2013) . Despite neither the left- nor right-wing parties in the S wedish parliament being fully satisfied with the public pen- sion reform in 1913, it remained unchanged until 1935. Public criticism of the increasing gap between high- and low-income earners fueled debates about reforming the pension system. Consequently, a universal flat-rate pension, the “ people’ s pension” (Swedish: 1 For a comprehensive historical re view, see Hagen, J. (2013). A history of the S wedish pension system. Uppsala: Uppsala University. 9 Folkpension) , replaced the public pension in 1948 . The universal flat - rate pension was complemented with relatively generous means- tested benefits and improved the overall living standards among retirees. However, it was still considered too low, especially among high- income earners, as the expected pension amounted to only one - third of the industrial sector’ s average wage in 1948. Another consequence of the flat - rate design of the people’ s pension was that pension benefits lagged behind the increasing real wages in the work- force and the increase in the general cost of living (Hagen, 2013) . At that time, certain income and occupational groups already had complementary solutions to supplement the flat- rate pension. The most common form was negotiated occupational pensions that typ- ically provided white- collar workers with additional financial security. In contrast, many blue- collar workers were still limited to the public pension option. Thus, after intense political debates at the end of the 1950s, dissatisfaction with the flat - rate pension’ s ability to provide decent pensions, in combination with the increasing gap between people with and without occupational pensions, ultimately led the Swedish parliament to vote for an earnings- related supplementary pension scheme— the ATP system (Swedish: Allmän Till äggspension). The ATP system was based on earnings - related benefits, i.e., on the 15 years during which the income was highest over 30 years of labor market participation. With the ATP system, disability pensions became an integrated part of the public pension system, and the legislated pensionable age was set to 65 years. However, claiming disabil- ity, workers could withdraw earlier , and retiring for non- medical reasons was possible from age 60 . Since the integration of disability pensions into the ATP system, the eligibility rules for disability pensions have changed over time. In times of higher unemployment rates on the labor market and recession since the 1970s, generous early retirement schemes sanctioned by employers or the state gave people an attractive and socially ac- ceptable opportunity to withdraw from the workforce in Sweden and many other Euro- pean countries (Hofäcker and Radl, 2016) . The access to disability pension can be divided into roughly three phases (Johansson et al., 2018) . In the first phase, before 1970, eligibility for disability pension was strict and required medical reasons. On some occasions, medical reasons in combination with long- term unemployment were also acceptable. In the sec- ond phase, between 1970 and 1997, eligibility rules became more generous and included medical and non- medical reasons, such as certain age groups and long- term unemploy- ment (unemployment was excluded in 1991). In the third phase, starting in 1997 , eligibility rules for disability were once again tightened. Disability pensions were primarily granted based on medical reasons and, on very few occasions, long - term unemployment. Today, disability pension is referred to as sickness compensation and, since 2003, has been part of the health insurance system. Thus, disability pensioners (i.e., people with sickness com- pensation) are automatically transferred from the health insurance system to the public pension system at age 65. Historically, disability pension has been a common retirement path for workers with reduced work ability due to health impairments, and, before 1991, also for long- term unemployment (Johansson et al., 2018) . Between 1968 and 1994 , for instance, disability pensions were the dominant pathway to retirement before the age of 8 levels, in the last decade, the Swedish government conducted two investigations (SOU, 2012; SOU, 20 13; SOU, 2020) and produced several thematic reports (e.g., SWEA, 2012b; SWEA, 2014) . As discussed above, the Swedish government’ s response to work- force aging has emphasized public policy adjustments of the overall social security system and work environment regulations. In practice, several private - and public- financed work- place interventions have been carried out, often research driven and seemingly inde- pendently of one another. Although no existing review cover s all such interventions in Sweden, several interventions can be traced in previous research articles and reports. For example, several Swedish municipalities have implemented reduced working time and fi- nancial benefits for older workers ( “ the 80 – 90 – 100 model” and “ Attraktiva seniorer ” ) and age- management education (“BestAgers”) for managers in public dental care (Domschke et al., 2014) . In the latest official report from the Swedish government concerning pro- longed working life entitled Older peopl e have never been younger (SOU, 2020) , the delegation for senior labor noted several favorable conditions for sustainably prolonging the working life. For instance, the workforce is healthier, more educated, and less exposed to hazard- ous working conditions, and extend ed employment participation is economically benefi- cial for the individual. Still, the delegation for senior labor identified several remaining obstacles related to age discrimination and ageism, namely, employers’ negative attitudes to older workers and a lack of age awareness in personnel policies, norms, and procedures that nudge people to retire at 65. The Swedish pension system: historical background Before the implementation of the public pension system in 1913, there was a mix ture of various pension arrangements and solutions in Sweden. 1 The earliest forms of pension originated in the Swedish Navy in the 17 th century and provided veterans with some in- come security after service. However, for most of the population living in the countryside, such as farmers, pension security relied on family and private assets up to and beyond the introduction of the public pension system. The rapid industrialization, urbanization, and expansion of the private and public sectors starting in the 19 th century paved the way for new private pension solutions. For instance, profession - specific pensions were financed through voluntary and mandatory contributions and attracted an increasing number of teachers, bakers, railway workers, civil servants, and other occupational groups. Many of these occupational pension arrangements survived and supplemented the public pension in the decades following the 1913 pension reform (Hagen, 2013) . Despite neither the left- nor right- wing parties in the Swedish parliament being fully satisfied with the public pen- sion reform in 1913, it remained unchanged until 1935. Public criticism of the increasing gap between high- and low- income earners fueled debates about reforming the pension system. Consequently, a universal flat- rate pension, the “ people’ s pension” (Swedish: 1 For a comprehensive historical re view, see Hagen, J. (2013). A history of the S wedish pension system. Uppsala: Uppsala University. 9 Folkpension) , replaced the public pension in 1948 . The universal flat -rate pension was complemented with relatively generous means-tested benefits and improved the overall living standards among retirees. However, it was still considered too low, especially among high- income earners, as the expected pension amounted to only one -third of the industrial sector’ s average wage in 1948. Another consequence of the flat - rate design of the people’ s pension was that pension benefits lagged behind the increasing real wages in the work- force and the increase in the general cost of living (Hagen, 2013) . At that time, certain income and occupational groups already had complementary solutions to supplement the flat-rate pension. The most common form was negotiated occupational pensions that typ- ically provided white-collar workers with additional financial security. In contrast, many blue-collar workers were still limited to the public pension option. Thus, after intense political debates at the end of the 1950s, dissatisfaction with the flat -rate pension ’ s ability to provide decent pensions, in combination with the increasing gap between people with and without occupational pensions, ultimately led the Swedish parliament to vote for an earnings-related supplementary pension scheme — the ATP system (Swedish: Allmän Till äggspension). The ATP system was based on earnings -related benefits, i.e., on the 15 years during which the income was highest over 30 years of labor market participation. With the ATP system, disability pensions became an integrated part of the public pension system, and the legislated pensionable age was set to 65 years. However, claiming disabil- ity, workers could withdraw earlier , and retiring for non-medical reasons was possible from age 60 . Since the integration of disability pensions into the ATP system, the eligibility rules for disability pensions have changed over time. In times of higher unemployment rates on the labor market and recession since the 1970s, generous early retirement schemes sanctioned by employers or the state gave people an attractive and socially ac- ceptable opportunity to withdraw from the workforce in Sweden and many other Euro- pean countries (Hofäcker and Radl, 2016) . The access to disability pension can be divided into roughly three phases (Johansson et al., 2018) . In the first phase, before 1970, eligibility for disability pension was strict and required medical reasons. On some occasions, medical reasons in combination with long-term unemployment were also acceptable. In the sec- ond phase, between 1970 and 1997, eligibility rules became more generous and included medical and non- medical reasons, such as certain age groups and long-term unemploy- ment (unemployment was excluded in 1991). In the third phase, starting in 1997 , eligibility rules for disability were once again tightened. Disability pensions were primarily granted based on medical reasons and, on very few occasions, long - term unemployment. Today, disability pension is referred to as sickness compensation and, since 2003, has been part of the health insurance system. Thus, disability pensioners (i.e., people with sickness com- pensation) are automatically transferred from the health insurance system to the public pension system at age 65. Historically, disability pension has been a common retirement path for workers with reduced work ability due to health impairments, and, before 1991, also for long-term unemployment (Johansson et al., 2018) . Between 1968 and 1994 , for instance, disability pensions were the dominant pathway to retirement before the age of 10 65 (Hagen, 2013) . During the recession in the 1990s, disability pensions became an early retirement option for many people. However, stricter eligibility requirements were im- posed in 2008 that dramatically reduced the number of disability pensioners , i.e., people receiving sickness compensation. Research indicates that these reforms left some people suffering from impaired health with few options but to continue working or to take the economic consequences on themselves and leav e the labor market prematurely (Kadefors et al., 2018a). The most critical policy reform was in 1999 when the current pension system, referred to as the Notional Defined Contribution System (NDC), was introduced (Kruse, 2010) . The NDC system increased the connection between the length of an individual’ s working life (i.e., contributions) and pension benefits. The NDC is financed through a combination of individual accounts and ongoing tax payments from the current workforce, in what is sometimes called a pay-as -you - go arrangement, and is based on intergenerational trans- fers. The current pension system consists of three parts : the state pension ( comprising the income pension, the premium pension scheme, and , in some instances, the minimum guaranteed pension, which is a means -tested pension for people with no or low prior labor-force attachment), the occupational pension , and private pension savings. 2 The legislated pensionable age of 65 years was abandoned. The new system provided some flexibility, as benefits from the income pension allowed workers to withdraw from work to retirement starting at age 61 and legislative rights in the Employment Protection Act allowed them to work until age 67. According to the 2019 pension agreement, however, these age thresholds for receiving old-age pension will gradually increase from 62 in 20 20 to 64 by 2026 and, after that, be linked to the average life expectancy advance in the Swedish pop- ulation (Ds, 2019:2) . Al though the pension system still allows for earlier or later with- drawal than at the age of 65, the 65 -year threshold still constitutes a strong normative reference point due to its historical origin (Johansson et al., 2018). The 65 -year threshold is often used to distinguish whether an individual’ s retirement is “ early, ” “ on time, ” or “ late” in the public debate and in the research literature (Fisher et al., 2016) . 2 The state pension primar ily consists of an income pension and premium pension. Every year, 16% of the pensionable income is allocated to the income pension, while 2.5% of the pensionable income is as signed to the premium pension ( the total contribution rate is 18.5%). The premium pensions are allocated to stock market funds, but individuals can allocate their money to the fund(s) of their own choice. Most employers on the Swe- dish labor market provide an occupational pension to their employees. There are several occupational pension schemes, depending on the branch and sector, which are negotiated between the unions, employers , and em- ployers’ federations. Outside the pension system, some people also have private savings that can be used once they retire. This is of particular importance for the self-employed or for workers without an occupational pension provided by their employer, as private savings replace the occupational pension ( Kruse, A. (2010). A stable pension system: The eighth wonder. In: Bengtsson, T. (ed.) Population Ageing -A Threat to the Welfare State? Switzerland: Springer. ) . 11 Structural developments and challenges in the Swedish labor market Since the end of the 20th century, globalization and technological advance s have contrib- uted to significant changes in the Swedish labor market. These structural developments also interacted with and conditioned employment for men and women. Particularly nota- ble is the decline in male- dominated industrial production in favor of service- based indus- tries. The decline in industrial pr oduction was accompanied by stricter and more compre- hensive work environment regulations during the 1970s and 1980s, the establishment of several national institutes of occupational safety and health, and new legislation requiring occupational health services in workplaces. These developments reduced the number of physically strenuous work tasks and common occupational hazards such as chemical risks and noise pollution (Johansson et al., 2018) . However, improvement in the work environ- ment has not been equally distributed in all labor market sectors. During the recession of the 1990s, Swedish public and private employers cut back on jobs and carried out a re- structuring that led to higher job demands at many workplace s. Retrospective studies of the 1997 – 2015 period indicate an overall stable trend in job demands and resources, alt- hough certain occupational groups deviated from this pattern (Corin et al., 2021) . More- over, two other longitudinal analyses show that , since the early 1990s, the Swedish labor market has witnessed increasing job dema nds, particularly in female - dominated occupa- tions such as health, social care , education (Aronsson et al., 2021; Cerdas et al., 2019) . In many of these female- dominated sectors, sick leave rates escalated quickly and have re- mained high until today (SSIA, 2015) . According to 2019 data, women have a 25 % higher risk of sick leave and a 41 % higher risk of stress- related mental illness than do men across all occupational groups (SSIA, 2020) . The life - course trajectory that introduced the first section— growing up, acquiring an ed- ucation, getting a job, and ultimately retiring — can be accused of being oversimplified and reflecting a largely male perspective on working life. Feminists and other scholars accuse prolonged working life policies and European pension systems of overlooking differences in work- life trajectories between men and women and of being designed based on a tra- ditional male breadwinner model of family life (Foster and Walker, 2014; Ni Leime and Street, 2019) . Although Swedish women’s and men’s working life courses are becoming more similar, historical and current differences still influence employment participation and retirement behavior. Women still have generally more interrupted work histories due to pregnancy and child care, more often occupy lower- status jobs, earn less, and have fewer opportunities to accumulate wealth in private/occupational pensions (Krekula et al., 2017) . Starting in the 1960 s, women’s employment participation saw a stable and steady increase, while the long - term negative trend for men was first reversed after the turn of the millennium (Johansson et al., 2018) . In monetary terms, how ever, women still obtain lower pensions than their male counterparts do. According to OECD estimates, the gender pension gap in Sweden is 28 % , which is close to the average pension gap for all 27 OECD countries of 25 % (OECD, 2020) . Besides cultural norms and gender roles, 10 65 (Hagen, 2013) . During the recession in the 1990s, disability pensions became an early retirement option for many people. However, stricter eligibility requirements were im- posed in 2008 that dramatically reduced the number of disability pensioners , i.e., people receiving sickness compensation. Research indicates that these reforms left some people suffering from impaired health with few options but to continue working or to take the economic consequences on themselves and leav e the labor market prematurely (Kadefors et al., 2018a) . The most critical policy reform was in 1999 when the current pension system, referred to as the Notional Defined Contribution System (NDC), was introduced (Kruse, 2010) . The NDC system increased the connection between the length of an individual’ s working life (i.e., contributions) and pension benefits. The NDC is financed through a combination of individual accounts and ongoing tax payments from the current workforce, in what is sometimes called a pay- as- you- go arrangement, and is based on intergenerational trans- fers. The current pension system consists of three parts : the state pension ( comprising the income pension, the premium pension scheme, and , in some instances, the minimum guaranteed pension, which is a means - tested pension for people with no or low prior labor- force attachment), the occupational pension , and private pension savings. 2 The legislated pensionable age of 65 years was abandoned. The new system provided some flexibility, as benefits from the income pension allowed workers to withdraw from work to retirement starting at age 61 and legislative rights in the Employment Protection Act allowed them to work until age 67. According to the 2019 pension agreement, however, these age thresholds for receiving old- age pension will gradually increase from 62 in 20 20 to 64 by 2026 and, after that, be linked to the average life expectancy advance in the Swedish pop- ulation (Ds, 2019:2) . Al though the pension system still allows for earlier or later with- drawal than at the age of 65, the 65 - year threshold still constitutes a strong normative reference point due to its historical origin (Johansson et al., 2018) . The 65 - year threshold is often used to distinguish whether an individual’ s retirement is “ early, ” “ on time, ” or “ late” in the public debate and in the research literature (Fisher et al., 2016) . 2 The state pension primar ily consists of an income pension and premium pension. Every year, 16% of the pensionable income is allocated to the income pension, while 2.5% of the pensionable income is as signed to the premium pension ( the total contribution rate is 18.5%). The premium pensions are allocated to stock market funds, but individuals can allocate their money to the fund(s) of their own choice. Most employers on the Swe- dish labor market provide an occupational pension to their employees. There are several occupational pension schemes, depending on the branch and sector, which are negotiated between the unions, employers , and em- ployers’ federations. Outside the pension system, some people also have private savings that can be used once they retire. This is of particular importance for the self- employed or for workers without an occupational pension provided by their employer, as private savings replace the occupational pension ( Kruse, A. (2010). A stable pension system: The eighth wonder. In: Bengtsson, T. (ed.) Population Ageing -A Threat to the Welfare State? Switzerland: Springer. ) . 11 Structural developments and challenges in the Swedish labor market Since the end of the 20th century, globalization and technological advance s have contrib- uted to significant changes in the Swedish labor market. These structural developments also interacted with and conditioned employment for men and women. Particularly nota- ble is the decline in male-dominated industrial production in favor of service -based indus- tries. The decline in industrial pr oduction was accompanied by stricter and more compre- hensive work environment regulations during the 1970s and 1980s, the establishment of several national institutes of occupational safety and health, and new legislation requiring occupational health services in workplaces. These developments reduced the number of physically strenuous work tasks and common occupational hazards such as chemical risks and noise pollution (Johansson et al., 2018) . However, improvement in the work environ- ment has not been equally distributed in all labor market sectors. During the recession of the 1990s, Swedish public and private employers cut back on jobs and carried out a re- structuring that led to higher job demands at many workplace s. Retrospective studies of the 1997– 2015 period indicate an overall stable trend in job demands and resources, alt- hough certain occupational groups deviated from this pattern (Corin et al., 2021) . More- over, two other longitudinal analyses show that , since the early 1990s, the Swedish labor market has witnessed increasing job dema nds, particularly in female -dominated occupa- tions such as health, social care , education (Aronsson et al., 2021; Cerdas et al., 2019) . In many of these female-dominated sectors , sick leave rates escalated quickly and have re- mained high until today (SSIA, 2015). According to 2019 data, women have a 25 % higher risk of sick leave and a 41 % higher risk of stress-related mental illness than do men across all occupational groups (SSIA, 2020) . The life -course trajectory that introduced the first section—growing up, acquiring an ed- ucation, getting a job, and ultimately retiring —can be accused of being oversimplified and reflecting a largely male perspective on working life. Feminists and other scholars accuse prolonged working life policies and European pension systems of overlooking differences in work-life trajectories between men and women and of being designed based on a tra- ditional male breadwinner model of family life (Foster and Walker, 2014; Ni Leime and Street, 2019) . Although Swedish women’s and men’s working life courses are becoming more similar, historical and current differences still influence employment participation and retirement behavior. Women still have generally more interrupted work histories due to pregnancy and child care, more often occupy lower-status jobs, earn less, and have fewer opportunities to accumulate wealth in private/occupational pensions (Krekula et al., 2017) . Starting in the 1960 s, women’s employment participation saw a stable and steady increase, while the long - term negative trend for men was first reversed after the turn of the millennium (Johansson et al., 2018) . In monetary terms, how ever, women still obtain lower pensions than their male counterparts do. According to OECD estimates, the gender pension gap in Sweden is 28 % , which is close to the average pension gap for all 27 OECD countries of 25% (OECD, 2020) . Besides cultural norms and gender roles, 12 some explanations are found in the highly gender-segregated Swedish labor market. Swe- dish men and women tend to work in different sectors and occupations ( i.e., horizontal segregation), and women are less likely to advance to higher positions in their careers ( i.e., vertical segregation). The gender division in the Swedish labor market influences earnings, physical and psychosocial exposures at work, the work environment, and opportunities to attain good health (Krekula et al., 2017) . So far, increasing demands for welfare services and labor shortages in specific sectors have been two of many consequences of population aging (Ahlberg et al., 2019; Brunello and Wruuck, 2019) . Th ese consequences have been manifested, for example, in Swedish wel- fare organizations that face difficulties maintaining skill and competence due to recruit- ment and retention problems. For instan ce, Swedish Association of Local Authorities and Regions (SAL AR) forecasts indicate that approximately 472 , 000 employees will be needed between 2019 and 2029 to compensate for forthcoming retirements and increase d service demands (SALAR, 2020) . SALAR has called on public and private employers to prioritize recruiting and retaining older healthcare workers while simultaneously becoming more attractive employers, implementing new technology and priorit izing sustainability in the work environment (SAL AR, 2018b; SALAR, 2020) . According to SALAR (2020) , welfare organizations’ abilities to recruit and sustain competence over upcoming decades will be decisive in preserving well-functioning welfare services such as preschool, school , public transportation, health care, and elderly care. This thesis focuses on Swedish welfare organ- izations and examines their organizational capability to manage and retain older employ- ees. In the next section, background information about the Sw edish public welfare organ- izations is outlined. The case of Swedish public welfare organizations Swedish welfare organizations are among the largest employers in the Swedish labor mar- ket, providing services in, for example, health and eldercare, education , sanitation, green spaces, and technical support . 3 About 80 % of workers in municipal and county councils are women (SALAR, 201 8b) , and nurses, nursing auxiliaries, and nursing assistants are all represented among the 30 largest occupations (Statistics Sweden, 2018). Swedish welfare organizations are predominantly public driven; however, the number of private welfare providers has increased substantially in recent decades. For example, since 2000, privately employed welfare workers have doubled to approximately 23 % of all welfare workers (SALAR, 2019) . 3 In Article s III and IV , public welfare organizations are referred to as “public -sector organizations.” However, it is worth noting that, although Swedish welfare organizations are predominantly public driven, private -owned companies also provide welfare services. 13 Public welfare organizations in Sweden, as in many other countries, became influenced by new public management (NPM ) in the 1980s, which led to the marketization of services, competition between public and private providers, instrumental control over processes, and a shift toward buyer– seller relationships (Bengtsson and Melke, 2014) . In the decades that followed, many welfare organizations experienced frequent restructurings and devel- opment work. According to Dellve and Eriksson (2016) , the recurring restructurings and development work often led to adverse outcomes in terms of employees’ motivation, sense of meaning, and health, although some organizations seemed to cope better with the changes than others. A common criti cism among debaters and scholars is that political decisions combined with new management models such as NPM have drained resources from welfare organizations, creating poorer work environments for employees and poorer service for their clients since the 1980s (Bengtsson and Melke, 2014) . Although much of the rationale underlying NPM originate d from private enterprises, several significant dif- ferences exist between organizations in the public and private sectors. Compared with private enterprises producing services or products based on free- market capitalism, public organizations, to a greater extent, must be transparent toward citizens and reflect demo- cratic values, and are heavily regulated by laws, policies, and guidelines. Swedish welfare organizations are governed at three levels: by parliament at the national level, by county councils at the regional level, and the county councils at the local level (Thylefors, 2016) . Several literature reviews and reports have pointed out that occupations with poor psy- chosocial working conditions and high physical demands are associated with ill health and earlier retirement, especially in contact jobs (see, for example, FORTE, 2015; SWEA, 2013) . Women often dominate health service occupations and, compared with men, more frequently experience higher job exposure s and occupational disorders, affecting not only sick- leave rates but also earlier retirement behavior (see, for example, Dellve et al., 2006) . As noted earlier, some studies have found a long- term increase in job demands in female - dominated jobs in the Swedish labor market. For inst ance, a recent longitudinal study covering the 1992 – 2013 period showed that this long- term trend was more apparent in care workers than in other occupations (Aronsson et al., 2021) . Another longitudinal study covering approximately the same period found a negative development of job demands and decision authority in women- dominated industries such as health, social care, and education (Cerdas et al., 2019) . The work environment and employment conditions in many welfare occupations are often up for public debate in Sweden. The Swedish Munic- ipal Workers’ Union (Kommunal) has argued that low status and poor working conditions explain the recruitment problems in these occupations. At the same time, employer s’ rep- resentatives highlight the general lack of potential personnel and high turnover as expla- nations for the recruitment difficulties (Odeberg, 2020) . Due to staffing needs, SALAR has initiated a national campaign to change the part- time work culture and transform part- time workers into full- time workers (SALAR, 2018a) . In line with these goals, the p ropor- 12 some explanations are found in the highly gender- segregated Swedish labor market. Swe- dish men and women tend to work in different sectors and occupations ( i.e., horizontal segregation), and women are less likely to advance to higher positions in their careers ( i.e., vertical segregation). The gender division in the Swedish labor market influences earnings, physical and psychosocial exposures at work, the work environment, and opportunities to attain good health (Krekula et al., 2017) . So far, increasing demands for welfare services and labor shortages in specific sectors have been two of many consequences of population aging (Ahlberg et al., 2019; Brunello and Wruuck, 2019) . Th ese consequences have been manifested, for example, in Swedish wel- fare organizations that face difficulties maintaining skill and competence due to recruit- ment and retention problems. For instan ce, Swedish Association of Local Authorities and Regions (SAL AR) forecasts indicate that approximately 472 , 000 employees will be needed between 2019 and 2029 to compensate for forthcoming retirements and increase d service demands (SALAR, 2020) . SALAR has called on public and private employers to prioritize recruiting and retaining older healthcare workers while simultaneously becoming more attractive employers, implementing new technology and priorit izing sustainability in the work environment (SAL AR, 2018b; SALAR, 2020) . According to SALAR (2020) , welfare organizations’ abilities to recruit and sustain competence over upcoming decades will be decisive in preserving well- functioning welfare services such as preschool, school , public transportation, health care, and elderly care. This thesis focuses on Swedish welfare organ- izations and examines their organizational capability to manage and retain older employ- ees. In the next section, background information about the Sw edish public welfare organ- izations is outlined. The case of Swedish public welfare organizations Swedish welfare organizations are among the largest employers in the Swedish labor mar- ket, providing services in, for example, health and eldercare, education , sanitation, green spaces, and technical support . 3 About 80 % of workers in municipal and county councils are women (SALAR, 201 8b) , and nurses, nursing auxiliaries, and nursing assistants are all represented among the 30 largest occupations (Statistics Sweden, 2018). Swedish welfare organizations are predominantly public driven; however, the number of private welfare providers has increased substantially in recent decades. For example, since 2000, privately employed welfare workers have doubled to approximately 23 % of all welfare workers (SALAR, 2019) . 3 In Article s III and IV , public welfare organizations are referred to as “public - sector organizations.” However, it is worth noting that, although Swedish welfare organizations are predominantly public driven, private - owned companies also provide welfare services. 13 Public welfare organizations in Sweden, as in many other countries, became influenced by new public management (NPM ) in the 1980s, which led to the marketization of services, competition between public and private providers, instrumental control over processes, and a shift toward buyer– seller relationships (Bengtsson and Melke, 2014) . In the decades that followed, many welfare organizations experienced frequent restructurings and devel- opment work. According to Dellve and Eriksson (2016) , the recurring restructurings and development work often led to adverse outcomes in terms of employees’ motivation, sense of meaning, and health, although some organizations seemed to cope better with the changes than others. A common criti cism among debaters and scholars is that political decisions combined with new management models such as NPM have drained resources from welfare organizations, creating poorer work environments for employees and poorer service for their clients since the 1980s (Bengtsson and Melke, 2014) . Although much of the rationale underlying NPM originate d from private enterprises, several significant dif- ferences exist between organizations in the public and private sectors. Compared with private enterprises producing services or products based on free-market capitalism, public organizations, to a greater extent, must be transparent toward citizens and reflect demo- cratic values, and are heavily regulated by laws, policies, and guidelines. Swedish welfare organizations are governed at three levels: by parliament at the national level, by county councils at the regional level, and the county councils at the local level (Thylefors, 2016) . Several literature reviews and reports have pointed out that occupations with poor psy- chosocial working conditions and high physical demands are associated with ill health and earlier retirement, especially in contact jobs (see, for example, FORTE, 2015; SWEA, 2013) . Women often dominate health service occupations and, compared with men, more frequently experience higher job exposure s and occupational disorders, affecting not only sick-leave rates but also earlier retire ment behavior (see, for example, Dellve et al., 2006) . As noted earlier, some studies have found a long-term increase in job demands in female - dominated jobs in the Swedish labor market. For inst ance, a recent longitudinal study covering the 1992 – 2013 period showed that this long- term trend was more apparent in care workers than in other occupations (Aronsson et al., 2021) . Another longitudinal study covering approximately the same period found a negative development of job demands and decision authority in women-dominated industries such as health , social care, and education (Cerdas et al., 2019) . The work environment and employment conditions in many welfare occupations are often up for public debate in Sweden. The Swedish Munic- ipal Workers’ Union (Kommunal) has argued that low status and poor working conditions explain the recruitment problems in these occupations. At the same time, employer s’ rep- resentatives highlight the general lack of potential personnel and high turnover as expla- nations for the recruitment difficulties (Odeberg, 2020) . Due to staffing needs, SALAR has initiated a national campaign to change the part-time work culture and transform part- time workers into full-time workers (SALAR, 2018a) . In line with these goals, the p ropor- 14 tion of full-time employees steadily increased between 2009 and 2019. However, accord- ing to 2019 data, a large share, approximately 30 % of all employees in Swedish municipal- ities, still work part time (SAL AR, 2020) . 15 3 Individual and organizational capability Leaving the labor market is a major life event for most people. Work offers occupational identity, colleagues, meaning, and earnings to sustain living expenses (Jahoda, 1981) . Jobs can simultaneously be both demanding and demeaning: working conditions can be poor, and the earnings inadequate to support a decent standard of living. Retirement might mean freedom and relief for some, but isolation and loss of meaning for others (Bengts- son and Flisbäck, 2017) . Every day, an average of approximately 317 Swedes take the critical step of leaving the labor market and retiring, most of them permanently. 4 The age at which individuals decide to retire is determined by the interaction of conditions at the personal, organizational, and societal le vels. To date, research into prolonged working life has spawned several concepts describing older workers’ ability to be active, productive, and age successfully at work (Zacher, 2015) . Also, numerous theoretical models have been developed to explain individuals’ retirement decision- making and behavior. These theo- retical models are interdisciplinary but originate from authors or research groups in spe- cific disciplines such as work psychology (Fisher et al., 2016; Wang and Shultz, 2010; Ybema et al., 2014) , occupational health (Dellve et al., 2003; Hasselhorn and Apt, 2015; Ilmarinen, 2006; Nilsson, 2020) , and sociology and work science (Dellve et al., 2021; Hofäcker and Radl, 2016) . Most theoretical conceptualizations o f retirement decision- making rest on an underlying exchange framework (Becker, 1965; Blau, 1964) in which individuals and organizations are assumed to assess the benefits and costs of maintaining the employment relationship. The e mployment relationship can be considered an exchange relationship in which em- ployees and employers negotiate and exchange resources to achieve their own or shared goals of value (Blau, 1964) . Resources consist of tangible or intangible goods, such as status, information, money, goods, and services (Foa and Foa, 1980) . Older workers can, for example, contribute experience and specific skills, and in return receive remuneration from the organization in terms of income, benefits, and access to social networks. How- ever, the notion that older people’ s retirement decisions are personal choices conceals the fact that people’s retirement decisions are formed and often limited by their opportunity structure, which depends on factors ranging from the resources they possess (e.g., health, education, and finances) to the contextual conditions under which they live (e.g., eligibility 4 The figure was calculated based on the average annual number of people, between 2018 and 2020, who ob- tained public pensions, divided by 365 days (i nformation retrieved from the Swedish Pension Agency , private correspondence, 2021- 03 - 25 ). 14 tion of full- time employees steadily increased between 2009 and 2019. However, accord- ing to 2019 data, a large share, approximately 30 % of all employees in Swedish municipal- ities, still work part time (SAL AR, 2020) . 15 3 Individual and organizational capability Leaving the labor market is a major life event for most people. Work offers occupational identity, colleagues, meaning, and earnings to sustain living expenses (Jahoda, 1981) . Jobs can simultaneously be both demanding and demeaning: working conditions can be poor, and the earnings inadequate to support a decent standard of living. Retirement might mean freedom and relief for some, but isolation and loss of meaning for others (Bengts- son and Flisbäck, 2017) . Every day, an average of approximately 317 Swedes take the critical step of leaving the labor market and retiring, most of them permanently. 4 The age at which individuals decide to retire is determined by the interaction of conditions at the personal, organizational, and societal le vels. To date, research into prolonged working life has spawned several concepts describing older workers’ ability to be active, productive, and age successfully at work (Zacher, 2015) . Also, numerous theoretical models have been developed to explain individuals’ retirement decision- making and behavior. These theo- retical models are interdisciplinary but originate from authors or research groups in spe- cific disciplines such as work psychology (Fisher et al., 2016; Wang and Shultz, 2010; Ybema et al., 2014) , occupational health (Dellve et al., 2003; Hasselhorn and Apt, 2015; Ilmarinen, 2006; Nilsson, 2020) , and sociology and work science (Dellve et al., 2021; Hofäcker and Radl, 2016) . Most theoretical conceptualizations o f retirement decision- making rest on an underlying exchange framework (Becker, 1965; Blau, 1964) in which individuals and organizations are assumed to assess the benefits and costs of maintaining the employment relationship. The e mployment relationship can be considered an exchange relationship in which em- ployees and employers negotiate and exchange resources to achieve their own or shared goals of value (Blau, 1964) . Resources consist of tangible or intangible goods, such as status, information, money, goods, and services (Foa and Foa, 1980) . Older workers can, for example, contribute experience and specific skills, and in return receive remuneration from the organization in terms of income, benefits, and access to social networks. How- ever, the notion that older people’ s retirement decisions are personal choices conceals the fact that people’s retirement decisions are formed and often limited by their opportunity structure, which depends on factors ranging from the resources they possess (e.g., health, education, and finances) to the contextual conditions under which they live (e.g., eligibility 4 The figure was calculated based on the average annual number of people, between 2018 and 2020, who ob- tained public pensions, divided by 365 days (i nformation retrieved from the Swedish Pension Agency , private correspondence, 2021-03-25). 16 age in the pension system, norms, and culture). For instance, Radl (2013) argued that older workers’ opportunity structure is primarily determined by their level of employability and pension entitlements. T o sum up, the objective of section 3 is to analytically examine the micro and meso levels and to illustrate how contextual circumstances at the macro level interact in an individual’ s retirement decision-making. Specifically, attention is directed toward the interaction be- tween personal and work-related conditions , on one hand, and the organization’ s HR practices that influence older workers’ retirement behavior, on the other . To do this, I use the capability approach as a theoretical framework. I t offers an analytical tool with which to understand and elucidate how surrounding circumstances condition an individual’s or or- ganization’ s capability to realize selected goals based on their ability to convert resources into capabilities. The following section s describe the capability concept and how it can be applied as a theoretical tool with which to understand conditions that influence older workers’ capability to continue working or organizations’ capabilities to implement age- management strategies. C apability in the context of work C apability refers to an individual’ s ability to reach goals and objectives that s/he has reason to value, such as working, being healthy, or maintainin g social relations (Sen, 2005; Sen, 2009) . An individual ’ s capability is determined by her/his access to resources and ability to convert these resources into attaining personal goals. Together, these condition s con- stitute the individual’ s capability set, i.e., ability to convert existing resources and perform actions to realize selected goals. In th is thesis, I am primarily interested in how people’ s capabilities are translated into work functions and how their capabilities influence older workers’ decisions to remain employed or leave the labor market. Later, I will also use the capability approach in an organizational setting and apply Kadefors et al.’ s (2020) notion of organizational capability, which refers to organizations ’ capability to implement age- management practices based on their existing resources and preferences. The individual and organizational capability approach es share the same underlying theory (Sen, 2005; Sen, 2009) , based on individuals’ and organizations’ access to various forms of resources and specific preferences to achieve selected goals. Resources exist at different societal levels, in people ’ s private lives in the form of, for example, health, finances, and social position ( the micro level), in their circumstances in their families and workplaces ( the meso level) , and in their contextual conditions such as social security systems and regulations in the societies where they live ( the macro level) (Dellve et al., 2021). T he next section presents an overview of retirement determinants at the societal, workplace, and individual levels. The configurations of th e different levels, i.e., the micro, meso, and macro levels, frame individuals ’ opportunities to realize chosen goals. However, although 17 two people may have access to the same resources, their ability to convert the se resources into valuable goals may differ. Low capability can be understood as a lack of congruence between the different components of the capability model, i.e., an imbalance or mismatch between resources, conversion factors, choices, and goals. For instance, if an individual ’ s goal is to leave the labor market at age 63, but eligibility thresholds in the pension system prohibit this or financial resources are scarce, their capability can be considered low. Given the objective of this thesis, it is essential to identify enabling and hindering circum- stances for older workers’ capability to continue working. Resources of particular interest are those that determine people’ s ability to work, for example, health, education, and workplace conditions (e.g., flexibility or opportunities for competence development) . People ’ s abilities to convert resources are referred to as conversion factors , which can some- times overlap with resources. For instance, collective agreements can be both a resource (e.g., understanding workplace rights and agreements) and a conversion factor (e.g., the ability to request certain working conditions, work adjustments , or competence develop- ment at the workplace). To illustrate , an older work er experiencing reduced work ability due to hip surgery may need alternative job tasks or ergonomic equipment during the healing process. T hus, in this example, ergonomic equipment is both a r esource and a conversion factor that strengthens individuals’ capabilities. T he capability approach has been questioned and criticized from different perspectives. A common criti cism is that the approach is too individualistic and thus downplays the importance of structures and institutions as well as groups and communities. Capability and other agency- based theories can be contrasted to more Marxist - oriented perspectives, which emphasize the imbalance of exchange in terms of resources and power between the employees (workers) and the employer (capitalist) (Blauner, 1964; Braverman, 1998; Edwards, 1 982) . Another related criticism of the capability approach is that it does not offer any tool with which to examine whether an individual’ s living and working condi- tions are objectively fair or equal . In stead, Sen (2005) and Nussbaum (1997) argued that external reference points are needed, such as universally agreed human rights, which they commonly employ in their research on human development. Against this limitation of the capability concept, I use the notion of “ sustainable working life” as a reference point. Government agencies (Regeringskansliet, 2019; SALAR, 2020; SWEA, 2016) and re- searchers (e.g., Dellve and Eriksson, 2016; Docherty et al., 2008; Kira, 2003; Nilsson, 2020; Van der Klink et al., 2016) often invoke sustainability to depict situations in which individuals’ work is stimulating and enriching and does not endanger their health or well- being. Sustainability is also a key term used in the United Nations ’ Sustainable Develop- ment Goals, the eighth of which concerns sustainability at work to “ promote sustained, inclusive and sustainable economic growth, full and productive employment and decent work for all” (United Nations , 2021) . 16 age in the pension system, norms, and culture). For instance, Radl (2013) argued that older workers’ opportunity structure is primarily determined by their level of employability and pension entitlements. T o sum up, the objective of section 3 is to analytically examine the micro and meso levels and to illustrate how contextual circumstances at the macro level interact in an individual’ s retirement decision- making. Specifically, attention is directed toward the interaction be- tween personal and work- related conditions, on one hand, and the organization’ s HR practices that influence older workers’ retirement behavior, on the other . To do this, I use the capability approach as a theoretical framework. I t offers an analytical tool with which to understand and elucidate how surrounding circumstances condition an individual’ s or or- ganization’ s capability to realize selected goals based on their ability to convert resources into capabilities. The following section s describe the capability concept and how it can be applied as a theoretical tool with which to understand conditions that influence older workers’ capability to continue working or organizations’ capabilities to implement age- management strategies. C apability in the context of work C apability refers to an individual’ s ability to reach goals and objectives that s/he has reason to value, such as working, being healthy, or maintainin g social relations (Sen, 2005; Sen, 2009) . An individual ’ s capability is determined by her/his access to resources and ability to convert these resources into attaining personal goals. Together, these condition s con- stitute the individual’ s capability set, i.e., ability to convert existing resources and perform actions to realize selected goals. In th is thesis, I am primarily interested in how people’ s capabilities are translated into work functions and how their capabilities influence older workers’ decisions to remain employed or leave the labor market. Later, I will also use the capability approach in an organizational setting and apply Kadefors et al. ’ s (2020) notion of organizational capability, which refers to organizations ’ capability to implement age- management practices based on their existing resources and preferences. The individual and organizational capability approach es share the same underlying theory (Sen, 2005; Sen, 2009) , based on individuals’ and organizations’ access to various forms of resources and specific preferences to achieve selected goals. Resources exist at different societal levels, in people ’ s private lives in the form of, for example, health, finances, and social position ( the micro level), in their circumstances in their families and workplaces ( the meso level) , and in their contextual conditions such as social security systems and regulations in the societies where they live ( the macro level) (Dellve et al., 2021) . T he next section presents an overview of retirement determinants at the societal, workplace, and individual levels. The configurations of th e different levels, i.e., the micro, meso, and macro levels, frame individuals ’ opportunities to realize chosen goals. However, although 17 two people may have access to the same resources, their ability to convert the se resources into valuable goals may differ. Low capability can be understood as a lack of congruence between the different components of the capability model, i.e., an imbalance or mismatch between resources, conversion factors, choices, and goals. For instance, if an individual ’ s goal is to leave the labor market at age 63, but eligibility thresholds in the pension system prohibit this or financial resources are scarce, their capability can be considered low. Given the objective of this thesis, it is essential to identify enabling and hindering circum- stances for older workers’ capability to continue working. Resources of particular interest are those that determine people’ s ability to work, for example, health, education, and workplace conditions (e.g., flexibility or opportunities for competence development) . People ’ s abilities to convert resources are referred to as conversion factors , which can some- times overlap with resources. For instance, collective agreements can be both a resource (e.g., understanding workplace rights and agreements) and a conversion factor (e.g., the ability to request certain working conditions, work adjustments , or competence develop- ment at the workplace). To illustrate , an older work er experiencing reduced work ability due to hip surgery may need alternative job tasks or ergonomic equipment during the healing process. T hus, in this example, ergonomic equipment is both a r esource and a conversion factor that strengthens individuals’ capabilities. T he capability approach has been questioned and criticized from different perspectives. A common criti cism is that the approach is too individualistic and thus downplays the importance of structures and institutions as well as groups and communities. Capability and other agency- based theories can be contrasted to more Marxist -oriented perspectives, which emphasize the imbalance of exchange in terms of resources and power between the employees (workers) and the employer (capitalist) (Blauner, 1964; Braverman, 1998; Edwards, 1 982) . Another related criticism of the capability approach is that it does not offer any tool with which to examine whether an individual’ s living and working condi- tions are objectively fair or equal . In stead, Sen (2005) and Nussbaum (1997) argued that external reference points are needed, such as universally agreed human rights, which they commonly employ in their research on human development. Against this limitation of the capability concept, I use the notion of “ sustainable working life” as a reference point. Government agencies (Regeringskansliet, 2019; SALAR, 2020; SWEA, 2016) and re- searchers (e.g., Dellve and Eriksson, 2016; Docherty et al., 2008; Kira, 2003; Nilsson, 2020; Van der Klink et al., 2016) often invoke sustainability to depict situations in which individuals’ work is stimulating and enriching and does not endanger their health or well- being. Sustainability is also a key term used in the United Nations ’ Sustainable Develop- ment Goals, the eighth of which concerns sustainability at work to “ promote sustained, inclusive and sustainable economic growth, full and productive employment and decent work for all” (United Nations , 2021) . 18 Individual capability In past literature on retirement decision-making , a distinction is often made between an individual’ s ability and willingness to continue working or to leave the labor market. Here, abilities refer to individuals’ physical and psychological functions as well as to their com- petence and skills. In contrast , individuals ’ willingness is assumed to be determined by, for example, circumstances at their workplace, (dis)incentives in the pension system, fi- nancial needs, and social relations. Th is categorization based on the different combina- tions of abilities and willingness offers a tool with which to understand what influences employment participation among older workers and to identify “ stronger” and “ weaker” groups. In Figure 1 , suggested typologies are visualized in terms of individuals’ abilities and willingness. I t is essential to recall that these typologies in fact simplify a complex reality and should not be considered constant. Instead, an individual’ s willingness and ability should be regarded as dynamic, affected by age-related changes in physical and psychological functions as well as by surrounding circumstances. Figure 1. Typology of i ndividual willingness and abilities to continue working. To illustrate the different combinations of abilities and willingness, I present four fictive people: Helena , Olov , Johnny , and Katrin . Helena is a 67 -year -old assistan t nurse employed at a public elderly care service. She is satisfied with her job and co-workers and is often a mentor for younger employees. Because her employer terminates employment contracts after age 67, Helena has few alternatives to retiring at age 67 ( Willing/Unable in Figure 1). Olov is a skilled 62 -year -old welder working at a sheet metal processing company where there is little room for autonomy or work flexibility. Olov is fully functional in terms of 19 abilities but looks forward to retirement to spend more time with his grandchildren and hobbies. Olov ’ s work ability is high and, despite continuous employment being somewhat involuntary, his financial need to continue working until the age of 64 is still considered a motivator to remain at work ( Willing/Able in Figure 1). Johnny is a 60 - year- old industrial worker who believes that he has “ done his duty” on the labor market after a long working life. After a lengthy rehabilitation process after hip surgery, he returned to his job but immediately experienced problems carrying out his ordinary work tasks. John ny’ s abilities and willingness to continue working can be considered low ( Unwilling /Unable in Figure 1). Katrin is a 65 - year- old manager in elderly care. After a major restructuring at Katrin ’ s workplace, including rationalizing personnel and new work proc edures, administrative work consumed much of the time spent with clients, leading to stress and feelings of meaninglessness. In this example, Katrin ’ s willingness to continue working is low, yet she can still carry out her job, despite the high workload ( U nwilling/Able in Figure 1). In the “ Discussion” section, I will return to the cases of Helena , Olov , Johnny , and Katrin . Organizational capability The age - management imperative advocates organizations’ active involvement, through HR practices, in meeting older workers’ needs and enhancing their work capability. How- ever, Kade fors et al. (2020) argued that organizations’ ability to implement such manage- ment practices depends not only on their willingness and preferences but also on their resources. Consequently, Kadefors et al. (2020) developed a typology of organizational capability and identified four ideal types of organizations’ capability to implement age- management practices based on their preferences and available resources: proactive , reactive , chained , and passive. Resources may refer to tangible and intangible goods, such as in - house HRM competence, the number of employees per line manager, and financial capacity to offer ergonomic equipment. On the other hand, preferences can be determined by an organization’ s priorities, recruitment needs, and attitudes. Figure 2 visualizes the typology; for example, a proactive organizati on has the willingness and resources to implement age management, while a chained organization lacks the resources to do so. 18 Individual capability In past literature on retirement decision- making, a distinction is often made between an individual’ s ability and willingness to continue working or to leave the labor market. Here, abilities refer to individuals’ physical and psychological functions as well as to their com- petence and skills. In contrast , individuals ’ willingness is assumed to be determined by, for example, circumstances at their workplace, (dis)incentives in the pension system, fi- nancial needs, and social relations. Th is categorization based on the different combina- tions of abilities and willingness offers a tool with which to understand what influences employment participation among older workers and to identify “ stronger” and “ weaker” groups. In Figure 1 , suggested typologies are visualized in terms of individuals’ abilities and willingness. I t is essential to recall that these typologies in fact simplify a complex reality and should not be considered constant. Instead, an individual’ s willingness and ability should be regarded as dynamic, affected by age- related changes in physical and psychological functions as well as by surrounding circumstances. Figure 1. Typology of i ndividual willingness and abilities to continue working. To illustrate the different combinations of abilities and willingness, I present four fictive people: Helena , Olov , Johnny , and Katrin . Helena is a 67 - year- old assistant nurse employed at a public elderly care service. She is satisfied with her job and co- workers and is often a mentor for younger employees. Because her employer terminates employment contracts after age 67, Helena has few alternatives to retiring at age 67 ( Willing/Unable in Figure 1). Olov is a skilled 62 - year- old welder working at a sheet metal processing company where there is little room for autonomy or work flexibility. Olov is fully functional in terms of 19 abilities but looks forward to retirement to spend more time with his grandchildren and hobbies. Olov ’ s work ability is high and, despite continuous employment being somewhat involuntary, his financial need to continue working until the age of 64 is still considered a motivator to remain at work ( Willing/Able in Figure 1). Johnny is a 60-year -old industrial worker who believes that he has “ done his duty” on the labor market after a long working life. After a lengthy rehabilitation process after hip surgery, he returned to his job but immediately experienced problems carrying out his ordinary work tasks. John ny’ s abilities and willingness to continue working can be considered low ( Unwilling /Unable in Figure 1). Katrin is a 65 -year -old manager in elderly care. After a major restructuring at Katrin ’ s workplace, including rationalizing personnel and new work proc edures, administrative work consumed much of the time spent with clients, leading to stress and feelings of meaninglessness. In this example, Katrin ’ s willingness to continue working is low, yet she can still carry out her job, despite the high workload ( U nwilling/Able in Figure 1). In the “ Discussion” section, I will return to the cases of Helena , Olov , Johnny , and Katrin . Organizational capability The age -management imperative advocates organizations ’ active involvement, through HR practices, in meeting older workers’ needs and enhancing their work capability. How- ever, Kade fors et al. (2020) argued that organizations’ ability to implement such manage- ment practices depends not only on their willingness and preferences but also on their resources. Consequently, Kadefors et al. (2020) developed a typology of organizational capability and identified four ideal types of organizations’ capability to implement age- management practices based on their preferences and available resources: proactive , reactive , chained , and passive. Resources may refer to tangible and intangible goods, such as in - house HRM competence, the number of employees per line manager, and financial capacity to offer ergonomic equipment. On the other hand, preferences can be determined by an organization’ s priorities, recruitment needs, and attitudes. Figure 2 visualizes the typology; for example, a proactive organizati on has the willingness and resources to implement age management, while a chained organization lacks the resources to do so. 20 Figure 2. Typology of o rganizational capability to implement age management based on the re- sources and preferences identified by Kadefors et al. (2020 ) . Because organizations’ capabilities to retain older workers are based on their preferences and access to resources (Kadefors et al., 2020) , it is reasonable to assume that the prereq- uisites for organizational capability vary between sectors and branches in the labor market. Contextual conditions such as business -cycle fluctuations, labor supply, evolving mana ge- ment ideas, technological advance s, and policy changes over time affect organizations ’ and older workers’ motivations, abilities , and need to continue the employment relation- ship. For example, in Swedish public welfare organizations, recruitment needs ma y trigger preferences for retaining older workers. Organizations ’ preferences must be understood in relation to some of the basic assumptions of organizational theory, which stipulates that organizations are goal-oriented, focus on maximizing productivity, strive for conti- nuity, and have a sound market position (Kalleberg et al., 1996) . Therefore, organizations ’ willingness to employ, manage , and retain older workers depends on whether they con- tribute to these goals. However, other factors such as laws, regulations, and attitudes to- ward older workers also influence organizations’ decisions. For instance, Conen et al. (2011) presented two reasons why organizations retain and recruit older workers: the workers have experience, specific skills, and can help ensure knowledge transfer; the em- ployers are concerned with avoiding labor shortages over time. Moreover, Malmqvist (2007) evaluated the potential costs and benefits for organizations of applying a defensive or offensive approach to implementing age management; additionally, he considered prac- tical experiences from four case studies in Swedish organizations. Malmqvist (2007) con- cluded that positive outcomes were more likely in organizations that more proactively implemented age-management practices. Furthermore, Malmqvist (2007) identified sev- 21 eral potential upsides of active age management, such as lower sickness absence and re- habilitation costs, exten ded employment, reduced turnover, higher productivity due to competence development and motivation, and ultimately external goodwill from the mar- ket. A model of individual– organizational capability By combining individuals’ work capabilities and organizations’ capability to implement age- management practices, we can see the preconditions for prolonged working life in Figure 3 . The diagonal arrow denote s the preconditions for a prolonged working life. This f igure makes it possible to distinguish and identify stronger and weaker positions or groups in the labor market. For instance, it is reasonable that the “ worst” situation is when an older individual is unwilling and unable to continue working (i.e., Incapable) while simultane- ously being employed in a passive organization. Conversely, the same individual would probably have better employment opportunities in a proactive organization. Another rea- sonable assumption is that a capable older worker is less likely to be affected in a passive organization. I t is important to emphasize that older workers’ capability set, in theory, simultaneously includes the existing conditions and resources available at the organizational level. However, individuals ’ and organizations’ capabilities are dynamic and can change depending on the time and place. For instance, an individual may experience a change in physical and psychological functions over time, while a committee decision to prioritize the retention of older workers may transform a passive organization into a proactive one, provided it has sufficient resources. In the “ Discussion” section, I will return to the indi- vidual– organizational model. Figure 3. Preconditions for a prolonged working life: organizational capability to implement age-management practices, combined with a typology of individual abilities and willingness. 20 Figure 2. Typology of o rganizational capability to implement age management based on the re- sources and preferences identified by Kadefors et al. (2020 ) . Because organizations’ capabilities to retain older workers are based on their preferences and access to resources (Kadefors et al., 2020) , it is reasonable to assume that the prereq- uisites for organizational capability vary between sectors and branches in the labor market. Contextual conditions such as business - cycle fluctuations, labor supply, evolving mana ge- ment ideas, technological advance s, and policy changes over time affect organizations ’ and older workers’ motivations, abilit ies, and need to continue the employment relation- ship. For example, in Swedish public welfare organizations, recruitment needs ma y trigger preferences for retaining older workers. Organizations ’ preferences must be understood in relation to some of the basic assumptions of organizational theory, which stipulates that organizations are goal- oriented, focus on maximizing productivity, strive for conti- nuity, and have a sound market position (Kalleberg et al., 1996) . Therefore, organizations ’ willingness to employ, manage , and retain older workers depends on whether they con- tribute to these goals. However, other factors such as laws, regulations, and attitudes to- ward older workers also influence organizations’ decisions. For instance, Conen et al. (2011) presented two reasons why organizations retain and recruit older workers: the workers have experience, specific skills, and can help ensure knowledge transfer; the em- ployers are concerned with avoiding labor shortages over time. Moreover, Malmqvist (2007) evaluated the potential costs and benefits for organizations of applying a defensive or offensive approach to implementing age management; additionally, he considered prac- tical experiences from four case studies in Swedish organizations. Malmqvist (2007) con- cluded that positive outcomes were more likely in organizations that more proactively implemented age- management practices. Furthermore, Malmqvist (2007) identified sev- 21 eral potential upsides of active age management, such as lower sickness absence and re- habilitation costs, exten ded employment, reduced turnover, higher productivity due to competence development and motivation, and ultimately external goodwill from the mar- ket. A model of individual– organizational capability By combining individuals’ work capabilities and organizations’ capability to implement age-management practices , we can see the preconditions for prolonged working life in Figure 3 . The diagonal arrow denote s the preconditions for a prolonged working life. This f igure makes it possible to distinguish and identify stronger and weaker positions or groups in the labor market. For instance, it is reasonable that the “ worst” situation is when an older individual is unwilling and unable to continue working (i.e., Incapable) while simultane- ously being employed in a passive organization. Conversely, the same individual would probably have better employment opportunities in a proactive organization. Another rea- sonable assumption is that a capable older worker is less likely to be affected in a passive organization. I t is important to emphasize that older workers’ capability set, in theory, simultaneously includes the existing conditions and resources available at the organizational level. However, individuals ’ and organizations’ capabilities are dynamic and can change depending on the time and place. For instance, an individual may experience a change in physical and psychological functions over time, while a committee decision to prioritize the retention of older workers may transform a passive organization into a proactive one, provided it has sufficient resources. In the “ Discussion” section, I will return to the indi- vidual– organizational model. Figure 3. Preconditions for a prolonged working life: organizational capability to implement age-management practices, combined with a typology of individual abilities and willingness. 22 4 Understanding the employment participation of older workers The effort to prolong the working life in many countries has prompted much research to identify conditions that hinder older workers from continuing to work, or enable them to do so. The current knowledge base regarding retirement determinants is relatively good, although research is fragmented among disciplines, domains, and contexts and depending on how retirement is defined and measured. A person’ s pathway from work to retirement is often determined by multiple factors at the societal, workplace , and individual levels. Also, interpersonal heterogeneity, interactions at different levels, and changes over time add to the complexity. In past research, the push – pull model was often applied to distin- guish factors that “ encourage” workers to stay in or leave a specific workplace or labor market position, or “discourage” them f rom doing so. However, in retirement research, the push– pull model refers to conditions influencing older workers’ earlier exit from the labor market. In this view, people either get pushed (involuntar ily) or pulled (voluntarily) out of the workforce. To e xemplify, occupational hazards, personal health problems, lack of recognition and social support in the workplace, and ageism in society are potential push factors. On the other hand, occupational pension schemes that incentivize early retire- ment, having a retired spouse, and having compelling leisure interests are examples of pull factors. In light of PWL policies, Hofäcker and Radl (2016) identified a gap in the push– pull model and added two dimensions that influence older workers’ retirement decisions, i.e., need and maintenance factors. Need factors refer to institutional and individual conditions that force older workers to continue working to meet living expenses. For instance, policy adjustments to raise the retiremen t age, reduce the replacement rate in the pension system, or tighten the eligibility criteria for the disability pension require some older workers to continue working. Maint enance factors refer to institutional conditions such as active labor market policies and lifelong learning intended to strengthen the employability of older workers. Maint enance factors also include factors at the organizational level, such as age - management strategies by which employers can accommodate, develop , or modify work according to older workers’ needs. Previous empirical research on working and retirement decisions is presented in terms of the capability model’ s different levels, i.e., the societal, workplace, and individual level s. Here, I apply the push – pull, need – maintain terminology mentioned above to elucidate factors influencing older workers to remain in employment or retire (Hofäcker and Ra dl, 23 2016) . The following sections are intended to offer an overview of factors influencing people’ s retirement intentions or behavior, not to provide a comprehensive review (For previous literature reviews, see, for example, Browne et al., 2019; Fisher et al., 2016; Phillipson and Smith, 2005; Scharn et al., 2018; SWEA, 2012b; Wang and Shultz, 2010) . After first discussing societal conditions, I then focus on workplace conditions and end with individual age- related changes that influence retirement behavior. Societal- level determinants Th e configurations of contextual conditions at the societal level, for example , the pension system, work environment laws , collective agreements, business - cycle fluctuations, devel- opments in the labor market, societal norms, and values, constitute the framework for individuals’ capabilities. For older as well as younger people, the need to uphold a certain standard of living is a main factor behind remaining active in the labor market. As previ- ously discussed, several policy changes have been imple mented to encourage prolonged working lives. For example, t he current pension system was fully implemented in 2003 and increased the connection between working life length and pension benefits. The pre- dicted replacement rate from the pension system, i.e., the difference between the final salary and the first pension payout, is another factor affecting people’ s retirement deci- sion- making. A Swedish governmental report estimates that the replacement rate will de- cline over time. For instance, people born in 19 30 have an expected replacement rate of 60 % ; among those born in 1945, the replacement rate is 50 % (Ds, 2019:2) , and among those born in 1992, the expected replacement rate will be below 40 % . These repl acement rates will be essential factors for many workers, as they will need to stay active longer to uphold a certain standard of living. In addition, t he employment tax deduction (Swedish: j obbskatteavdraget ) implemented stepwise between 2008 and 2019 has made employment participation more profitable. A consequ ence of implementing the employment tax de- duction is that, for example, pensioners pay more tax in relative terms, and critics are referring to it as an additional tax on retirees ( Swedish: pensionärsskatt ) . The employment tax deduction maintains older workers in the workforce for financial reasons. T he Employment Protection Act (LAS, 1982:80) offers older workers statutory rights to continue working until age 6 7 , a threshold that will be gradually increased to 69 by 2023. (Swedish Parliament, 2019) . The consequences of increasing th is age in the Act have pre- viously been evaluated in a Swedish study. Kadefors et al. (2014) followed employees in two large public- sector organizations and examined what happened after the employment protection threshold was raised from 65 to 67 in 2001. The ir findings indicated an overall extension of employment participation but also revealed gender and socioeconomic dif- ferences. Males with higher education capitalized on the opportunities to remain em- 22 4 Understanding the employment participation of older workers The effort to prolong the working life in many countries has prompted much research to identify conditions that hinder older workers from continuing to work, or enable them to do so. The current knowledge base regarding retirement determinants is relatively good, although research is fragmented among disciplines, domains, and contexts and depending on how retirement is defined and measured. A person’ s pathway from work to retirement is often determined by multiple factors at the societal, workplace , and individual levels. Also, interpersonal heterogeneity, interactions at different levels, and changes over time add to the complexity. In past research, the push – pull model was often applied to distin- guish factors that “ encourage” workers to stay in or leave a specific workplace or labor market position, or “discourage” them f rom doing so. However, in retirement research, the push– pull model refers to conditions influencing older workers’ earlier exit from the labor market. In this view, people either get pushed (involuntar ily) or pulled (voluntarily) out of the workforce. To e xemplify, occupational hazards, personal health problems, lack of recognition and social support in the workplace, and ageism in society are potential push factors. On the other hand, occupational pension schemes that incentivize early retire- ment, having a retired spouse, and having compelling leisure interests are examples of pull factors. In light of PWL policies, Hofäcker and Radl (2016) identified a gap in the push– pull model and added two dimensions that influence older workers’ retirement decisions, i.e., need and maintenance factors. Need factors refer to institutional and individual conditions that force older workers to continue working to meet living expenses. For instance, policy adjustments to raise the retiremen t age, reduce the replacement rate in the pension system, or tighten the eligibility criteria for the disability pension require some older workers to continue working. Maint enance factors refer to institutional conditions such as active labor market policies and lifelong learning intended to strengthen the employability of older workers. Maint enance factors also include factors at the organizational level, such as age - management strategies by which employers can accommodate, develop , or modify work according to older workers’ needs. Previous empirical research on working and retirement decisions is presented in terms of the capability model’ s different levels, i.e., the societal, workplace, and individual level s. Here, I apply the push – pull, need – maintain terminology mentioned above to elucidate factors influencing older workers to remain in employment or retire (Hofäcker and Ra dl, 23 2016) . The following sections are intended to offer an overview of factors influencing people’ s retirement intentions or behavior, not to provide a comprehensive review (For previous literature reviews, see, for example, Browne et al., 2019; Fisher et al., 2016; Phillipson and Smith, 2005; Scharn et al., 2018; SWEA, 2012b; Wang and Shultz, 2010) . After first discussing societal conditions, I then focus on workplace conditions and end with individual age-related changes that influence retirement behavior. Societal- level determinants Th e configurations of contextual conditions at the societal level, for example , the pension system, work environment laws , collective agreements, business -cycle fluctuations , devel- opments in the labor market, societal norms, and values, constitute the framework for individuals’ capabilities. For older as well as younger people, the need to uphold a certain standard of living is a main factor behind remaining active in the labor market. As previ- ously discussed, several policy changes have been imple mented to encourage prolonged working lives. For example, t he current pension system was fully implemented in 2003 and increased the connection between working life length and pension benefits. The pre- dicted replacement rate from the pension system, i.e., the difference between the final salary and the first pension payout, is another factor affecting people’ s retirement deci- sion-making. A Swedish governmental report estimates that the replacement rate will de- cline over time. For instance, people born in 19 30 have an expected replacement rate of 60% ; among those born in 1945, the replacement rate is 50 % (Ds, 2019:2) , and among those born in 1992, the expected replacement rate will be below 40 % . These repl acement rates will be essential factors for many workers, as they will need to stay active longer to uphold a certain standard of living. In addition, t he employment tax deduction (Swedish: j obbskatteavdraget ) implemented stepwise between 2008 and 2019 has made employment participation more profitable. A consequ ence of implementing the employment tax de- duction is that, for example, pensioners pay more tax in relative terms, and critics are referring to it as an additional tax on retirees ( Swedish: pensionärsskatt ) . The employment tax deduction maintains older workers in the workforce for financial reasons. T he Employment Protection Act (LAS, 1982:80) offers older workers statutory rights to continue working until age 6 7 , a threshold that will be gradually increased to 69 by 2023. (Swedish Parliament, 2019) . The consequences of increasing th is age in the Act have pre- viously been evaluated in a Swedish study. Kadefors et al. (2014) followed employees in two large public-sector organizations and examined what happened after the employment protection threshold was raised from 65 to 67 in 2001. The ir findings indicated an overall extension of employment participation but also revealed gender and socioeconomic dif- ferences. Males with higher education capitalized on the opportunities to remain em- 24 ployed, while members of the female-dominated occupations assistant nurses, adminis- trators, and nurses were less likely to continue working beyond age 65. Th e increased age threshold in the Act can be considered both a push and a maintenance factor depending on the demographic conditions and whether or not an individual intends to continue working beyond the age threshold specified in the Act. The Swedish Work Environment Act requires organizations to conduct health and safety work at the workplace, in what is better known as systematic work environment manage- ment ( SAM). Most Swedish public and private organizations have either an internal or external occupational health service (OHS) provider. The role of the OHS is to support employers and their managerial representatives in occupational health and safety manage- ment (OHS M) to prevent hazardous work environments from forming, safeguard em- ployees’ health and well- being, and provide rehabilitation schemes for those in need. In addition, well- functioning implementation of the Work Environment Act maintains and strengthens older workers’ capability to continue working. However, according to data from the Swedish Work Environment Authority , the presence of OHS providers in Swe- dish workplaces has gradually decreased since the year 2000 (SWEA, 2012a; SWEA, 2020a) , and on ly half of participants stated that systematic work on SAM is pursued at their workplaces (SWEA, 2016:2). Ageism and age discrimination are caused by how people and employers think, feel, and act regarding older workers and their competence, knowledge, and abilities (Marques et al., 2020) . It is well established in previous research that ageism and negative attitudes among employers are a push factor that restricts older people from, for example, recruit- ment opportunities in the labor market (Ahmed et al., 2012; Conen et al., 2012; Furunes and Mykletun, 2007; Harris et al., 2017; Kadefors and Hanse, 2012) . Furthermore, norms and values still play an important role in retirement decisions. For instance, even though individuals’ health and working environments have improved in recent decades, the 65 - year threshold still constitutes a strong normative reference point and retirement age for many people (Johansson et al., 2015) . These normative and cultural notions are examples of push factors encouraging older individuals to leave the labor market. Also, state legis- lation can contribute to unequal opportunities and age discrimination, pushing people out of employment participation. For instance, the age thresholds in the Employment Pro- tection Act are usually the benchmarks for employment agreements stipulating the age at which the employment relationship ends. Even though the configurations of social secu- rity and pension systems are essential to older workers’ retirement behavior, personal and organizational factors also interact in these relationships, as we will see in the next two sections. 25 Workplace- level determinants The sustainable extension of working life requires that individuals maintain their work ability and motivation in order to continue working. Likewise, organizations must recog- nize older workers as a resource and be able to manage their heterogeneous needs and desires. Today, it is well established that the design or content of work is strongly related to employees’ health and well- being as well as to their motivation, engagement, and productivity. In line with this, the person – job fit literature (Edwards, 1991; Edwards et al., 1998) suggests that early or late retirement behavior can be understood as a result of the (mis)fit between older workers ’ personal characteristics (e.g., abilit ies, preferences, personality, health, and financial resources) and the characteristics of their job s, co - work- ers, organization s, and managers. To exemplify, an older worker might experience reduced work ability that requires some form of work adaptation. However, the organiz ation may not provide such work adaptation, so the individual experiences a misfit. Such misfits are likely to increase older workers’ inclination to take early retirement. Li nking back to the concept of capability, the aforementioned example illustrates that employers play a key role in its materialization. This is not least because of employers’ potential efforts to con- vert “ misfits” into “ fits” through, for example, work adaptation, which is an essential conversion factor for the individual worker and thus a means to strengthen his/her capa- bility (Wang and Shultz, 2010) . Preju dices toward older workers Emp loyers’ willingness and ability to retain and recruit older workers are fundamental preconditions for older workers’ opportunities to prolong their working lives (Conen et al., 2012; Conen et al., 2014; Vick erstaff et al., 2003) . For instance, a longitudinal study in the Netherlands found that having a supervisor who encourage d employees to continue working helped keep older workers longer in the workforce (Van Solinge and Henkens, 2014) . Also, the implementation of age- management activities is reliant on line managers’ knowledge, beliefs, and attitudes toward older workers (Furunes et al., 2011) . The idea that employers can have prior knowledge, expec tations, or perception s of the productivity of certain groups of workers (e.g., older workers) relates to what Phelps (1972) has called “ previous statistical experience.” To maximize expected utility, the employer ’ s beliefs and access to a priori information on certain group averages can lead to discrimination toward individual workers, regardless of individual differences. The existence of negative stereo- types, prejudice, and age discrimination ha s been highlighted in the earlier research litera- ture and considered a push factor that impedes older workers’ employment opportunities (Harris e t al., 2017; Kadefors and Hanse, 2012; Taylor and Walker, 1998; Van Dalen et al., 2010) . For instance, a Norwegian study in the hospitality indus try (e.g., hotels and motels) showed that while a combination of positive and negative attitude s exists among managers, underlying prejudice regarding senior workers and a preference for younger 24 ployed, while members of the female- dominated occupations assistant nurses, adminis- trators, and nurses were less likely to continue working beyond age 65. Th e increased age threshold in the Act can be considered both a push and a maintenance factor depending on the demographic conditions and whether or not an individual intends to continue working beyond the age threshold specified in the Act . The Swedish Work Environment Act requires organizations to conduct health and safety work at the workplace, in what is better known as systematic work environment manage- ment ( SAM). Most Swedish public and private organizations have either an internal or external occupational health service (OHS) provider. The role of the OHS is to support employers and their managerial representatives in occupational health and safety manage- ment (OHS M) to prevent hazardous work environments from forming, safeguard em- ployees’ health and well- being, and provide rehabilitation schemes for those in need. In addition, well- functioning implementation of the Work Environment Act maintains and strengthens older workers’ capability to continue working. However, according to data from the Swedish Work Environment Authority , the presence of OHS providers in Swe- dish workplaces has gradually decreased since the year 2000 (SWEA, 2012a; SWEA, 2020a) , and on ly half of participants stated that systematic work on SAM is pursued at their workplaces (SWEA, 2016:2) . Ageism and age discrimination are caused by how people and employers think, feel, and act regarding older workers and their competence, knowledge, and abilities (Marques et al., 2020) . It is well established in previous research that ageism and negative attitudes among employers are a push factor that restricts older people from, for example, recruit- ment opportunities in the labor market (Ahmed et al., 2012; Conen et al., 2012; Furunes and Mykletun, 2007; Harris et al., 2017; Kadefors and Hanse, 2012) . Furthermore, norms and values still play an important role in retirement decisions. For instance, even though individuals’ health and working environments have improved in recent decades, the 65 - year threshold still constitutes a strong normative reference point and retirement age for many people (Johansson et al., 2015) . These normative and cultural notions are examples of push factors encouraging older individuals to leave the labor market. Also, state legis- lation can contribute to unequal opportunities and age discrimination, pushing people out of employment participation. For instance, the age thresholds in the Employment Pro- tection Act are usually the benchmarks for employment agreements stipulating the age at which the employment relationship ends. Even though the configurations of social secu- rity and pension systems are essential to older workers’ retirement behavior, personal and organizational factors also interact in these relationships, as we will see in the next two sections. 25 Workplace- level determinants The sustainable extension of working life requires that individuals maintain their work ability and motivation in order to continue working. Likewise, organizations must recog- nize older workers as a resource and be able to manage their heterogeneous needs and desires. Today, it is well established that the design or content of work is strongly related to employees’ health and well-being as well as to their motivation, engagement, and productivity. In line with this, the person – job fit literature (Edwards, 1991; Edwards et al., 1998) suggests that early or late retirement behavior can be understood as a result of the (mis)fit between older workers ’ personal characteristics (e.g., abilit ies, preferences, personality, health, and financial resources) and the characteristics of their job s, co -work- ers, organization s, and managers. To exemplify, an older worker might experience reduced work ability that requires some form of work adaptation. However, the organiz ation may not provide such work adaptation, so the individual experiences a misfit. Such misfits are likely to increase older workers’ inclination to take early retirement. Li nking back to the concept of capability, the aforementioned example illustrates that employers play a key role in its materialization. This is not least because of employers’ potential efforts to con- vert “ misfits” into “ fits” through, for example, work adaptation, which is an essential conversion factor for the individual worker and thus a means to strengthen his/her capa- bility (Wang and Shultz, 2010) . Preju dices toward older workers Emp loyers’ willingness and ability to retain and recruit older workers are fundamental preconditions for older workers’ opportunities to prolong their working lives (Conen et al., 2012; Conen et al., 2014; Vick erstaff et al., 2003) . For instance, a longitudinal study in the Netherlands found that having a supervisor who encourage d employees to continue working helped keep older workers longer in the workforce (Van Solinge and Henkens, 2014) . Also, the implementation of age-management activities is reliant on line managers’ knowledge, beliefs, and attitudes toward older workers (Furunes et al., 2011) . The idea that employers can have prior knowledge, expec tations, or perception s of the productivity of certain groups of workers (e.g., older workers) relates to what Phelps (1972) has called “ previous statistical experience.” To maximize expected utility, the employer ’ s beliefs and access to a priori information on certain group averages can lead to discrimination toward individual workers, regardless of individual differences. The existence of negative stereo- types, prejudice, and age discrimination ha s been highlighted in the earlier research litera- ture and considered a push factor that impedes older workers’ employment opportunities (Harris e t al., 2017; Kadefors and Hanse, 2012; Taylor and Walker, 1998; Van Dalen et al., 2010) . For instance, a Norwegian study in the hospitality indus try (e.g., hotels and motels) showed that while a combination of positive and negative attitude s exists among managers, underlying prejudice regarding senior workers and a preference for younger 26 people during recruitment are plausible explanations for why diversity management fails (Furunes and Mykletun, 2007) . A study among employers and employees showed that older workers above the age of 50 were rated substantially lower than their younger coun- terparts in performance, flexibility, physical and mental capacity, and willingness to learn new technologies (Van Dalen et al., 2010) . Moreover, European employers believe that aging leads to losses in productivity (Conen et al., 2012) and that aging workers are more expensive to employ. However, recent evaluations in a Swedish context concluded that people above the age of 65 generally cost less to employ due to lower social fees, taxes, and pension payments (SOU, 2020) . Working conditions Several work-related factors have been identified as associated with retirement in previous research. F or example, research shows that a supportive workplace climate (Bal et al., 2012; Browne et al., 2019; Naegele and Walker, 2006; Zacher and Yang, 2016) and healthy working conditions keep older workers in employment (Carr et al., 2016; Robroek et al., 2013) . Also, empirical evidence shows that job recognition (Carr et al., 2016) and job satisfaction (Kautonen et al., 2012; Topa et al., 2009) are essential factors in keeping older workers in employment. However, a recent review suggests that the associations between social support at work and retirement are mixed (Browne et al., 2019) . In a Swe- dish study, high work demands and low job satisfaction were push factors affecting older workers’ labor market participation (Berglund et al., 2017) . However, g ood health and economic incentives keep older workers in employment, while life partners’ retirement decisions pull older workers toward early retirement (Nilsson et al., 2011) . A European meta-analysis found that job involvement (i.e., individuals ’ psychological identification with their work) a nd job satisfaction keep older workers in employment and postpone retirement intentions (Topa et al., 2009 ) . In addition, older employees’ opportunities for skill development are important for their employability and motivation to continue work- ing (Van Solinge and Henkens, 2014; Zaniboni et al., 2010) . However, a large Danish study by Jensen (2020) found no association between older workers’ access to competence development and their retirement behavior. Despite extensive support for the job de- mands– control model (i.e., jobs as high -strain, active, passive, or low -strain) in occupa- tional stress research (Magnusson Hanson et al., 2008; Van der D oef and Maes, 1999 ) , the job demands – control model appears to lack association with retirement, according to two Nordic studies (Virtanen et al., 2014; Örestig et al., 2013) . Poor working conditions such as strain are a push factor for early labor market withdrawal (Robroek et al., 2013) . Another and perhaps more important condition affecting older workers’ labor market participation is decision latitude (sometimes referred to as job control or work autonomy) , which refers to the extent to which workers can make decisions and exercise control over 27 their work during the working day (Karasek , 1979) . In practice, decision latitude is as- sessed based on workers’ influence over their job s ( in terms of influence on, e.g., working hours, pace, and planning) and their freedom to conduct work tasks (when, what, and how). It is well established in prior literature that high decision latitude is important to buffer high job demands (See the job demand - control model, Karasek, 1979; Van der Doef and Maes, 1999) and affects older workers’ retirement preferences and actual retire- ment (Browne et al., 2019; Robroek et al., 2013) . Decision latitude offers adjustment opportunities when workers experience poor health (see the illness flexibility model, Johansson and Lundberg, 2004) and enhances their abilities to redesign and craft the con- tent of their jobs (see job cr afting, Kooij et al., 2020) . In other words, d ecision latitude has the potential to strengthen (older) workers ’ capabilities by providing opportunities to make temporary or permanent adjustments to current job s to match personal preferences or abilities. Consequently, high decision latitude enhances people ’ s opportunities to self- manage, so that the job fits current needs beyond preexisting organizational HR practices. According to the life- span development theorists Baltes and Baltes (1993 ) , decision lati- tude is particularly important for older workers because it allows them to select , optimize , and compensate for age- related changes in function and motivation. However, the degree of decision latitude is not randomly distributed; certain occupations in the labor market, of- ten in higher socioeconomic positions, have higher autonomy and ability to exercise con- trol over work tasks. Employment contracts usually stipulate agreed working conditions, compensation, working hours, tasks, and responsibilities, and assign employees a certain degree of decision latitude. Therefore, decision latitude or job control is a central compo- nent of neo- Weberian class analysis (see relationships between skill, autonomy, and class in Erikson and Goldthorpe, 1992) . Opportunities for individualized work arrangements While the age- management literature suggests formalized policies and practices to be car- ried out by organizations and their management staff, an alternative is negotiated individ- ualized work arrangements between employer and employees— often called idiosyncratic deals (I - deals). Individual ized work arrangements can be regarded as a “ third way” that goes beyond specific HR strategies targeting older workers or standardized HR practices for all employees. I - deals are negotiated and agreed upon between employees and em- ployers, tailored to individuals’ heterogeneous needs, and assumed to strengthen the per- son– job fit (Hornung et al., 2010) . I diosyncratic deals may concern, for example, task and work responsibilities, workload reduction, schedule flexibility, location flexibility, and fi- nancial incentives. P revious research has shown that employees’ abilities to strike idiosyn- cratic deals are associated with enhanced employability (Oostrom et al., 2016) , productiv- ity (Hornung et al., 2008) , and working beyond retirement (Bal et al., 2012) . However, the existing research on I - deals and retirement is relatively scarce, particularly in a Nordic context in which collective agreement and HR policies geared toward universal solutions 26 people during recruitment are plausible explanations for why diversity management fails (Furunes and Mykletun, 2007) . A study among employers and employees showed that older workers above the age of 50 were rated substantially lower than their younger coun- terparts in performance, flexibility, physical and mental capacity, and willingness to learn new technologies (Van Dalen et al., 2010) . Moreover, European employers believe that aging leads to losses in productivity (Conen et al., 2012) and that aging workers are more expensive to employ. However, recent evaluations in a Swedish context concluded that people above the age of 65 generally cost less to employ due to lower social fees, taxes, and pension payments (SOU, 2020) . Working conditions Several work- related factors have been identified as associated with retirement in previous research. F or example, research shows that a supportive workplace climate (Bal et al., 2012; Browne et al., 2019; Naegele and Walker, 2006; Zacher and Yang, 2016) and healthy working conditions keep older workers in employment (Carr et al., 2016; Robroek et al., 2013) . Also, empirical evidence shows that job recognition (Carr et al., 2016) and job satisfaction (Kautonen et al., 2012; Topa et al., 2009) are essential factors in keeping older workers in employment. However, a recent review suggests that the associations between social support at work and retirement are mixed (Browne et al., 2019) . In a Swe- dish study, high work demands and low job satisfaction were push factors affecting older workers’ labor market participation (Berglund et al., 2017) . However, g ood health and economic incentives keep older workers in employment, while life partners’ retirement decisions pull older workers toward early retirement (Nilsson et al., 2011) . A European meta- analysis found that job involvement (i.e., individuals ’ psychological identification with their work) a nd job satisfaction keep older workers in employment and postpone retirement intentions (Topa et al., 2009 ) . In addition, older employees’ opportunities for skill development are important for their employability and motivation to continue work- ing (Van Solinge and Henkens, 2014; Zaniboni et al., 2010) . However, a large Danish study by Jensen (2020) found no association between older workers’ access to competence development and their retirement behavior. Despite extensive support for the job de- mands– control model (i.e., jobs as high - strain, active, passive, or low - strain) in occupa- tional stress research (Magnusson Hanson et al., 2008; Van der D oef and Maes, 1999 ) , the job demands – control model appears to lack association with retirement, according to two Nordic studies (Virtanen et al., 2014; Örestig et al., 2013) . Poor working conditions such as strain are a push factor for early labor market withdrawal (Robroek et al., 2013) . Another and perhaps more important condition affecting older workers’ labor market participation is decision latitude (sometimes referred to as job control or work autonomy) , which refers to the extent to which workers can make decisions and exercise control over 27 their work during the working day (Karasek , 1979) . In practice, decision latitude is as- sessed based on workers’ influence over their job s ( in terms of influence on, e.g., working hours, pace, and planning) and their freedom to conduct work tasks (when, what, and how). It is well established in prior literature that high decision latitude is important to buffer high job demands (See the job demand -control model, Karasek, 1979; Van der Doef and Maes, 1999) and affects older workers’ retirement preferences and actual retire- ment (Browne et al., 2019; Robroek et al., 2013). Decision latitude offers adjustment opportunities when workers experience poor health (see the illness flexibility model, Johansson and Lundberg, 2004) and enhances their abilities to redesign and craft the con- tent of their jobs (see job cr afting, Kooij et al., 2020) . In other words, d ecision latitude has the potential to strengthen (older) workers ’ capabilities by providing opportunities to make temporary or permanent adjustments to current job s to match personal preferences or abilities. Consequently, high decision latitude enhances people ’ s opportunities to self- manage, so that the job fits current needs beyond preexisting organizational HR practices. According to the life-span development theorists Baltes and Baltes (1993 ) , decision lati- tude is particularly important for older workers because it allows them to select , optimize , and compensate for age-related changes in function and motivation. However, the degree of decision latitude is not randomly distributed; certain occupations in the labor market, of- ten in higher socioeconomic positions, have higher autonomy and ability to exercise con- trol over work tasks. Employment contracts usually stipulate agreed working conditions, compensation, working hours, tasks, and responsibilities, and assign employees a certain degree of decision latitude. Therefore, decision latitude or job control is a central compo- nent of neo- Weberian class analysis (see relationships between skill, autonomy, and class in Erikson and Goldthorpe, 1992) . Opportunities for individualized work arrangements While the age-management literature suggests formalized policies and practices to be car- ried out by organizations and their management staff, an alternative is negotiated individ- ualized work arrangements between employer and employees—often called idiosyncratic deals (I -deals). Individual ized work arrangements can be regarded as a “ third way” that goes beyond specific HR strategies targeting older workers or standardized HR practices for all employees. I -deals are negotiated and agreed upon between employees and em- ployers, tailored to individuals’ heterogeneous needs, and assumed to strengthen the per- son– job fit (Hornung et al., 2010) . I diosyncratic deals may concern, for example, task and work responsibilities, workload reduction, schedule flexibility, location flexibility, and fi- nancial incentives. P revious research has shown that employees’ abilities to strike idiosyn- cratic deals are associated with enhanced employability (Oostrom et al., 2016) , productiv- ity (Hornung et al., 2008) , and working beyond retirement (Bal et al., 2012) . However, the existing research on I -deals and retirement is relatively scarce, particularly in a Nordic context in which collective agreement and HR policies geared toward universal solutions 28 are strong and regulate conditions of employment and the employer– employee relation- ship. Occupational health and safety management Th e Swedish Work Environment Act (AML , 1977 :1160) stipulates employers’ responsi- bility to adapt work needs and provide rehabilitation when employees experience illness or injury . Also , employers are responsible for preventing unhealthy workloads, working times, and victimization (SWEA, 2016 ) . To prevent or treat musculoskeletal disorders or burnout, for instance, it is necessary to have well -functioning collaboration between the employer, HR, OHS providers, managers, safety representatives, and employees at a workplace. Well-organized collaboration between these actors would help keep older workers in employment. However, past studies raise several questions about whether oc- cupational health and safety management (OHSM) in Swedish workplaces meet s the leg- islative requirements . A study conducted in eleven Swedish public-sector organizations found that OHS providers were mainly employed in the aftermath of an incident (e.g., for rehabilitation and wellness activities), being less involved in preventive measures or im- provements in the work environment (Schmidt et al., 2017) . The same study also found that traditional OHS responsibilities are incre asingly being handed over to HR. Because HR personnel often lack the expertise that usually belongs to the OHS professions (e.g., ergonomists, physicians, nurses, and psychologists) , they might fail to identify or lack time to prevent hazardous working conditions. In cases in which OHS providers are used, their primary contact is with the HR staff, limiting their independence and ability to gather sufficient knowledge of the workplace. Schmidt et al. (2017) concluded that the present use of OHS measures in the public sector has a poor chance of improving the work en- vironment as intended by legislation. P oor collaboration between different OHS stake- holders is a potential push factor for older workers and premature retirement. Most of the retirement determinants presented above ha ve been identified from popula- tion- based studies. In contrast , the current knowledge base concerning effective work- place interventions to promote older workers’ employability, work ability , and health is sparse (Cloostermans et al., 2015; Hilsen and Midtsundstad, 2015; Söderbacka et al., 2020; Truxillo et al., 2015) . For instance, a Norwegian study demonstrated that workplace interventions to support older workers in Norway showed weak or no associations be- tween wellbeing, health, work ability and attitudes, and retirement behavior (Hilsen and Midtsundstad, 2015) . Cloostermans et al. (2015) review found a similar lack of evidence for the favorable effects of workplace interventions. However, Söderbacka et al. (2020) found that interventions promoting older workers’ health did enhance work ability and help maintain employment. 29 Individual - level determinants The factors identified in the previous section are known to be associated with an individ- ual’ s retirement preferences, actual retirement behavior, or both. Many of these factors relate to age- related changes in older workers’ psychological and physical functions and motives for working. Improvements in health are one driver of increasing longevity and decreasing mortality. They allow people to live longer and have more active and healthy years. Aging is multidimensional in terms of how it affects our bodies and mind s. Inter- personal variation in age makes chronological age somewhat less relevant as an indicator of physiological and physical capacity (Kanfer and Ackerman, 2004; Truxillo et al., 2015) . However , as people become older , they generally experience a decline in physical and mental function caused by declines in eyesight, hearing, muscular strength, bone structure, and flexibility (Maertens, Putter, Chen, Diehl, & Huang, 2012). Consequently, as people age, their unique character traits become stronger, resulting in greater heterogeneity in terms of needs and motivations at the aggregate level (Carstensen, 1995 ; Kanfer and Ackerman, 2004; Rowe and Kahn, 1987) . Moreover, aging is linked to both growth and decline in cognitive performance (Harada, Love, & Triebel, 2013), and increases the risks of experiencing poor health and reduced work ability (Ilmarinen, 2006) . T he association between poor health and earlier withdrawal from the labor market is well established in the earlier research literature (Siegrist et al., 2007; Sousa - Ribeiro et al., 2021; SWEA, 2014; Van Rijn et al., 2014) . Also , past studies have demonstrated that poor work ability is related to earlier retirement behavior (Boissonneault and de Beer, 2018; Jääskeläinen et al., 2016) . A decline in physical and psychological functions is a common push factor for older workers’ involuntary exit from the workforce. However, a lthough aging entails a general decline in physical and psychological functions, it is not directly related to work ability. A literature survey on age and productivity by Skirbekk (2004) showed that older individuals tend to maintain high productivity in jobs requiring experience and verbal abil- ities; however, their productivity decreases strongly in jobs requiring problem- solving, learning, and speed . T he negative consequences of aging are not randomly distributed in the working popul a- tion; instead, there are social class differences (Halleröd and Gustafsson, 2011; Hasselhorn, 2020; König et al., 2020) . For insta nce, the probability of experiencing poor health is higher in lower socioeconomic groups (Kunst et al., 2005; Mackenbach et al., 2008) . T herefore, PWL policies are likely to have a more substantial impact on certain groups that, due to financial need, have to continue working to sustain living standards or be pushed out of the workforce due to, for example, health reasons (Hofäcker and Radl, 2016) . 28 are strong and regulate conditions of employment and the employer– employee relation- ship. Occupational health and safety management Th e Swedish Work Environment Act (AML , 1977 :1160) stipulates employers’ responsi- bility to adapt work needs and provide rehabilitation when employees experience illness or injury . Also , employers are responsible for preventing unhealthy workloads, working times, and victimization (SWEA, 2016 ) . To prevent or treat musculoskeletal disorders or burnout, for instance, it is necessary to have well - functioning collaboration between the employer, HR, OHS providers, managers, safety representatives, and employees at a workplace. Well- organized collaboration between these actors would help keep older workers in employment. However, past studies raise several questions about whether oc- cupational health and safety management (OHSM) in Swedish workplaces meet s the leg- islative requirements . A study conducted in eleven Swedish public- sector organizations found that OHS providers were mainly employed in the aftermath of an incident (e.g., for rehabilitation and wellness activities), being less involved in preventive measures or im- provements in the work environment (Schmidt et al., 2017) . The same study also found that traditional OHS responsibilities are incre asingly being handed over to HR. Because HR personnel often lack the expertise that usually belongs to the OHS professions (e.g., ergonomists, physicians, nurses, and psychologists) , they might fail to identify or lack time to prevent hazardous working conditions. In cases in which OHS providers are used, their primary contact is with the HR staff, limiting their independence and ability to gather sufficient knowledge of the workplace. Schmidt et al. (2017) concluded that the present use of OHS measures in the public sector has a poor chance of improving the work en- vironment as intended by legislation. P oor collaboration between different OHS stake- holders is a potential push factor for older workers and premature retirement. Most of the retirement determinants presented above ha ve been identified from popula- tion- based studies. In contrast , the current knowledge base concerning effective work- place interventions to promote older workers’ employability, work ability , and health is sparse (Cloostermans et al., 2015; Hilsen and Midtsundstad, 2015; Söderbacka et al., 2020; Truxillo et al., 2015) . For instance, a Norwegian study demonstrated that workplace interventions to support older workers in Norway showed weak or no associations be- tween wellbeing, health, work ability and attitudes, and retirement behavior (Hilsen and Midtsundstad, 2015) . Cloostermans et al. (2015) review found a similar lack of evidence for the favorable effects of workplace interventions. However, Söderbacka et al. (2020) found that interventions promoting older workers’ health did enhance work ability and help maintain employment. 29 Individual - level determinants The factors identified in the previous section are known to be associated with an individ- ual’ s retirement preferences, actual retirement behavior, or both. Many of these factors relate to age-related changes in older workers ’ psychological and physical functions and motives for working. Improvements in health are one driver of increasing longevity and decreasing mortality. They allow people to live longer and have more active and healthy years. Aging is multidimensional in terms of how it affects our bodies and mind s. Inter- personal variation in age makes chronological age somewhat less relevant as an indicator of physiological and physical capacity (Kanfer and Ackerman, 2004; Truxillo et al., 2015) . However , as people become older , they generally experience a decline in physical and mental function caused by declines in eyesight, hearing, muscular strength, bone structure, and flexibility (Maertens, Putter, Chen, Diehl, & Huang, 2012). Consequently, as people age, their unique character traits become stronger, resulting in greater heterogeneity in terms of needs and motivations at the aggregate level (Carstensen, 1995 ; Kanfer and Ackerman, 2004; Rowe and Kahn, 1987) . Moreover, aging is linked to both growth and decline in cognitive performance (Harada, Love, & Triebel, 2013), and increases the risks of experiencing poor health and reduced work ability (Ilmarinen, 2006) . T he association between poor health and earlier withdrawal from the labor market is well established in the earlier research literature (Siegrist et al., 2007; Sousa -Ribeiro et al., 2021; SWEA, 2014; Van Rijn et al., 2014) . Also , past studies have demonstrated that poor work ability is related to earlier retirement behavior (Boissonneault and de Beer, 2018; Jääskeläinen et al., 2016) . A decline in physical and psychological functions is a common push factor for older workers’ involuntary exit from the workforce. However, a lthough aging entails a general decline in physical and psychological functions, it is not directly related to work ability. A literature survey on age and productivity by Skirbekk (2004) showed that older individuals tend to maintain high productivity in jobs requiring experience and verbal abil- ities; however, their productivity decreases strongly in jobs requiring problem- solving, learning, and speed . T he negative consequences of aging are not randomly distributed in the working popul a- tion; instead, there are social class differences (Halleröd and Gustafsson, 2011; Hasselhorn, 2020; König et al., 2020) . For insta nce, the probability of experiencing poor health is higher in lower socioeconomic groups (Kunst et al., 2005; Mackenbach et al., 2008) . T herefore, PWL policies are likely to have a more substantial impact on certain groups that, due to financial need, have to continue working to sustain living standards or be pushed out of the workforce due to, for example, health reasons (Hofäcker and Radl, 2016) . 30 O lder workers’ retirement timing is not only important for themselves, but also has im- plications for families, organizations, and society. Factors outside work such as social re- lations, family, and spousal retirement are pull factors influencing older workers’ early retirement decisions (Eismann et al., 2019; Elovainio et al., 2003; Litwin and Tur -Sinai, 2015; Nilsson et al., 2011) . For example, a Swedish study showed that responsibilities to care for aging parents increased the probability of early labor market withdrawal in both men and women (Kridahl and Silverstein, 2019) . Also, the ability to balance leisure with work can be either a push or pull factor for older workers (Richert-Kaźmierska and Stankiewicz, 2016) . To sum up, social relations outside work or the lack thereof can either keep older workers in employment or pull them toward early retirement. 31 5 Data and methods This thesis combines quantitative and qualitative methods. The data ha ve been collected in Sweden from a population- based representative sample of employed Swedes (Article s I and II), public welfare employees in the city of Gothenburg (Article III) , and lower managers and their H R partners in the public healthcare sector from two regions in Swe- den (Article IV). Table 1 presents a summary of the studies. All studies focus directly or indirectly on older workers defined as individuals aged 55 years or older. Table 1. Overview of data and the articles. ARTIC LE I II III IV TITLE Preferred versus ac- tual retirement and their personal and work- related predic- tors: A comparison using representative Swedish survey and register data Work despite poor health? A 14 - year follow- up of how individual work ac- commodations are extending the time to retirement for workers with poor health conditions Matching the pieces: The pres- ence of idiosyn- cratic deals and their impact on re- tirement prefer- ences among older workers Organizational hindrances to the retention of older healthcare workers DATA Survey/ register data: ULF/PSAE , 2002/2003 ; LISA , 2002– 2015 Survey/ register data: ULF/PSAE , 2002/2003 ; LISA , 2002– 2015 Survey data: HEART S– LEXLI V, 2019 Data from Ar- betsK raft: rec- orded and tran- scribed interviews, 2017 – 2018 DESIGN Cross - sectional and longitudinal register data Cross - sectional and longitudinal register data Cross - sectional survey data Intervention pro- gram with follow- up interviews SAMPLE A representative sample of Swedes born 1938 – 1943, n = 1143 A representa tive sample of Swedes born 1938 – 1943, n = 1143 P ublic welfare em- ployees n = 4499 Six organizations in healthcare, nine line managers, and 11 HR partners ANALYSIS Cox proportional hazard and linear re- gressions Cox proportional hazard regression Confirmatory fac- tor analysis/ struc- tural equation modeling Qualitative the- matic analysis OUTCOME Self- reported retire- ment preferences; actual retirement measured by a drop in work- related in- come Actual retirement measured by a drop in work- re- lated income Self- reported re- tirement prefer- ences 30 O lder workers’ retirement timing is not only important for themselves, but also has im- plications for families, organizations, and society. Factors outside work such as social re- lations, family, and spousal retirement are pull factors influencing older workers’ early retirement decisions (Eismann et al., 2019; Elovainio et al., 2003; Litwin and Tur - Sinai, 2015; Nilsson et al., 2011) . For example, a Swedish study showed that responsibilities to care for aging parents increased the probability of early labor market withdrawal in both men and women (Kridahl and Silverstein, 2019) . Also, the ability to balance leisure with work can be either a push or pull factor for older workers (Richert - Kaźmierska and Stankiewicz, 2016) . To sum up, social relations outside work or the lack thereof can either keep older workers in employment or pull them toward early retirement. 31 5 Data and methods This thesis combines quantitative and qualitative methods. The data ha ve been collected in Sweden from a population-based representative sample of employed Swedes (Article s I and II), public welfare employees in the city of Gothenburg (Article III) , and lower managers and their H R partners in the public healthcare sector from two regions in Swe- den (Article IV). Table 1 presents a summary of the studies. All studies focus directly or indirectly on older workers defined as individuals aged 55 years or older. Table 1. Overview of data and the articles. ARTIC LE I II III IV TITLE Preferred versus ac- tual retirement and their personal and work-related predic- tors: A comparison using representative Swedish survey and register data Work despite poor health? A 14 -year follow-up of how individual work ac- commodations are extending the time to retirement for workers with poor health conditions Matching the pieces: The pres- ence of idiosyn- cratic deals and their impact on re- tirement prefer- ences among older workers Organizational hindrances to the retention of older healthcare workers DATA Survey/ register data: ULF/PSAE , 2002/2003 ; LISA , 2002– 2015 Survey/ register data: ULF/PSAE , 2002/2003 ; LISA , 2002– 2015 Survey data: HEART S– LEXLI V, 2019 Data from Ar- betsK raft: rec- orded and tran- scribed interviews, 2017 – 2018 DESIGN Cross - sectional and longitudinal register data Cross - sectional and longitudinal register data Cross - sectional survey data Intervention pro- gram with follow- up interviews SAMPLE A representative sample of Swedes born 1938 – 1943, n = 1143 A representa tive sample of Swedes born 1938 – 1943, n = 1143 P ublic welfare em- ployees n = 4499 Six organizations in healthcare, nine line managers, and 11 HR partners ANALYSIS Cox proportional hazard and linear re- gressions Cox proportional hazard regression Confirmatory fac- tor analysis/ struc- tural equation modeling Qualitative the- matic analysis OUTCOME Self- reported retire- ment preferences; actual retirement measured by a drop in work-related in- come Actual retirement measured by a drop in work-re- lated income Self- reported re- tirement prefer- ences 32 PREDICTORS Personal factors ( seven items) , work- related resources ( eight items) , and work-related de- mands ( five items) ; see Table 4 for de- tails. Health (illness, dis- ease, and self-rated health) ; decision latitude in terms of influence ( working hours, work pace, and planning) and freedom ( when, how, and what ) Task and work re- sponsibilities, workload reduc- tion, schedule flex- ibility, location flexibility, and fi- nancial incentives COVARIATES Age, cohort, SES, status, education Age, cohort, SES, status, education Age, cohort, SES, status, education, self-rated health, organizational ten- ure The quantitative studies Before presenting the quantitative data sets, study sample s, and data collection, error sources in the survey and sampling methodology are described based on Scheaffer et al. (2011) ; after that, some of these error sources are discussed in relation to each dataset. Panel Survey of Ag eing and the Elderly The data used in Article s I and II come from two sources, a cross- sectional survey con- ducted from 2002 to 2003 and a longitudinal data register containing income information from 2002 to 2015 (see Figure 4 ). The cross - sectional data come from the first wave of the Panel Survey on Ageing and the Elderly (PSAE), 2002– 2003 , which is integrated with Statistics Sweden’ s annual population-based survey of living conditions (ULF/SILC) . PSAE is population based, but the 65+ age category has been oversampled and recruited from a panel previously enrolled in the ULF/SILC survey program. PSAE covers many variables capturing living conditions, working condition s, health, and social relation s, but only a limited number were used in the present analysis. In addition, t he PSAE dataset is linked to data from the Longitudinal Integration Data base for Health Insurance and La- bor Market Studies (LISA) u sing the participants’ social security numbers. From LISA , income information was obtained to determine an individual’ s connection to the labor market between 2002 and 2015 ( the follow-up period). LISA contains detailed infor- mation on individuals’ income and compensation sources (e.g., unemployment insurance, parental leave, development allowance, activity grant, disability pension , and old-age pen- sion). 33 Study sample Statistics Sweden collected PSAE data in 2002 and 2003. Based on information from Statistics Sweden (2020b) , the participants in PSA E were selected using simple random sampling from the Total Population Register (Registret över totalbefolkningen, RTB ) and by recruiting respondents previously included in earlier waves of ULF in 1993 and 1994 , but the 65 + age category was over- sampled. T he total sample consists of a total of 12, 68 5 participants, out whom 5374 are from the 55+ group. Many of these participants were excluded from the final analysis as they were considered outside the target population. For instance, farmers, self - employed, and respondents younger than 55 and older than 64 years were excluded. T herefore, t he study sample consists of employed individuals aged 55 – 64 years in 2002 – 2003 who were employed and had a work- related income above the price basic amount (PBA) at the time of data collection. Data collection Statistics Sweden collected the PSAE data by combining face- to- face and telephone per- sonal interviews. The response rate for ULF , 2002– 2003 , was 75 % (Statistics Sweden, 2020b) ; however, as far as I am aware, there are no existing response analyses of the data. Figure 4. Flow chart of the PSAE dataset . 32 PREDICTORS Personal factors ( seven items) , work- related resources ( eight items) , and work- related de- mands ( five items) ; see Table 4 for de- tails. Health (illness, dis- ease, and self- rated health) ; decision latitude in terms of influence ( working hours, work pace, and planning) and freedom ( when, how, and what ) Task and work re- sponsibilities, workload reduc- tion, schedule flex- ibility, location flexibility, and fi- nancial incentives COVARIATES Age, cohort, SES, status, education Age, cohort, SES, status, education Age, cohort, SES, status, education, self- rated health, organizational ten- ure The quantitative studies Before presenting the quantitative data sets, study sample s, and data collection, error sources in the survey and sampling methodology are described based on Scheaffer et al. (2011) ; after that, some of these error sources are discussed in relation to each dataset. Panel Survey of Ag eing and the Elderly The data used in Article s I and II come from two sources, a cross- sectional survey con- ducted from 2002 to 2003 and a longitudinal data register containing income information from 2002 to 2015 (see Figure 4 ). The cross - sectional data come from the first wave of the Panel Survey on Ageing and the Elderly (PSAE), 2002– 2003 , which is integrated with Statistics Sweden’ s annual population- based survey of living conditions (ULF/SILC) . PSAE is population based, but the 65+ age category has been oversampled and recruited from a panel previously enrolled in the ULF/SILC survey program. PSAE covers many variables capturing living conditions, working condition s, health, and social relation s, but only a limited number were used in the present analysis. In addition, t he PSAE dataset is linked to data from the Longitudinal Integration Data base for Health Insurance and La- bor Market Studies (LISA) u sing the participants’ social security numbers. From LISA , income information was obtained to determine an individual’ s connection to the labor market between 2002 and 2015 ( the follow- up period). LISA contains detailed infor- mation on individuals’ income and compensation sources (e.g., unemployment insurance, parental leave, development allowance, activity grant, disability pension , and old- age pen- sion). 33 Study sample Statistics Sweden collected PSAE data in 2002 and 2003. Based on information from Statistics Sweden (2020b) , the participants in PSA E were selected using simple random sampling from the Total Population Register (Registret över totalbefolkningen, RTB ) and by recruiting respondents previously included in earlier waves of ULF in 1993 and 1994 , but the 65 + age category was over-sampled. T he total sample consists of a total of 12,68 5 participants, out whom 5374 are from the 55+ group. Many of these participants were excluded from the final analysis as they were considered outside the target population. For instance, farmers, self -employed, and respondents younger than 55 and older than 64 years were excluded. Therefore, t he study sample consists of employed individuals aged 55– 64 years in 2002 – 2003 who were employed and had a work-related income above the price basic amount (PBA) at the time of data collection. Data collection Statistics Sweden collected the PSAE data by combining face- to-face and telephone per- sonal interviews. The response rate for ULF , 2002– 2003, was 75% (Statistics Sweden, 2020b) ; however, as far as I am aware, there are no existing response analyses of the data. Figure 4. Flow chart of the PSAE dataset . 34 H earts– Lexliv dataset The data used in Article III are based on the Hearts – Lexliv web-based survey covering public welfare employees above the age of 55 in the city of Gothenburg ( see F igure 5 and Table 2). The participants were invited to participate in the survey and were informed of the aim of the study, the condition of voluntary participation, their right to withdraw without penalty, and the guarantee of anonymity and confidentiality. The dataset covers a vast number of items regarding, for instance, preferences and motivations in relation to retirement, health, working conditions , and personal life. A more detailed overview of the Hearts – Lexliv data was presented by Seldén et al. (2020) ; the information regarding the study sample and data collection presented below is based on that source. Study sample The sample was selected and data were collected by the Lexliv research group at the Uni- versity of Gothenburg (Seldén et al., 2020) . Al l monthly and permanently employed em- ployees in the city of Gothenburg were asked to participate in an online survey ( n = 10, 485). Out of the 10 , 485 employees in the target population, 4499 employees completed the survey; they had various occupations, such as social workers, teachers, technical sup- port staff, assistan t nurses, and engineers in various operations in the city of Gothenburg. Data collection In September 2019, a ll employees in the target population were asked to participate in an online survey. Before and after data collection, employees were prepared and encouraged by their closest managers to fill out the questionnaire during working hours. Also, infor- mation about the survey was available on the city of Gothenburg’ s internal website. After obtaining an email list of all employees in the city of Gothenburg, the survey was sent out on 20 September 2019 using the Qualtrics survey platform. After the first email- ing of the survey, two reminders were sent to non-responders between 20 September and 20 November 2019. Thirty -six employees actively withdrew from participation for various reasons, often due to lack of time or interest. A total of 5 940 employees chose not to respond at all. After data collection, a limited response analysis was conducted, which indicated that employees with higher education were overrepresented, probably because this group uses computers and email daily. Likewise, women were also overrepresented in the sample. 35 Figure 5. Flow chart of the Hearts– Lexliv dataset . Table 2. Descriptive statistics for PSAE and Hearts– Lexliv . Sociodemographic variables PSAE , n (%) H earts – Lexliv , n (%) Sex Male 537 (46) 974 (2 2) Female 606 (54) 3528 (78 ) Age at baseline 55 – 59 743 (65) 2199 ( 49 ) 60 – 64 400 (35) 1892 (42 ) >65 408 (9 ) Socioeconomic status Higher managerial/ higher- grade service class 194 (17) 674 (15 ) Middle managerial/ lower- grade service class 308 (27) 1096 (24 ) Lower managerial/ skilled workers 209 (18) 1551 ( 35 ) Skilled manual/ unskilled workers 169 (15) 1178 ( 26 ) Unskilled manual 263 (23) - Education University or higher 397 (35) 2250 (63 ) High school 491 (43) 1159 (3 3 ) Elementary school or less 251 (22) 144 (4 ) Operationalization of the dependent variable In past studies, retirement has been operationalized in terms of intentions, plans, thoughts, preferences , and actual retirement (Denton and Spencer, 2009) . A distinction 34 H earts– Lexliv dataset The data used in Article III are based on the Hearts – Lexliv web- based survey covering public welfare employees above the age of 55 in the city of Gothenburg ( see F igure 5 and Table 2). The participants were invited to participate in the survey and were informed of the aim of the study, the condition of voluntary participation, their right to withdraw without penalty, and the guarantee of anonymity and confidentiality. The dataset covers a vast number of items regarding, for instance, preferences and motivations in relation to retirement, health, working conditions , and personal life. A more detailed overview of the Hearts – Lexliv data was presented by Seldén et al. (2020) ; the information regarding the study sample and data collection presented below is based on that source. Study sample The sample was selected and data were collected by the Lexliv research group at the Uni- versity of Gothenburg (Seldén et al., 2020) . Al l monthly and permanently employed em- ployees in the city of Gothenburg were asked to participate in an online survey ( n = 10, 485). Out of the 10 , 485 employees in the target population, 4499 employees completed the survey; they had various occupations, such as social workers, teachers, technical sup- port staff, assistan t nurses, and engineers in various operations in the city of Gothenburg. Data collection In September 2019, a ll employees in the target population were asked to participate in an online survey. Before and after data collection, employees were prepared and encouraged by their closest managers to fill out the questionnaire during working hours. Also, infor- mation about the survey was available on the city of Gothenburg’ s internal website. After obtaining an email list of all employees in the city of Gothenburg, the survey was sent out on 20 September 2019 using the Qualtrics survey platform. After the first email- ing of the survey, two reminders were sent to non- responders between 20 September and 20 November 2019. Thirty - six employees actively withdrew from participation for various reasons, often due to lack of time or interest. A total of 5 940 employees chose not to respond at all. After data collection, a limited response analysis was conducted, which indicated that employees with higher education were overrepresented, probably because this group uses computers and email daily. Likewise, women were also overrepresented in the sample. 35 Figure 5. Flow chart of the Hearts– Lexliv dataset . Table 2. Descriptive statistics for PSAE and Hearts– Lexliv . Sociodemographic variables PSAE , n (%) H earts – Lexliv , n (%) Sex Male 537 (46) 974 (2 2) Female 606 (54) 3528 (78 ) Age at baseline 55 – 59 743 (65) 2199 ( 49 ) 60 – 64 400 (35) 1892 (42 ) >65 408 (9 ) Socioeconomic status Higher managerial/ higher- grade service class 194 (17) 674 (15 ) Middle managerial/ lower- grade service class 308 (27) 1096 (24 ) Lower managerial/ skilled workers 209 (18) 1551 ( 35 ) Skilled manual/ unskilled workers 169 (15) 1178 ( 26 ) Unskilled manual 263 (23) - Education University or higher 397 (35) 2250 (63 ) High school 491 (43) 1159 (3 3 ) Elementary school or less 251 (22) 144 (4 ) Operationalization of the dependent variable In past studies, retirement has been operationalized in terms of intentions, plans, thoughts, preferences , and actual retirement (Denton and Spencer, 2009) . A distinction 36 can be made as to whether retirement is measured subjective ly or objective ly and prospec- tively or retrospectively (see Table 3 ). R etirement preferences were measured using a survey- based question , while actual retirement was measured using income information from regis- ter data, with a drop in work-related income data being taken as a sign of actual retirement. Below , I present the oper ationalization of each dependent variable. Table 3. D ifferent ways of measuring retirement. Prospective Retrospective Subjective Self- reported survey question: retirement preferences (Article s I and III) Self- reported survey question: date of retirement from work Objective Income information from register data: “ actual” retirement (Article s I and II) The question about the dependent variable retirement preferences in Article I was the follow- ing: “ Today , you have a certain degree of influence on the timing of retirement. You can retire before or after the age of 65. Taking your health, your economic condition s, your work, and the present retirement regulations into account, at what age do you want to retire? ” Responders answered using a continuous scale ranging from 50 to 75 years. The question about retirement preferences in Article III differed slightly from that in Article I: “ Today, you have the opportunity to choose when you want to retire. It can be either before or after you turn 65. If you think about your situation today, at what age do you want to retire (completely) ? ” Responses were given on a restricted continuous scale: <60, 61 … 69, >70 years. Actual retirement was assessed in Article s I and II by classifying individuals as “ active” or “ retired” using income information from LISA over the follow -up period. LISA contains detailed information on an individual’ s income and compensation sources, helping us to determine whether the total annual income is from work-related ( e.g., unemployment in- surance, parental leave, development allowance, or activity grant) or pension -related ( e.g., disability pension and old age pension) sources. Participants were classified as retired when they had a work-rel ated income below the price basic amount ( P BA ) for two con- secutive years, and were otherwise classified as active (Johansson et al., 2018) . The PBA reflects changes in the general price level and is annually adjus ted by Statistics Sweden in line with changes in the consumer price index. For example, the PBA was SEK 37 ,900 in 2002 and SEK 44,500 in 2015 ( Statistics Sweden, 2019) . T o illustrate, the PBA can be compared with the average annual salary in 2002 of SEK 345,600 in the private sector (full -time non -manual workers in mining, quarrying, and manufacturing) and SEK 228,000 in the public municipal s ector (Statistics Sweden, 2020a) . The information for assessing the year of retirement was used in Article s I and I I to construct the continuous 37 component (i.e., years at risk of retirement) and the dichotomous component (i.e., ac- tive/retired) of the Cox regression models (see section “ Statistical methods” ). Operationalization of the independent variables All personal and work - related predictors used in Article s I and II are presented in Table 4 , together with their question s (translated from Swedish to English) and coding. Table 4. Coding of independent variables . English translation Coding Personal factors Self -rated health Question: “ How do you assess your general health condition? ” ( ordinal scale 1– 5) 1 = “ Very good ,” 5 = “ Very poor ” Disease Question: “ Do you have any long- term disease, trouble after an accident, disability, or other frailty? ” Dummy variable: 1 = “ Yes ,” 0 = “ No ” Sickness “ How many days of sick leave have you taken in the previous 12 months?” ( ordinal scale 1– 5) “ Good” (sickness absence < 8 days/12 months), “ Bad ” (sickness absence • 8 days/12 months) Pain in the neck or shoul- ders Question: “ Do you have either of the following [health problems]: pain in the neck or shoul- ders? ” Dummy variable: 1 = “ Yes , se- vere/light ,” 0 = “ No ” Physical work ability Question: “ Rate your physical work ability in re- lation to the physical demands of your current work. ” ( visual analog scale, 0– 10) 1 = “ Very bad ,” 10 = “ V ery good” Psych ological work ability Question: “ Rate your psychological work ability in relation to the psychological demands of your current work. ” ( visual analog scale, 0– 10) 1 = “ Very bad ,” 10 = “ V ery good” Caring re- sponsibility Question: “ Do you regularly help another person who is old, sick , or handicapped and who lives either here in your household or elsewhere? ” (or- dinal scale 1– 5) Dummy variable: 1 = “ Yes ” ( “ Daily, ” “ Several times per week,” “ Some times a week,” “ Some times a month” ), 0 = “ No ” No cash re- serve Question: “ If you suddenly end ed up in an un- foreseen situation in which you have to raise SEK 14,000 in one week, could you manage it? ” Dummy variable: 1 = “ No, ” 0 = “ Yes ” Job resources Organiza- tional climate Question s: “ Do you think that the workplace co- hesion is good? ” ; “ Do you get along well with your superior managers? ” ; and “ Do you think that there are often troubles and conflicts at work? ” ( ordinal scale 1– 4, all three it ems were summed ranging from 3 to 12) Higher values represent higher levels of organizational climate; the first five points were merged, and the indicator ranges from 1 to 6, ordinal scale. 36 can be made as to whether retirement is measured subjective ly or objective ly and prospec- tively or retrospectively (see Table 3 ). R etirement preferences were measured using a survey- based question , while actual retirement was measured using income information from regis- ter data, with a drop in work- related income data being taken as a sign of actual retirement. Below , I present the oper ationalization of each dependent variable. Table 3. D ifferent ways of measuring retirement. Prospective Retrospective Subjective Self- reported survey question: retirement preferences (Article s I and III) Self- reported survey question: date of retirement from work Objective Income information from register data: “ actual” retirement (Article s I and II) The question about the dependent variable retirement preferences in Article I was the follow- ing: “ Today , you have a certain degree of influence on the timing of retirement. You can retire before or after the age of 65. Taking your health, your economic condition s, your work, and the present retirement regulations into account, at what age do you want to retire? ” Responders answered using a continuous scale ranging from 50 to 75 years. The question about retirement preferences in Article III differed slightly from that in Article I: “ Today, you have the opportunity to choose when you want to retire. It can be either before or after you turn 65. If you think about your situation today, at what age do you want to retire (completely) ? ” Responses were given on a restricted continuous scale: <60, 61 … 69, >70 years. Actual retirement was assessed in Article s I and II by classifying individuals as “ active” or “ retired” using income information from LISA over the follow - up period. LISA contains detailed information on an individual’ s income and compensation sources, helping us to determine whether the total annual income is from work- related ( e.g., unemployment in- surance, parental leave, development allowance, or activity grant) or pension - related ( e.g., disability pension and old age pension) sources. Participants were classified as retired when they had a work- related income below the price basic amount ( P BA ) for two con- secutive years, and were otherwise classified as active (Johansson et al., 2018) . The PBA reflects changes in the general price level and is annually adjus ted by Statistics Sweden in line with changes in the consumer price index. For example, the PBA was SEK 37 ,900 in 2002 and SEK 44,500 in 2015 ( Statistics Sweden, 2019) . T o illustrate, the PBA can be compared with the average annual salary in 2002 of SEK 345,600 in the private sector (full - time non- manual workers in mining, quarrying, and manufacturing) and SEK 228,000 in the public municipal s ector (Statistics Sweden, 2020a) . The information for assessing the year of retirement was used in Article s I and I I to construct the continuous 37 component (i.e., years at risk of retirement) and the dichotomous component (i.e., ac- tive/retired) of the Cox regression models (see section “ Statistical methods” ). Operationalization of the independent variables All personal and work -related predictors used in Article s I and II are presented in Table 4 , together with their question s (translated from Swedish to English) and coding. Table 4. Coding of independent variables . English translation Coding Personal factors Self -rated health Question: “ How do you assess your general health condition? ” ( ordinal scale 1– 5) 1 = “Very good ,” 5 = “Very poor ” Disease Question: “ Do you have any long- term disease, trouble after an accident, disability, or other frailty? ” Dummy variable: 1 = “Yes ,” 0 = “No ” Sickness “How many days of sick leave have you taken in the previous 12 months?” ( ordinal scale 1– 5) “ Good” (sickness absence < 8 days/12 months), “ Bad ” (sickness absence • 8 days/12 months) Pain in the neck or shoul- ders Question: “ Do you have either of the following [health problems]: pain in the neck or shoul- ders? ” Dummy variable: 1 = “Yes , se- vere/light ,” 0 = “No ” Physical work ability Question: “Rate your physical w ork ability in re- lation to the physical demands of your current work. ” ( visual analog scale, 0– 10) 1 = “Very bad ,” 10 = “V ery good” Psych ological work ability Question: “Rate your psycholog ical work ability in relation to the psychological demands of your current work. ” ( visual analog scale, 0– 10) 1 = “Very bad ,” 10 = “V ery good” Caring re- sponsibility Question: “ Do you regularly help another person who is old, sick , or handicapped and who lives either here in your household or elsewhere?” (or- dinal scale 1– 5) Dummy variable: 1 = “ Yes ” ( “ Daily, ” “Several times per week ,” “Some times a week,” “Some times a month”), 0 = “ No ” No cash re- serve Question: “ If you suddenly end ed up in an un- foreseen situation in which you have to raise SEK 14,000 in one week, could you manage it? ” Dummy variable: 1 = “No, ” 0 = “Yes ” Job resources Organiza- tional climate Question s: “D o you think that the workplace co- hesion is good? ”; “ Do you get along well with your superior managers?”; and “ Do you think that there are often troubles and conflicts at work? ” ( ordinal scale 1– 4, all three it ems were summed ranging from 3 to 12) Higher values represent higher levels of organizational climate; the first five points were merged, and the indicator ranges from 1 to 6, ordinal scale. 38 Supervisor support Question: “ if you need, do you get support from the closest manager” ( ordinal scale 1– 3) Dummy variable: 1 = “ Yes, to a large ex- tent/to some extent,” 0 = “ No ” Co -worker support Question: “ if you need, do you get support from colleagues when needed” ( ordinal scale 1– 3) Dummy variable: 1 = “ Yes, to a large ex- tent/to some extent,” 0 = “ No ” Supervisor recognition Question: “ Do you think that your work is rec- ognized by your superior managers? ” ( ordinal scale 1– 3) Dummy variable: 1 = “Yes, to a large ex- tent/to some extent,” 0 = “No ” Co -worker recognition Question: “ Do you think that your work is rec- ognized by your co- workers? ” ( ordinal scale 1– 3) Dummy variable: 1 = “ Yes, to a large ex- tent/to some extent,” 0 = “ No ” Job satisfac- tion Question: “How do you feel when you are on your way to work?” ( ordinal scale 1– 6) 1 = “ Feel strong uneasiness before work,” 6 = “ Feel happy and satisfied at the thought of the work that awaits” Growth op- portunities Question: “ Is your work such that you have many opportunities to learn new things? ” Dummy variable: 1 = “Yes ,” 0 = “No ” Instrumental view of work Question: “Which of these sentences best matches what you think about the work you have now? ” Dummy variable, answers : 1 = “ This job is like any other job. You do your own thing, but the only thing that matters is the salary”; 0 = “This job is something special. Besides the salary, it gives me a feeling of personal satisfaction. ” Work– life imbalance Question: “How often is your work difficult to combine with family and leisure time? ” ( ordinal scale 1– 5) 1 = “ Never,” 5 = “Almost every day ” Job demands Two q uestions: “ How often does your work re- quire you to work very hard? ” ; and “How often does your work require far too great an effort? ” (ordinal scale 1 – 5) 1 = “N ever,” 5 = “B asically every day”; the two items were summed, with h igher values representing higher j ob demands Job strain This item is based on a combination of “Job de- mands” and “Job control.” Both these indexes were dichotomized at the median value (Job de- mands = 10, Job control = 6) and coded into four categories: High-strain jobs ( high job de- mands, low job control), Passive jobs ( low job de- mands, high job control), Active jobs ( high job de- mands, high job control) , and Low -strain jobs ( low job demands, low j ob control). A similar ap- proach using median values has been applied in earlier studies (cf. Jonsson et al., 2013) . Only Job strain was included in the analysis. Categorical dummy variable, Job strain Physical work exposure The variable consists of nine indicators of job characteristics: repeated and one-sided move- ments ( two items), crooked and twisted postures ( two items), becoming sweaty on a daily basis ( two items), shaking and vibrations ( two items), and heavy lifting ( one item). The number of indi- cators summed, ordinal 0– 9. 0 = “ No to all items,” 1 = “ Yes to a sin- gle item,” 2 = “Y es to two items, ” up to nine items 39 Personal h ealth indicators In Article II, three indicators of poor health ( i.e., disease, illness, and sickness ) w ere in- cluded in the analysis (see Table 4 ) to capture a multidimensional and holistic understand- ing of health. Marinker ’ s (1975 ) tripartite schema of disease, illness , and sickness captures the broader objective, subjective, and contextual aspects of poor health. Disease can be defined as health conditions, diagnosed by a doctor into standardized or specific codes, that often can be seen, touched, measured, or smelled. These conditions are often known and treatable but can sometimes be unspecific due to subjective information and the in- terrelationship between different body parts and organs. For example, illness is a personal, self- reported temporary or permanent mental or physical condition that sometimes may precede disease. Illness i s often a subjective interior feeling or experience that can range from being a minor to a more serious health- related problem for the individual. Finally, sickness is the social role of poor health and the contextual meaning of being “ sick” in the surrounding society; sickness is often related to an individual’ s absence from work. Decision latitude PSAE include s several questions about individuals ’ freedom and influence over their work (Article s I and II) . P articipants were asked to rate their freedom to decide when to carry out their work, how to carry out their work, and what tasks to engage in on a scale ranging from 1 (none) to 4 (great). Subsequently, they were asked to rate their level of influence on working hours, wor king pace, and planning of work on a scale ranging from 1 (none) to 3 (great) (see Table 5 ) . In Article I I , the question s were dichotomized in the following way: low influence (no influence, some influence) and high influence (great influence); and low freedom (no, little, and some freedom) and high freedom (great freedom). The dichotomization was conducted to construct typologies based on influence/freedom and health condition. In Article I, the questions regarding freedom at work were used to con- struct an index. 38 Supervisor support Question: “ if you need, do you get support from the closest manager” ( ordinal scale 1– 3) Dummy variable: 1 = “ Yes, to a large ex- tent/to some extent,” 0 = “ No ” Co -worker support Question: “ if you need, do you get support from colleagues when needed” ( ordinal scale 1– 3) Dummy variable: 1 = “ Yes, to a large ex- tent/to some extent,” 0 = “ No ” Supervisor recognition Question: “ Do you think that your work is rec- ognized by your superior managers? ” ( ordinal scale 1– 3) Dummy variable: 1 = “ Yes, to a large ex- tent/to some extent,” 0 = “ No ” Co -worker recognition Question: “ Do you think that your work is rec- ognized by your co- workers? ” ( ordinal scale 1– 3) Dummy variable: 1 = “ Yes, to a large ex- tent/to some extent,” 0 = “ No ” Job satisfac- tion Question: “ How do you feel when you are on your way to work? ” ( ordinal scale 1– 6) 1 = “ Feel strong uneasiness before work,” 6 = “ Feel happy and satisfied at the thought of the work that awaits” Growth op- portunities Question: “ Is your work such that you have many opportunities to learn new things? ” Dummy variable: 1 = “ Yes ,” 0 = “ No ” Instrumental view of work Question: “ Which of these sentences best matches what you think about the work you have now? ” Dummy variable, answers : 1 = “ This job is like any other job. You do your own thing, but the only thing that matters is the salary” ; 0 = “ This job is something special. Besides the salary, it gives me a feeling of personal satisfaction. ” Work– life imbalance Question: “ How often is your work difficult to combine with family and leisure time? ” ( ordinal scale 1– 5) 1 = “ Never ,” 5 = “ Almost every day ” Job demands Two q uestions: “ How often does your work re- quire you to work very hard? ” ; and “ How often does your work require far too great an effort? ” (ordinal scale 1 – 5) 1 = “N ever,” 5 = “B asically every day”; the two items were summed, with h igher values representing higher j ob demands Job strain This item is based on a combination of “ Job de- mands” and “ Job control.” Both these indexes were dichotomized at the median value (Job de- mands = 10, Job control = 6) and coded into four categories: High-strain jobs ( high job de- mands, low job control), Passive jobs ( low job de- mands, high job control), Active jobs ( high job de- mands, high job control) , and Low -strain jobs ( low job demands, low j ob control). A similar ap- proach using median values has been applied in earlier studies (cf. Jonsson et al., 2013) . Only Job strain was included in the analysis. Categorical dummy variable, Job strain Physical work exposure The variable consists of nine indicators of job characteristics: repeated and one- sided move- ments ( two items), crooked and twisted postures ( two items), becoming sweaty on a daily basis ( two items), shaking and vibrations ( two items), and heavy lifting ( one item). The number of indi- cators summed, ordinal 0– 9. 0 = “ No to all items,” 1 = “ Yes to a sin- gle item,” 2 = “ Y es to two items, ” up to nine items 39 Personal h ealth indicators In Article II, three indicators of poor health ( i.e., disease, illness, and sickness ) w ere in- cluded in the analysis (see Table 4 ) to capture a multidimensional and holistic understand- ing of health. Marinker ’ s (1975 ) tripartite schema of disease, illness , and sickness captures the broader objective, subjective, and contextual aspects of poor health. Disease can be defined as health conditions, diagnosed by a doctor into standardized or specific codes, that often can be seen, touched, measured, or smelled. These conditions are often known and treatable but can sometimes be unspecific due to subjective information and the in- terrelationship between different body parts and organs. For example, illness is a personal, self-reported temporary or permanent mental or physical condition that sometimes may precede disease. Illness i s often a subjective interior feeling or experience that can range from being a minor to a more serious health-related problem for the individual. Finally, sickness is the social role of poor health and the contextual meaning of being “ sick” in the surrounding society; sickness is often related to an individual’ s absence from work. Decision latitude PSAE include s several questions about individuals ’ freedom and influence over their work (Article s I and II) . P articipants were asked to rate their freedom to decide when to carry out their work, how to carry out their work, and what tasks to engage in on a scale ranging from 1 (none) to 4 (great). Subsequently, they were asked to rate their level of influence on working hours, wor king pace, and planning of work on a scale ranging from 1 (none) to 3 (great) (see Table 5 ) . In Article II , the question s were dichotomized in the following way: low influence (no influence, some influence) and high influence (great influence); and low freedom (no, little, and some freedom) and high freedom (great freedom). The dichotomization was conducted to construct typologies based on influence/freedom and health condition. In Article I, the questions regarding freedom at work were used to con- struct an index. 40 Table 5. Questions regarding d ecision latitude in terms of influence and freedom at work. English translation Influence at work Q1 To what extent do you have influence over your working hours? Q2 To what extent do you have influence over your working pace? Q3 To what extent do you have influence over the planning of your work? Freedom at work Q4 How much freedom do you have to decide when your work is to be carried out? Q5 How much freedom do you have to decide what to do? Q6 How much freedom do you have to decide what should be carried out? Idiosyncratic deals and face validity In Article III , negotiated individualized work arrangements ( idiosyncratic deals) were cap- tured using 15 questions covering five areas : task and work responsibilities, workload re- duction, schedule flexibility, location flexibility, and financial incentives . This question- naire originated from the Ex Post I -deals Scale (Rosen et al., 2013) and was translated into Swedish by Mikael Stattin and colleagues at Umeå University . T he original questionnaire was slightly modified to better fit the cultural and contextual elements of Swedish em- ployment relationships and working life ( e.g., superlative wording was toned down) . Also, two questions concerning workload reduction were added to the questionnaire to capture employee– employer negotiation regarding alternative and/or fewer work tasks (s ee Table 6 for the specific question s). Because this was the first time the I -deal instrument was empirically tested in Sweden and in a Nordic context, we conducted a post hoc face validity analysis to examine individual in- terpretations of the I -deals instrument and the response scale. The respondents were liv- ing and working in Sweden. In selecti ng respondents, we used a strategic sampling ap- proach to achieve variety in terms of occupation, gender , and age (above 55 years) ( n = 13). The respondents were asked to (a) complete the questionnaire and (b) describe how they interpreted the questions and wh y they had responded as they did; (c) the respond- ents were finally interviewed about their understanding of the questions and the response scales. Results of this face validity analysis are presented in Article III and summarized in the “Summary of results” section below. 41 Table 6. Idiosyncratic deals questionnaire in English and Swedish . English translation In Swedish Q1 I have agreed with my employer (manager) to be assigned tasks that better match my competence and experiences. Jag har kommit överens med min arbetsgivare (chef) att få arbetsuppgifter som bättre motsva- rar min kompetens och min erfarenhet Q2 I have agreed with my employer (manager) to be assigned tasks that better develop my knowledge. Jag har kommit överens med min arbetsgivare (chef) att få arbetsuppgifter som bättre utveck- lar mina kunskaper. Q3 I have discussed the possibility of being assigned tasks that better match my skills and abilities. with my employer (manager) Jag har diskuterat med min arbetsgivare (chef) om att erhålla arbetsuppgifter som bättre mat- char mina kunskaper och förmågor. Q4 My employer has offered me the opportunity to work with tasks that are outside my usual area of respon- sibility. Min arbetsgivare har erbjudit mig möjligheten att få arbetsuppgifter som li gger utanför mitt egentliga ansvarsområde. Q5 My employer (manager) has granted mea higher degree of flexibility con- cerning the execution of my work tasks. Jag har av min arbetsgivare (chef) fått en högre grad av flexibilitet när det gäller arbetets utfö- rande. Q6 My employer is open to discussing my working conditions based on my individual needs and desires. Min arbetsgivare är öppen för att diskutera mina arbetsvillkor utifrån mina individuella be- hov och önskemål. Q7 My supervisor (manager) and I have agreed that I will be assigned fewer tasks in order to reduce my work- load. Min arbetsledning (chef) och jag har kommit överens om färre arbetsuppgifter för att minska min arbetsbelastning Q8 My employer (manager) and I have agreed to find alternative tasks to re- duce my workload. Min arbetsledning (chef) och jag har kommit överens om att hitta alternativa arbetsuppgifter för att minska min arbetsbelastning Q9 My manager takes into account my personal requests for working hours (schedules) . Min chef tar hänsyn till mina personliga önske- mål om arbetstid (scheman) Q1 0 Based on my wishes, my supervisor (manager) takes into account my private situation (outside of work) when determining my working hours. Utifrån mina önskemål tar min arbetsledning (chef) hänsyn till min privata situation (utanför arbetet) när min arbetstid bestäms. Q1 1 I have permission from my supervi- sor (manager) to take time off for non- work- related matters. Jag har medgivande av min arbetsledning (chef) att ta ledigt för icke - arbetsrelaterade ärenden Q1 2 Based on my individual needs, I can fulfill parts of my tasks outside the workplace (for example, at home), in agreement with my supervisor. Utifrån mina individuella behov kan jag via överenskommelse med min arbetsledning full- göra delar av mina arbetsuppgifter utanför ar- betsplatsen (till exempel i hemmet) 40 Table 5. Questions regarding d ecision latitude in terms of influence and freedom at work. English translation Influence at work Q1 To what extent do you have influence over your working hours? Q2 To what extent do you have influence over your working pace? Q3 To what extent do you have influence over the planning of your work? Freedom at work Q4 How much freedom do you have to decide when your work is to be carried out? Q5 How much freedom do you have to decide what to do? Q6 How much freedom do you have to decide what should be carried out? Idiosyncratic deals and face validity In Article III , negotiated individualized work arrangements ( idiosyncratic deals) were cap- tured using 15 questions covering five areas : task and work responsibilities, workload re- duction, schedule flexibility, location flexibility, and financial incentives . This question- naire originated from the Ex Post I - deals Scale (Rosen et al., 2013) and was translated into Swedish by Mikael Stattin and colleagues at Umeå University . T he original questionnaire was slightly modified to better fit the cultural and contextual elements of Swedish em- ployment relationships and working life ( e.g., superlative wording was toned down) . Also, two questions concerning workload reduction were added to the questionnaire to capture employee– employer negotiation regarding alternative and/or fewer work tasks (s ee Table 6 for the specific question s) . Because this was the first time the I - deal instrument was empirically tested in Sweden and in a Nordic context, we conducted a post hoc face validity analysis to examine individual in- terpretations of the I - deals instrument and the response scale. The respondents were liv- ing and working in Sweden. In selecti ng respondents, we used a strategic sampling ap- proach to achieve variety in terms of occupation, gender , and age (above 55 years) ( n = 13). The respondents were asked to (a) complete the questionnaire and (b) describe how they interpreted the questions and wh y they had responded as they did; (c) the respond- ents were finally interviewed about their understanding of the questions and the response scales. Results of this face validity analysis are presented in Article III and summarized in the “Summary of results” section below. 41 Table 6. Idiosyncratic deals questionnaire in English and Swedish . English translation In Swedish Q1 I have agreed with my employer (manager) to be assigned tasks that better match my competence and experiences. Jag har kommit överens med min arbetsgivare (chef) att få arbetsuppgifter som bättre motsva- rar min kompetens och min erfarenhet Q2 I have agreed with my employer (manager) to be assigned tasks that better develop my knowledge. Jag har kommit överens med min arbetsgivare (chef) att få arbetsuppgifter som bättre utveck- lar mina kunskaper. Q3 I have discussed the possibility of being assigned tasks that better match my skills and abilities. with my employer (manager) Jag har diskuterat med min arbetsgivare (chef) om att erhålla arbetsuppgifter som bättre mat- char mina kunskaper och förmågor. Q4 My employer has offered me the opportunity to work with tasks that are outside my usual area of respon- sibility. Min arbetsgivare har erbjudit mig möjligheten att få arbetsuppgifter som li gger utanför mitt egentliga ansvarsområde. Q5 My employer (manager) has granted mea higher degree of flexibility con- cerning the execution of my work tasks. Jag har av min arbetsgivare (chef) fått en högre grad av flexibilitet när det gäller arbetets utfö- rande. Q6 My employer is open to discussing my working conditions based on my individual needs and desires. Min arbetsgivare är öppen för att diskutera mina arbetsvillkor utifrån mina individuella be- hov och önskemål. Q7 My supervisor (manager) and I have agreed that I will be assigned fewer tasks in order to reduce my work- load. Min arbetsledning (chef) och jag har kommit överens om färre arbetsuppgifter för att minska min arbetsbelastning Q8 My employer (manager) and I have agreed to find alternative tasks to re- duce my workload. Min arbetsledning (chef) och jag har kommit överens om att hitta alternativa arbetsuppgifter för att minska min arbetsbelastning Q9 My manager takes into account my personal requests for working hours (schedules) . Min chef tar hänsyn till mina personliga önske- mål om arbetstid (scheman) Q1 0 Based on my wishes, my supervisor (manager) takes into account my private situation (outside of work) when determining my working hours. Utifrån mina önskemål tar min arbetsledning (chef) hänsyn till min privata situation (utanför arbetet) när min arbetstid bestäms. Q1 1 I have permission from my supervi- sor (manager) to take time off for non- work- related matters. Jag har medgivande av min arbetsledning (chef) att ta ledigt för icke -arbets relaterade ärenden Q1 2 Based on my individual needs, I can fulfill parts of my tasks outside the workplace (for example, at home), in agreement with my supervisor. Utifrån mina individuella behov kan jag via överenskommelse med min arbetsledning full- göra delar av mina arbetsuppgifter utanför ar- betsplatsen (till exempel i hemmet) 42 Q1 3 Because of my special skills, my em- ployer accepts that I work outside my regular workplace. På grund av min speciella kompetens så godtar min arbetsgivare att jag kan arbeta utanför min normala arbetsplats. Q1 4 Based on my special skills, my em- ployer is open to discussing my sal- ary and other benefits. Utifrån min speciella kompetens är min arbets- givare öppen för att diskutera min lön och andra förmåner. Q1 5 My employer has increased my sal- ary because of my good perfor- mance. Min arbetsgivare har höjt min lön på grund av mina goda prestationer Operationalization of covariate variables The multivariate analysis includes several covariates to control for potential confounding effects. Some of the se covariates are presented in Table 5 . Socioeconomic status was op- erationalized in two ways; in Article s I and II , socioeconomic status was operationalized according to the Erikson and Goldthorpe (1992) class schema: higher, managerial, middle managerial, lower managerial, skilled , and unskilled manual. In Article I II , the Swedish Standard Classification of Occupations 2012 (SSYK12) (Statistics Sweden, 2012) was transformed into ISCO08 codes (ILO, 2012) and translated to the (Oesch, 2003; Oesch, 2006) socioeconomic class schema: higher-grade service class, lower-grade service class, skilled workers, and unskilled workers. The main difference between these two class sche- mas is that Erikson and Goldthorpe (1992) operationalized class by means of hierarchical division. In comparison, Oesch (2003) emphasized differences in “ work logics” (i.e., tech- nical vs. organiz ational vs. interpersonal) and developed the schema to capture horizontal and often gendered class cleavages between service, routine sales, and production occu- pations. The short version of the Oesch (2003) class schema was found to differentiate between the service occupations of public welfare employees better than the Erikson and Goldthorpe (1992) class schema. Because the data only cover public welfare employees, the Oesch (2003) class schema was selected for the analysis. Statistical methods In Article s I and II, time to event was examined using Co x proportional hazard regression. Cox regression could accommodate data with right-censored cases , i.e., individuals who were still participating in the labor market at the final data collection point ( in total, 194 individuals out of 1143 in Article s I and II) . Hazard ratios (HRs) with 95% confidence intervals were calculated using the Efron method to handle ties, i.e., two events occurring at the same time (Box -Steffensmeier and Jones, 2004) . A hazard ratio of 1.084 is inter- preted as an 8.4 % higher risk of earlier retirement (1.084 – 1), while a hazard ratio of 0.98 is interpreted as a 2% lower risk of early retirement (1 – 0.98), with all other factors kept constant. 43 For the C ox regression models, a continuous component (i.e., years at risk of retirement) and a dichotomous component (i.e., active/retired) were constructed. The continuous component of the dependent variable is discrete with a 14 - year length (i.e., 2002/2003 to 2015), meaning that the participant could be active at age 77 if s/he was 64 years old at baseline. The dependent variable and its operationalization were primarily intended to capture whether or not individuals were still working, with the relative change (drop) in work- related income relative to pension- related income indicating withdrawal from the labor market. It is quite common for individuals to continue working while simultaneously receiving pension benefits, often due to private pension agreements and the absence of an earnings test in the Swedish pension system (i.e., reduc ed pension benefits depending on income). The price basic amount (PB A) is used as a threshold ; it reflects changes in the general price level and is annually adjusted by Statistics Sweden after adjust ing the consumer price index. When the work- related income drops below the PBA, individuals need income from other sources, i.e., pension - related income, to meet their living ex- penses. The income classification at retirement was evaluated in a recent Swedish study (referred to as “ earnings from labor” ) and shown to be adequate in predicting full - time retirement (Eyj ólfsdóttir et al., 2019b) . In Article III , confirmatory factor analysis (CFA ) and structural equation modeling (SEM) were employed. CFA/SEM has the advantage of including both manifest variables ( e.g., ob- servable/measured via surveys) and latent variables ( i.e., theoretical concepts such as “ job strain” based on manifest indicators). Each latent construct comprises the covariance be- tween a set of manifest indicator variables, which means that “ [they represent] what the indicators share” (Little, 2013) , allowing the more error- free measurement of the theoret- ical concept. In Article III , the analysis focused on the association between five types of idiosyncratic deals and retirement preferences. T his methodological approach has several advantages as it enables the incorporation of both observed and unobserved (latent) var- iables. Also, it allow s the analyst to assess and model measurement errors (Byrne, 2013) . We carried out the analysis in two steps. First, the measurement model was estimated using CFA to establish the factor structure. The n, the measurement model w as expanded by specifying the structural model (SEM). An equality constraint was put on the factor loadings for latent constructs with only two indicators (Little et al., 1999 ) . The quantitative studies were conducted using the statistical software programs SAS 9.1 (Article s I and II ), Stata 15.1 ( Article III) , and Mplus 8.1 (Article III ) . The qualitative study The qualitative study is an interview study with line managers and their HR par tners from six healthcare organizations in Sweden (Article IV) . Before the interview study, the par- ticipants took part in an age- management intervention program, ArbetsKraft (English : 42 Q1 3 Because of my special skills, my em- ployer accepts that I work outside my regular workplace. På grund av min speciella kompetens så godtar min arbetsgivare att jag kan arbeta utanför min normala arbetsplats. Q1 4 Based on my special skills, my em- ployer is open to discussing my sal- ary and other benefits. Utifrån min speciella kompetens är min arbets- givare öppen för att diskutera min lön och andra förmåner. Q1 5 My employer has increased my sal- ary because of my good perfor- mance. Min arbetsgivare har höjt min lön på grund av mina goda prestationer Operationalization of covariate variables The multivariate analysis includes several covariates to control for potential confounding effects. Some of the se covariates are presented in Table 5 . Socioeconomic status was op- erationalized in two ways; in Article s I and II , socioeconomic status was operationalized according to the Erikson and Goldthorpe (1992) class schema: higher, managerial, middle managerial, lower managerial, skilled , and unskilled manual. In Article I II , the Swedish Standard Classification of Occupations 2012 (SSYK12) (Statistics Sweden, 2012) was transformed into ISCO08 codes (ILO, 2012) and translated to the (Oesch, 2003; Oesch, 2006) socioeconomic class schema: higher- grade service class, lower- grade service class, skilled workers, and unskilled workers. The main difference between these two class sche- mas is that Erikson and Goldthorpe (1992) operationalized class by means of hierarchical division. In comparison, Oesch (2003) emphasized differences in “ work logics” (i.e., tech- nical vs. organiz ational vs. interpersonal) and developed the schema to capture horizontal and often gendered class cleavages between service, routine sales, and production occu- pations. The short version of the Oesch (2003) class schema was found to differentiate between the service occupations of public welfare employees better than the Erikson and Goldthorpe (1992) class schema. Because the data only cover public welfare employees, the Oesch (2003) class schema was selected for the analysis. Statistical methods In Article s I and II, time to event was examined using Co x proportional hazard regression. Cox regression could accommodate data with right- censored cases, i.e., individuals who were still participating in the labor market at the final data collection point ( in total, 194 individuals out of 1143 in Article s I and II) . Hazard ratios (HRs) with 95% confidence intervals were calculated using the Efron method to handle ties, i.e., two events occurring at the same time (Box - Steffensmeier and Jones, 2004) . A hazard ratio of 1.084 is inter- preted as an 8.4 % higher risk of earlier retirement (1.084 – 1), while a hazard ratio of 0.98 is interpreted as a 2% lower risk of early retirement (1 – 0.98), with all other factors kept constant. 43 For the C ox regression models, a continuous component (i.e., years at risk of retirement) and a dichotomous component (i.e., active/retired) were constructed. The continuous component of the dependent variable is discrete with a 14 -year length (i.e., 2002/2003 to 2015), meaning that the participant could be active at age 77 if s/he was 64 years old at baseline. The dependent variable and its operationalization were primarily intended to capture whether or not individuals were still working, with the relative change (drop) in work-related income relative to pension-related income indicating withdrawal from the labor market. It is quite common for individuals to continue working while simultaneously receiving pension benefits, often due to private pension agreements and the absence of an earnings test in the Swedish pension system (i.e., reduc ed pension benefits depending on income). The price basic amount (PB A) is used as a threshold ; it reflects changes in the general price level and is annually adjusted by Statistics Sweden after adjust ing the consumer price index. When the work-related income drops below the PBA, individuals need income from other sources, i.e., pension -related income, to meet their living ex- penses. The income classification at retirement was evaluated in a recent Swedish study (referred to as “ earnings from labor” ) and shown to be adequate in predicting full -time retirement (Eyj ólfsdóttir et al., 2019b) . In Article III , confirmatory factor analysis (CFA ) and structural equation modeling (SEM) were employed. CFA/SEM has the advantage of including both manifest variables ( e.g., ob- servable/measured via surveys) and latent variables ( i.e., theoretical concepts such as “job strain” based on manifest indicators). Each latent construct comprises the covariance be- tween a set of manifest indicator variables, which means that “ [they represent] what the indicators share” (Little, 2013) , allowing the more error-free measurement of the theoret- ical concept. In Article III , the analysis focused on the association between five types of idiosyncratic deals and retirement preferences. T his methodological approach has several advantages as it enables the incorporation of both observed and unobserved (latent) var- iables. Also, it allow s the analyst to assess and model measurement errors (Byrne, 2013) . We carried out the analysis in two steps. First, the measurement model was estimated using CFA to establish the factor structure. The n, the measurement model w as expanded by specifying the structural model (SEM). An equality constraint was put on the factor loadings for latent constructs with only two indicators (Little et al., 1999 ) . The quantitative studies were conducted using the statistical software programs SAS 9.1 (Article s I and II ), Stata 15.1 ( Article III) , and Mplus 8.1 (Article III ) . The qualitative study The qualitative study is an interview study with line managers and their HR par tners from six healthcare organizations in Sweden (Article IV) . Before the interview study, the par- ticipants took part in an age- management intervention program, ArbetsKraft (English : 44 workforce), undertaken between 2016 and 2018 in Skåne and the city of Gothenburg. Line managers and HR representatives were selected to be interviewed because imple- menting age management usually requires collaboration between line managers and HR personnel/department s. Li ne managers are responsible for daily operations (e.g., budget, staff, and work environment), while HR personnel support line managers as well as de- velop, implement, and oversee organizational policies (Leisink and Knies, 2011) . Further- more, earlier studies have claimed that the line managers’ knowledge, beliefs, and attitudes toward older workers are critical in maintaining and strengthening their work ability and motivation to continue working (Furunes and Mykletun, 2011; Ilmarinen and Rantanen, 1999) . Using prior contacts and networks in the healthcare sector (i.e., a purposive sam- pling approach) , higher HR functions in municipalities in Skåne and Gothenburg were contacted and asked to identify healthcare and eldercare organizations interested in par- ticipating in the Arbets K raft project. Unfortunately, s everal organizations declined to par- ticipate due to lack of time, staff shortages, and lack of interest. However, all line managers and HR representatives in the intervention program agreed to participate in the follow - up interview study. The objective of the Arbets K raft program was to transfer knowledge and stimulate line managers and their HR representative s to take measures that would increase older work- ers’ opportunities to work until and beyond pensionable age. The program was inspired by an organizational development method (Norbäck et al., 2006) that gathers representa- tives from different organizations to raise common issues and discuss solutions, with in- put from invited experts and under the leadership of experienced researchers. The experts ’ role was to articulate problems, answer questions, give advice, and suggest solutions. The participants in Arbets K raft raised various topics for discussion, such as: health and diag- noses; physical work; environmental factors; psychosocial work environmental factors; working time, work pace, and recovery opportunit ies; personal finance; family and leisure activities; leadership, social inclusion, participation, support, and discrimination ; motiva- tion, stimulating tasks, and work satisfaction; and skills and potential knowledge develop- ment. Six months after the final meeting in the intervention program, between November 2017 and F ebruary 2018, semi-structured interviews were conducted with the line managers and HR partners. The interviews were independent of the intervention itself. All partici- pants but one participated in the follow-up interview. Eight line managers and eleven HR partners at five eldercare organizations and one hospital ward ( see Table 4 ) were inter- viewed face to face using an interview guide with questions concerning current age -man- agement policies, practices, incentives, and motivation s to retain older workers (s ee Ap- pendix A ). The studied eldercare organizations consisted of both nursing homes and homecare centers. Field notes from the intervention program facilitated the formulation of the interview guide. The intervention program was followed by audio-recorded face- to-face interviews lasting an average of one hour and six minutes. For consistency, the 45 same interviewer carried out the interviews at a time and location chosen by the partici- pant. All interviews were transcribed verbatim. Table 7. Location s of participants and organizations in ArbetsK raft. Site Organization Type Line manager ( n) HR ( n) Total Site 1 (Skåne) 1 Eldercare 3 1 4 Site 1 (Skåne) 2 Eldercare 1 2 3 Site 1 (Skåne) 3 Eldercare - 2 2 Site 1 (Skåne) 4 Eldercare - 2 2 Site 2 ( Gothenburg) 5 Hospital ward 1 3 4 Site 2 ( Gothenburg) 6 Eldercare 3 1 4 Total 8 11 19 From voice and transcripts to themes The qualitative thematic analysis was conducted to identify themes in the data, guided by the recommendations of Braun and Clarke (2006) and Nowell et al. (2017) . The transcripts were transferred to the qualitative analysis software N V ivo version 11 (QSR International Pty Ltd) , which facilitated the data management and coding. During the interviews, the interviewer wrote memos, which helped us structure ideas and differentiate and define potential themes. The transcribed interviews were read and reread to gain better insights into the data. Guided by our research question s, our first analytical step started with id en- tifying parts of the interviews concerning the organizational level; the corresponding pas- sages of the transcripts were then coded. In the second step, we examined how the codes fit together to form broader themes. We then summarized the contents of each theme and checked them for coherence, consistency, and uniqueness . For example, the inform- ants described the challenges of (i) making work accommodations, (ii) individual variation in work ability, and (iii) treating everyone equally. Although these three codes, to some extent, have distinct meanings, they relate to an underlying theme of addressing individual needs via universal solutions. Another example is how the codes for (i) top - level lack of interest, (ii) resistance from central HR departments, and (iii) budget restrictions generated the overarching theme of age management “ under the radar. ” Ethical considerations All studies received ethical ap proval. Article s I and I I were approved by the Ethics Com- mittee of the Gothenburg Region (ref: 848 – 14) , which also approved Article III (ref: 2019 – 02934). Article I V was approved by the Regional Ethical Review Board in Lund (ref : 2013/722). 44 workforce), undertaken between 2016 and 2018 in Skåne and the city of Gothenburg. Line managers and HR representatives were selected to be interviewed because imple- menting age management usually requires collaboration between line managers and HR personnel/department s. Li ne managers are responsible for daily operations (e.g., budget, staff, and work environment), while HR personnel support line managers as well as de- velop, implement, and oversee organizational policies (Leisink and Knies, 2011) . Further- more, earlier studies have claimed that the line managers’ knowledge, beliefs, and attitudes toward older workers are critical in maintaining and strengthening their work ability and motivation to continue working (Furunes and Mykletun, 2011; Ilmarinen and Rantanen, 1999) . Using prior contacts and networks in the healthcare sector (i.e., a purposive sam- pling approach) , higher HR functions in municipalities in Skåne and Gothenburg were contacted and asked to identify healthcare and eldercare organizations interested in par- ticipating in the Arbets K raft project. Unfortunately, s everal organizations declined to par- ticipate due to lack of time, staff shortages, and lack of interest. However, all line managers and HR representatives in the intervention program agreed to participate in the follow - up interview study. The objective of the Arbets K raft program was to transfer knowledge and stimulate line managers and their HR representative s to take measures that would increase older work- ers’ opportunities to work until and beyond pensionable age. The program was inspired by an organizational development method (Norbäck et al., 2006) that gathers representa- tives from different organizations to raise common issues and discuss solutions, with in- put from invited experts and under the leadership of experienced researchers. The experts ’ role was to articulate problems, answer questions, give advice, and suggest solutions. The participants in Arbets K raft raised various topics for discussion, such as: health and diag- noses; physical work; environmental factors; psychosocial work environmental factors; working time, work pace, and recovery opportunit ies; personal finance; family and leisure activities; leadership, social inclusion, participation, support, and discrimination ; motiva- tion, stimulating tasks, and work satisfaction; and skills and potential knowledge develop- ment. Six months after the final meeting in the intervention program, between November 2017 and F ebruary 2018, semi- structured interviews were conducted with the line managers and HR partners. The interviews were independent of the intervention itself. All partici- pants but one participated in the follow- up interview. Eight line managers and eleven HR partners at five eldercare organizations and one hospital ward ( see Table 4 ) were inter- viewed face to face using an interview guide with questions concerning current age - man- agement policies, practices, incentives, and motivation s to retain older workers (s ee Ap- pendix A ). The studied eldercare organizations consisted of both nursing homes and homecare centers. Field notes from the intervention program facilitated the formulation of the interview guide. The intervention program was followed by audio- recorded face- to- face interviews lasting an average of one hour and six minutes. For consistency, the 45 same interviewer carried out the interviews at a time and location chosen by the partici- pant. All interviews were transcribed verbatim. Table 7. Location s of participants and organizations in ArbetsK raft. Site Organization Type Line manager ( n) HR ( n) Total Site 1 (Skåne) 1 Eldercare 3 1 4 Site 1 (Skåne) 2 Eldercare 1 2 3 Site 1 (Skåne) 3 Eldercare - 2 2 Site 1 (Skåne) 4 Eldercare - 2 2 Site 2 ( Gothenburg) 5 Hospital ward 1 3 4 Site 2 ( Gothenburg) 6 Eldercare 3 1 4 Total 8 11 19 From voice and transcripts to themes The qualitative thematic analysis was conducted to identify themes in the data, guided by the recommendations of Braun and Clarke (2006) and Nowell et al. (2017) . The transcripts were transferred to the qualitative analysis software N V ivo version 11 (QSR International Pty Ltd) , which facilitated the data management and coding. During the interviews, the interviewer wrote memos, which helped us structure ideas and differentiate and define potential themes. The transcribed interviews were read and reread to gain better insights into the data. Guided by our research question s, our first analytical step started with id en- tifying parts of the interviews concerning the organizational level; the corresponding pas- sages of the transcripts were then coded. In the second step, we examined how the codes fit together to form broader themes. We then summarized the contents of each theme and checked them for coherence, consistency, and uniqueness . For example, the inform- ants described the challenges of (i) making work accommodations, (ii) individual variation in work ability, and (iii) treating everyone equally. Although these three codes, to some extent, have distinct meanings, they relate to an underlying theme of addressing individual needs via universal solutions. Another example is how the codes for (i) top -level lack of interest, (ii) resistance from central HR departments, and (iii) budget restrictions generated the overarching theme of age management “ under the radar. ” Ethical considerations All studies received ethical ap proval. Article s I and I I were approved by the Ethics Com- mittee of the Gothenburg Region (ref: 848 – 14) , which also approved Article III (ref: 2019– 02934). Article I V was approved by the Regional Ethical Review Board in Lund (ref : 2013/722). 46 The studies have been carried out following the Swedish Research Council ’ s ethical guide- lines (Swedish Research Council, 2002) . Participants in all studies, both quantitative and qualitative , were guaranteed anonymity and confidentiality, informed of the aim of the studies, informed of the data collection procedure, and given the option to withdraw from participation without any conseque nces. In the qualitative study, participants signed a for- mal consent on paper. After the interviews, data were anonymized by giving participants fictive names in the transcripts. After the qualitative analysis, data were transferred to a secure depository. In the quantitative studies, all participants ’ identification numbers were replaced with random ID keys. Only researchers connected to the research projects and studies have been granted access to the data. Given the existence of ageism, negative attitudes, and discrimination against older people in general, it is worth considering the potential ethical aspects of studying “ older workers” in the Swedish labor market. The research questions addressed here could reinforce such stereotypes by depicting older people as vulnerable or having particular needs and desires that deviate from those of other age groups in the workforce. At the same t ime, older workers have been described positively in many ways, such as having valuable skills and competence and constituting a vital labor supply to help meet personnel needs in several sectors. Because research findings usually have a limited audience, it is difficult to assess whether the research questions could fuel ageism in society. An alternative approach would be to use broader concepts such as “ life-course management ” that do not single out older workers, but instead address all age groups (Toppinen -Tanner and Vuori, 2018) . 47 6 Summary of results In this section , the findings of the four empirical studies are summarized. The first three quantitative studies focus on personal and work- related conditions that shorten or lengthen older workers’ employment participation. In the fourth and final qualitative study, line managers and their HR partners are interviewed to examine public welfare organizations’ barriers to retaining older healthcare workers and opportunities to do so. Article I: Personal and work- related determinants of retirement behavior Jonsson, R. (2021) Preferred versus actual retirement and their personal and work - related predictors: A comparison using representative Swedish survey and register data Submitted manuscript The objective of the first study was to examine the association between personal and work- related factors and two outcome measures, retirement preferences and actual retire- ment. First, s elf- reported retirement preferences identified from survey data were com- pared with actual retirement behavior defined by a drop in work- related income obtained from register data. Second, t his study also referred to past efforts to identify conditions that hinder older workers from continuing to work or enable them to do so. A total of 20 factors originating from people’ s work and private life were included in the comparison. The analysis was based on the Panel Survey of Aging and the E lderly (PSAE) conducted in 2002– 2003 , combined with income information, 2002– 2015 , from the Longitudinal Integration Database for Health Insurance and Labour Market Studies (LISA). Retirement preferences and predictors were derived from the baseline survey, while actual retirement was measured using income activity from register- based data. The empirical analysis was conducted using linear regression and C ox proportional hazard regression and was ad- just ed for demographic variables. Findings suggest that retirement preferences were an important determinant of actual re- tirement behavior, although people tended to retire later than initially preferred. The in- crease in people’ s retirement age over the collection period can partially be explained by policy changes, structural changes in the labor market, and improvements in population health. Moreover, factors at the personal level (e.g., health and work ability) and work- related factors (e.g., physical work exposure and job satisfaction ) were related to both 46 The studies have been carried out following the Swedish Research Council ’ s ethical guide- lines (Swedish Research Council, 2002) . Participants in all studies, both quantitative and qualitative , were guaranteed anonymity and confidentiality, informed of the aim of the studies, informed of the data collection procedure, and given the option to withdraw from participation without any conseque nces. In the qualitative study, participants signed a for- mal consent on paper. After the interviews, data were anonymized by giving participants fictive names in the transcripts. After the qualitative analysis, data were transferred to a secure depository. In the quantitative studies, all participants ’ identification numbers were replaced with random ID keys. Only researchers connected to the research projects and studies have been granted access to the data. Given the existence of ageism, negative attitudes, and discrimination against older people in general, it is worth considering the potential ethical aspects of studying “ older workers” in the Swedish labor market. The research questions addressed here could reinforce such stereotypes by depicting older people as vulnerable or having particular needs and desires that deviate from those of other age groups in the workforce. At the same t ime, older workers have been described positively in many ways, such as having valuable skills and competence and constituting a vital labor supply to help meet personnel needs in several sectors. Because research findings usually have a limited audience, it is difficult to assess whether the research questions could fuel ageism in society. An alternative approach would be to use broader concepts such as “ life- course management” that do not single out older workers, but instead address all age groups (Toppinen - Tanner and Vuori, 2018) . 47 6 Summary of results In this section , the findings of the four empirical studies are summarized. The first three quantitative studies focus on personal and work-related conditions that shorten or lengthen older workers’ employment participation. In the fourth and final qualitative study, line managers and their HR partners are interviewed to examine public welfare organizations’ barriers to retaining older healthcare workers and opportunities to do so. Article I: Personal and work- related determinants of retirement behavior Jonsson, R. (2021) Preferred versus actual retirement and their personal and work -related predictors: A comparison using representative Swedish survey and register data Submitted manuscript The objective of the first study was to examine the association between personal and work-related factors and two outcome measures, retirement preferences and actual retire- ment. First, s elf-reported retirement preferences identified from survey data were com- pared with actual retirement behavior defined by a drop in work-related income obtained from register data. Second, t his study also referred to past efforts to identify conditions that hinder older workers from continuing to work or enable them to do so. A total of 20 factors originating from people’ s work and private life were included in the comparison. The analysis was based on the Panel Survey of Aging and the E lderly (PSAE) conducted in 2002– 2003 , combined with income information, 2002– 2015 , from the Longitudinal Integration Database for Health Insurance and Labour Market Studies (LISA). Retirement preferences and predictors were derived from the baseline survey, while actual retirement was measured using income activity from register-based d ata. The empirical analysis was conducted using linear regression and C ox proportional hazard regression and was ad- just ed for demographic variables. Findings suggest that retirement preferences were an important determinant of actual re- tirement behavior, although people tended to retire later than initially preferred. The in- crease in people’ s retirement age over the collection period can partially be explained by policy changes, structural changes in the labor market, and improvements in population health. Moreover, factors at the personal level (e.g., health and work ability) and work- related factors (e.g., physical work exposure and job satisfaction ) were related to both 48 retirement preferences and actual retirement. However, social support and recognition were only relevant to retirement preferences. These findings offer some guidance to pol- icymakers and employers, who could, for example, promote health and work ability, re- duce physical work exposure, and strengthen job satisfaction. Also, they highlight meth- odological discrepancies between different operationalizations of retirement and show that, by using retirement preferences instead of actual retirement, researchers risk overes- timating the impact of specific retirement predictors, particular ly the relationship between social factors at work and actual retirement. Table 8. Additional statistics from the PSA E dataset , including arguments made in the “D iscussion” section. Decision l atitude* n Influence index , mean Freedom index , mean Socioeconomic status Higher managerial 194 7.8 6.4 Middle managerial 308 7.4 5.8 Lower managerial 209 7.2 5.7 Skilled manual 169 6.6 4.9 Unskilled manual 263 6.2 4.7 Sex Male 537 7.2 5.7 Female 606 6.9 5.3 * The indices are based on the summation of the variables (influence and freedom) presented in Table 6 in the “ Data and methods” section. Article II: Work accommodation to support older workers with poor health Jonsson, R. , Dellve, L. , and Halleröd, B. (2019) Work despite poor health? A 14 -year follow-up of how individual work accommodations are extending the time to retirement for workers with poor health conditions. S S M Population Health Journal of Sampl es , Vol. 9, Article No. 1005141. I mprovements in population health are one of the drivers of increasing longevity and population aging. Still, aging increases the risks of poor health and reduced work ability, which are factors known to be related to people’ s early retirement behavior. Mor eover, given the goals of extended working lives, a considerable share of the workforce is ex- pected to work while suffering from chronic disease. The second article investigates the role of health in people’ s retirement behavior using Marinker (1975) tripartite schema of disease, illness, and sickness to capture the multidimensional aspects of health. In addi- tion, it examines whether decision latitude can moderate the cons equences of poor health conditions. High decision latitude allows older workers to make permanent or temporary work accommodations in line with their preferences or abilities. Such work accommoda- tions refer to an individual’ s ability to influence work hours, pace, and plan ning, and to 49 the freedom to determine when, how, and what work tasks are pursued , which is particu- larly important for this group of older workers. The analysis was based on PSAE , 2002– 2003 , combined with income information from LISA , 2002– 2015 . Retirement preferences and predictors were derived from the baseline survey, while actual retirement was measured by income activity from register - based data. Cox proportional hazard regression was employed, and regression models were adjusted for demographic and work- related covariates. Results indicate that disease, illness, and sickness increase d the risk of actual retirement, and that poor health in terms of disease and sickness in combination with low decision latitude increased the risk of early retirement. However, the associations were not equally strong in the case of illness. This study is aligned with prior studies in advocating the importance of health conditions and decision latitude in understanding individual retire- ment behavior. Article III: The importance of skill matching for retirement preferences Jonsson, R. , Hasselgren, C. , Dellve, L. , Seldén, D. , Larsson, D. , and Stattin, M. (2020) Matching the pieces: The presence of idiosyncratic deals and their impact on retirement preferences among older workers. Work, Aging and Retirem ent , Vol. 7 , No. 3 , pp. 240 – 255. 5 This study set out to examine individualized work arrangements — so- called idiosyncratic deals (I - deals) negotiated within the employee – employer relationship. Individual ized work arrangements were assumed to enhance the person– job fit and support older workers in extending their working lives. Five distinct domains of I - deals and their associations with retirement preferences were in focus: task and work responsibilities, workload reduction, schedule flexibility, location flexibility, and financial incentives . This article u ses cross- sectional data from the Hearts – Lexliv study covering 4499 partici- pants aged 55 years or older employed in the city of Gothenburg. I - deal domains were captured using a translated and adapted questionnaire. Confirmatory factor analysis (CFA ) and structural equation modeling (SEM) were employed to examine the association be- tween the I - deal dimensions and retirement preferences. After the analysi s, a post hoc face validity analysis was conducted to explore individual interpretations of the I - deal questionnaire (see the section “M ethodological considerations”) . Results indicate that I - deals are less prevalent among employees who occupy lower soci- oeconomic positions, experien ce poorer health, are female, are older and have shorter organizational tenure. Task and work responsibilities were found to be related to later 5 Robin Jonsson and Caroline Hasselgren are joint first authors. 48 retirement preferences and actual retirement. However, social support and recognition were only relevant to retirement preferences. These findings offer some guidance to pol- icymakers and employers, who could, for example, promote health and work ability, re- duce physical work exposure, and strengthen job satisfaction. Also, they highlight meth- odological discrepancies between different operationalizations of retirement and show that, by using retirement preferences instead of actual retirement, researchers risk overes- timating the impact of specific retirement predictors, particular ly the relationship between social factors at work and actual retirement. Table 8. Additional statistics from the PSA E dataset , including arguments made in the “D iscussion” section. Decision l atitude* n Influence index , mean Freedom index , mean Socioeconomic status Higher managerial 194 7.8 6.4 Middle managerial 308 7.4 5.8 Lower managerial 209 7.2 5.7 Skilled manual 169 6.6 4.9 Unskilled manual 263 6.2 4.7 Sex Male 537 7.2 5.7 Female 606 6.9 5.3 * The indices are based on the summation of the variables (influence and freedom) presented in Table 6 in the “ Data and methods” section. Article II: Work accommodation to support older workers with poor health Jonsson, R. , Dellve, L. , and Halleröd, B. (2019) Work despite poor health? A 14 - year follow- up of how individual work accommodations are extending the time to retirement for workers with poor health conditions. S S M Population Health Journal of Sampl es , Vol. 9, Article No. 1005141. I mprovements in population health are one of the drivers of increasing longevity and population aging. Still, aging increases the risks of poor health and reduced work ability, which are factors known to be related to people’ s early retirement behavior. Mor eover, given the goals of extended working lives, a considerable share of the workforce is ex- pected to work while suffering from chronic disease. The second article investigates the role of health in people’ s retirement behavior using Marinker (1975) tripartite schema of disease, illness, and sickness to capture the multidimensional aspects of health. In addi- tion, it examines whether decision latitude can moderate the cons equences of poor health conditions. High decision latitude allows older workers to make permanent or temporary work accommodations in line with their preferences or abilities. Such work accommoda- tions refer to an individual’ s ability to influence work hours, pace, and plan ning, and to 49 the freedom to determine when, how, and what work tasks are pursued , which is particu- larly important for this group of older workers. The analysis was based on PSAE , 2002– 2003, combined with income information from LISA , 2002– 2015 . Retirement preferences and predictors were derived from the baseline survey, while actual retirement was measured by income activity from register -based data. Cox proportional hazard regression was employed, and regression models were adjusted for demographic and work-related covariates. Results indicate that disease, illness, and sickness increase d the risk of actual retirement, and that poor health in terms of disease and sickness in combination with low decision latitude increased the risk of early retirement. However, the associations were not equally strong in the case of illness. This study is aligned with prior studies in advocating the importance of health conditions and decision latitude in understanding individual retire- ment behavior. Article III: The importance of skill matching for retirement preferences Jonsson, R. , Hasselgren, C. , Dellve, L. , Seldén, D. , Larsson, D. , and Stattin, M. (2020) Matching the pieces: The presence of idiosyncratic deals and their impact on retirement preferences among older workers. Work, Aging and Retirem ent , Vol. 7 , No. 3 , pp. 240– 255. 5 This study set out to examine individualized work arrangements —so -called idiosyncratic deals (I -deals) negotiated within the employee – employer relationship. Individual ized work arrangements were assumed to enhance the person– job fit and support older workers in extending their working lives. Five distinct domains of I -deals and their association s with retirement preferences were in focus: task and work responsibilities, workload reduction, schedule flexibility, location flexibility, and financial incentives . This article u ses cross- sectional data from the Hearts – Lexliv study covering 4499 partici- pants aged 55 years or older employed in the city of Gothenburg. I -deal do mains were captured using a translated and adapted questionnaire. Confirmatory factor analysis (CFA ) and structural equation modeling (SEM) were employed to examine the association be- tween the I -deal dimensions and retirement preferences. After the analysi s, a post hoc face validity analysis was conducted to explore individual interpretations of the I - deal questionnaire (see the section “M ethodological considerations”) . Results indicate that I -deals are less prevalent among employees who occupy lower soci- oeconomic positions, experien ce poorer health, are female, are older and have shorter organizational tenure. Task and work responsibilities were found to be related to later 5 Robin Jonsson and Caroline Hasselgren are joint first authors. 50 retirement age. In contrast , opportunities to negotiate reduced workload were found to be related to preferences for earlier retirement, probably because individuals who received workload reductions also reported poor health. This finding suggests that there might be room for additional interventions beyond workload reduction to retain older public-sector employees with health impairments or disabilities. To sum up, a good match between work and individuals’ competence, experience, and growth opportunities is important for public-sector employees ’ retirement preferences. These compon ents may be arranged out- side regular HRM policies or negotiated within the employment relationship. Another important implication of this study is that contextual circumstances regarding collective agreements and strategic HR policies may hinder negotiated individualized work arrange- ments. Article IV: An organizational perspective on the retention of older workers Jonsson, R. , Lindegård, A. , Bj örk, L. , and Nilsson, K. (2020) Organizational hindrances to the retention of older healthcare workers. Nordic Journal of Working Life S tudies , Vol. 10, No. 1, pp. 41– 58. In the final study, the obstacles and opportunities in the Swedish healthcare sector affect- ing the retention and recruitment of older healthcare employees were investigated based on interviews with line managers and their HR partners. The Swedish health care sector faces current and forecasted future recruitment shortfalls, and older health care workers constitute a potential labor reserve. Interviews were carried out with participants previously engaged in the Arbets K raft age- management intervention program between 2017 and 2018. In total, eight line managers and 11 HR representatives in the health care sector ( i.e., eldercare and hospital wards) from two locations in Sweden were asked about existing age-management policies, practices, incentives, and motivations to retain older workers. The interviews were audio-recorded and transcribed verbatim; the data were then subjected to a thematic analysis. 51 Figure 6. Organizational hindrance s to the retention of older healthcare workers: a model developed in Article IV . Results indicate overall positive attitudes among the participants toward older healthcare workers, especially in times of scarce supply of per sonnel, and they are regarded as valu- able for their knowledge and role as mentors to younger employees. The term “ age man- agement” was not part of the line managers’ and their HR partners ’ vocabulary or termi- nology. 6 This indicates a potential discrepancy between the academic debate on age man- agement and its practical application in the labor market, at least in the examined organi- zations. With regard to maintai ning and retaining older healthcare workers, no formalize d age- related HR practices were found in the examined organizations. Despite negotiations between older employees and line managers, there were no indications of any formalized or systematic negotiations at a larger scale. Several hindering factors were identified and seemed to reinforce one another, leading to a vicious circle and hindering lower managers from focusing on retaining older healthcare workers (see flowchart in Figure 4 to under- stand how the organizational hindrances are related). The overall labor shortage led to 6 It is worth pointing out that, since the data collection of Article IV, several Swedish municipalities have imple- mented the so- called “80 - 90 - 100” model (or modified forms thereof). Th is model offers welfare workers a re- duction of working time of 20%, 90% of the ir earnings, and 100% pension pa yments. In Article IV, the 80 - 90 - 100 model was tested in one of the organizations on a limited number of employees. However, since then, the number of Swedish municipalities implementing the model has increased rapidly. 50 retirement age. In contrast , opportunities to negotiate reduced workload were found to be related to preferences for earlier retirement, probably because individuals who received workload reductions also reported poor health. This finding suggests that there might be room for additional interventions beyond workload reduction to retain older public- sector employees with health impairments or disabilities. To sum up, a good match between work and individuals’ competence, experience, and growth opportunities is important for public- sector employees’ retirement preferences. These compon ents may be arranged out- side regular HRM policies or negotiated within the employment relationship. Another important implication of this study is that contextual circumstances regarding collective agreements and strategic HR policies may hinder negotiated individualized work arrange- ments. Article IV: An organizational perspective on the retention of older workers Jonsson, R. , Lindegård, A. , Bj örk, L. , and Nilsson, K. (2020) Organizational hindrances to the retention of older healthcare workers. Nordic Journal of Working Life S tudies , Vol. 10, No. 1, pp. 41 – 58 . In the final study, the obstacles and opportunities in the Swedish healthcare sector affect- ing the retention and recruitment of older healthcare employees were investigated based on interviews with line managers and their HR partners. The Swedish health care sector faces current and forecasted future recruitment shortfalls, and older health care workers constitute a potential labor reserve. Interviews were car ried out with participants previously engaged in the Arbets K raft age- management intervention program between 2017 and 2018. In total, eight line managers and 11 H R representatives in the healthcare sector ( i.e., eldercare and hospital wards) from two locations in Sweden were asked about existing age- management policies, practices, incentives, and motivations to retain older workers. The interviews were audio- recorded and transcribed verbatim; the data were then subjected to a thematic analysis. 51 Figure 6. Organizational hindrance s to the retention of older healthcare workers: a model developed in Article IV . Results indicate overall positive attitudes among the participants toward older healthcare workers, especially in times of scarce supply of per sonnel, and they are regarded as valu- able for their knowledge and role as mentors to younger employees. The term “ age man- agement” was not part of the line managers’ and their HR partners ’ vocabulary or termi- nology. 6 This indicates a potential discrepancy between the academic debate on age man- agement and its practical application in the labor market, at least in the examined organi- zations. With regard to maintai ning and retaining older healthcare workers, no formalize d age-related HR practices were found in the examined organizations. Despite negotiations between older employees and line managers, there were no indications of any formalized or systematic negotiations at a larger scale. Several hindering factors were identified and seemed to reinforce one another, leading to a vicious circle and hindering lower managers from focusing on retaining older healthcare workers (see flowchart in Figure 4 to under- stand how the organizational hindrances are related). The overall labor shortage led to 6 It is worth pointing out that, since the data collection of Article IV, several Swedish municipalities have imple- mented the so-called “80 -90-100” model (or modified forms thereof). Th is model offers welfare workers a re- duction of working time of 20%, 90% of the ir earnings, and 100% pension pa yments. In Article IV, the 80 -90- 100 model was tested in one of the organizations on a limited number of employees. However, since then, the number of Swedish municipalities implementing the model has increased rapidly. 52 time-consuming recruitment and administrative activities, which contributed to a high workload, particularly for the line managers. This strained work situation force d the man- agers to prioritize day-to -day activities over strategic pl anning. For instance, as the reten- tion of older workers in the organization is paid insufficient attention, no age -management strategies are developed. In the absence of such strategies, managers rely on universal policies that do not provide specific tools for managing older employees. Without such tools, the organization cannot offer sufficient incentives or individualized adaptations to its older workers, exacerbating the labor shortage. Line -managers’ high workloads, inflex- ible HR strategies, and lack o f elaborated age-management strategies and coordination within the organization limit the public welfare employers’ opportunities to gain from the potential labor reserve older employees constitute. 53 7 Discussion This thesis aimed to examine how older workers ’ decisions to prolong or end their work- ing lives are formed through interaction between, on one han d, personal and work - related conditions and, on the other, organizations ’ capabilities to implement age- management practices. Using a capability approach as a theoretical framework, conditions that influ- ence individuals’ capabilities to work and organizations’ capabilities to implement age- management measures were studied both separately and in combination. Below, I return to the thesis’ s main research questions and present the empirical findings in relation to previous research. Personal and work -related determinants of workers older workers’ ability and willing- ness to continue working In line with previous research (Browne et al., 2019; Edge et al., 2017; Fisher et al., 2016) , findings reported in Article I show that poor health and high physical work exposure push older workers toward early retirement, while job satisfaction and high work ability keep older workers in the workforce (Article I). P oor health conditions in terms of disease, illness, and sickness were found to reduce individuals’ physical and psychological work ability and increase the risk of earlier actual retirement (Article II) . One assumption is that higher decision latitude might permit self- management, giving older workers opportuni- ties to adapt their work in accordance with their abilities and preferences. Findings indi- cate that poor health conditions in combination with few opportunities to accommodate work through offering decision latitude increased the risk and likelihood of earlier actual retirement. However, these associations were not as strong in the case of illness (Article I I) . Another assumption was that employees’ opportunities to negotiate individualized work arrangements, i.e., idiosyncratic deals, may strengthen individuals’ capability and the person– job fit . Findings show that skill matching is of particular importance, i.e., idiosyn- cratic deals concerning employees’ tasks and responsibilities were related to public- sector employees’ later preferred retirement age (Article III). To sum up, o lder workers’ access to high decision latitude moderated poor health, and opportunities to negotiate individu- alized work arrangements concerning task and work responsibilities helped keep older workers in employment. Taken together, this indicates that the importance of certain fac- 52 time- consuming recruitment and administrative activities, which contributed to a high workload, particularly for the line managers. This strained work situation force d the man- agers to prioritize day- to- day activities over strategic planning. For instance, as the reten- tion of older workers in the organization is paid insufficient attention, no age - management strategies are developed. In the absence of such strategies, managers rely on universal policies that do not provide specific tools for managing older employees. Without such tools, the organization cannot offer sufficient incentives or individualized adaptations to its older workers, exacerbating the labor shortage. Line - managers’ high workloads, inflex- ible HR strategies, and lack o f elaborated age- management strategies and coordination within the organization limit the public welfare employers’ opportunities to gain from the potential labor reserve older employees constitute. 53 7 Discussion This thesis aimed to examine how older workers ’ decisions to prolong or end their work- ing lives are formed through interaction between, on one han d, personal and work -related conditions and, on the other, organizations ’ capabilities to implement age-management practices. Using a capability approach as a theoretical framework, conditions that influ- ence individuals’ capabilities to work and organizations’ capabilities to implement age- management measures were studied both separately and in combination. Below, I return to the thesis’ s main research questions and present the empirical findings in relation to previous research. Personal and work -related determinants of workers older workers’ ability and willing- ness to continue working In line with previous research (Browne et al., 2019; Edge et al., 2017; Fisher et al., 2016) , findings reported in Article I show that poor health and high physical work exposure push older workers toward early retirement, while job satisfaction and high work ability keep older workers in the workforce (Article I). P oor health conditions in terms of disease, illness, and sickness were found to reduce individuals’ physical and psychological work ability and increase the risk of earlier actual retirement (Article II) . One assumption is that higher decision latitude might permit self-manag ement, giving older workers opportuni- ties to adapt their work in accordance with their abilities and preferences. Findings indi- cate that poor health conditions in combination with few opportunities to accommodate work through offering decision latitude increased the risk and likelihood of earlier actual retirement. However, these associations were not as strong in the case of illness (Article I I) . Another assumption was that employees’ opportunities to negotiate individualized work arrangements, i.e., idiosyncratic deals, may strengthen individuals’ capability and the person– job fit . Findings show that skill matching is of particular importance, i.e., idiosyn- cratic deals concerning employees’ tasks and responsibilities were related to public-sector employees’ later preferred retirement age (Article III). To sum up, o lder workers’ access to high decision latitude moderated poor health, and opportunities to negotiate individu- alized work arrangements concerning task and work responsibilities helped keep older workers in employment. Taken together, this indicates that the importance of certain fac- 54 tors for individuals’ capabilities to continue working is partly context dependent and de- termined in relation to workplace-specific conditions (i.e., organizational capability). I re- turn to this argument below. It is worth noting that PWL policies are known to have divergent effects among certain sociodemographic groups on the labor market differing in, for example, gender, health, and socioeconomic position (Kadefors et al., 2018b; König et al., 2018; McAllister et al., 2020) . For instance, early actual retirement behavior was generally found among women, older age groups, and individuals with lower educational attainment and/or lower socio- economic positions (Article s I and II) . Similar relationships were found in Article III, in which women, individuals with lower socioeconomic position s, and individuals with poorer health reported preferences for earlier retirement. The gender segregation in the Swedish labor market is an underlying factor influencing people’ s retirement decisions. It affects men’s and women’ s earnings, physical and psychosocial exposure s at work, work environments, and opportunities to attain good health over the life course. The internal segregation in public welfare organizations can partly explain the differences in retirement preferences between men and women. For instance, the female-dominate d occupations (e.g., in healthcare, eldercare, and school) in the studied public welfare organizations ap- pear to have earlier retirement patterns than do the male-dominate d occupations (e.g., in technical operations) (Seldén et al., 2020) . Findings of Article III showed that the oppor- tunities to negotiate individualized work arrangements were lower among women, older workers in lower socioeconomic positions, those experiencing poor health, and those with shorter organizational tenure. In addition, decision latitude appears to be lower among people in lower socioeconomic positions and among women in the general population (see additional table in the “ Results” section of Article II). The availability of flexible and individualized work arrangements, however, is not solely dependent on individual factors ; rather, the se arrangements are also known to be heavily influenced by factors such as job type—for example, some work tasks require physical presence and/or must be completed during specific hours (Damman and Henkens, 2018; Hornung et al., 2009) . In light of these results, the uneven distribution of work-related resources such as idiosyncratic deals or decision latitude could give rise to unequal opportunities to prolong working lives. Public welfare organizations ’ (in)capability to retain older workers The findings of the qualitative study, reported in Article IV, suggest that multiple chal- lenges can reinforce one another in an organization, similar to a vicious circle . In the case of Swedish healthcare, empirical results suggest that organizations’ capability to retain older workers is hindered by line managers’ high workload, inflexible HR strategies, ab- sence of elaborated age-management strategies , and lack of coordination within the or- ganization. These are barriers to organizations’ capability to manage interpersonal varia- tion in abilities and preferences among older workers. Because a chain is no stronger than 55 its weakest link, implement ing age- management policies requires th at different organiza- tional functions be coordinated, in t erms of both prioritizing (i.e., preferences) and allo- cating resources (i.e., adequate resources to achieve goals). A notable theoretical and em- pirical finding was that an organization’ s implementation of age management is not ex- clusively determined by its preferences or resources, as claimed by Kadefors et al. (2020) . Inst ead, the implementation of age- management strategies also requires that different lev- els within the organization be coordinated. Drawing on Kadefors et al. (2020) typology of organizational capability, the examined health care organizations could be categorized as both passive with respect to the strategic HR function and chained with respect to their line managers’ limited resources. The lack of coordination demonstrates that organizational capability is the product of the combined operations extending from top management to the street level, where line managers perform da ily work. The potential consequence of the lack of coordination and the absence of “ age- management” strategies is that the re- tention of older healthcare workers is suboptimal. Here, I put age management within quotation marks because none of the participants interviewed for Article IV used the term “ age management” or described such measures being applied where they worked. This points to a potential discrepancy between the academic and societal debates about age management and its practical application in the labor market. Evidence of this discrepancy has indirectly been manifested by the fact that employers seem to take a passive or re- stricted approach to such measures (Conen et al., 2012; Fuertes et al., 2 013; Jensen, 2020; Principi et al., 2020; Vickerstaff et al., 2003) . The Swedish health care sector’ s lack of age awareness is a potential push factor at the individual level that affects both workers’ abil- ities and motivation to continue employment. Also, it counteracts the key national goal of the Swedish Association of Local Authorities and Regions of retaining older workers (SALAR, 2020) . The absence of organizational policies has led to the responsibility for retaining older workers being transferred to lower managers on the line. Earlier research in Sweden has found similar developments in public welfare organizations in which line managers encounter goal conflict and are torn between the organization’ s goals and what the employees or the operation require (Corin and Björk, 2 016; Kankkunen et al., 2014) . What could enhance the Swedish healthcare sector’ s capability to retain older workers, making it shift from a passive to active approach? Here, the strategic HR function would need to articulate the necessity and priority to higher decision- makers so that sufficient resources are allocated and line managers are provided with appropriate tools, instruc- tions, and authority to take the necessary actions. However, higher managers in the healthcare sector often face complex decisions pertaining to balancing budgets, maintain- ing operations, and accomplishing contradictory organizational goals. Therefore, it might be necessary for strategic HR to provide decision guid ance material that discloses the potential costs and benefits of implementing age- management practices. One illustrative example of such decision guidance material is the cost– benefit analysis carried out by Malmqvist (2007 ) , who evaluated two approaches to age management and identified some practical applications. A strategic organizational approach to the older workforce will 54 tors for individuals’ capabilities to continue working is partly context dependent and de- termined in relation to workplace- specific conditions (i.e., organizational capability). I re- turn to this argument below. It is worth noting that PWL policies are known to have divergent effects among certain sociodemographic groups on the labor market differing in, for example, gender, health, and socioeconomic position (Kadefors et al., 2018b; König et al., 2018; McAllister et al., 2020) . For instance, early actual retirement behavior was generally found among women, older age groups, and individuals with lower educational attainment and/or lower socio- economic positions (Article s I and II) . Similar relationships were found in Article III, in which women, individuals with lower socioeconomic position s, and individuals with poorer health reported preferences for earlier retirement. The gender segregation in the Swedish labor market is an underlying factor influencing people’ s retirement decisions. It affects men’s and women’ s earnings, physical and psychosocial exposure s at work, work environments, and opportunities to attain good health over the life course. The internal segregation in public welfare organizations can partly explain the differences in retirement preferences between men and women. For instance, the female- dominated occupations (e.g., in healthcare, eldercare, and school) in the studied public welfare organizations ap- pear to have earlier retirement patterns than do the male- dominated occupations (e.g., in technical operations) (Seldén et al., 2020) . Findings of Article III showed that the oppor- tunities to negotiate individualized work arrangements were lower among women, older workers in lower socioeconomic positions, those experiencing poor health, and those with shorter organizational tenure. In addition, decision latitude appears to be lower among people in lower socioeconomic positions and among women in the general population (see additional table in the “ Results” section of Article II). The availability of flexible and individualized work arrangements, however, is not solely dependent on individual factors ; rather, the se arrangements are also known to be heavily influenced by factors such as job type— for example, some work tasks require physical presence and/or must be completed during specific hours (Damman and Henkens, 2018; Hornung et al., 2009) . In light of these results, the uneven distribution of work- related resources such as idiosyncratic deals or decision latitude could give rise to unequal opportunities to prolong working lives. Public welfare organizations ’ (in)capability to retain older workers The findings of the qualitative study, reported in Article IV, suggest that multiple chal- lenges can reinforce one another in an organization, similar to a vicious circle . In the case of Swedish healthcare, empirical results suggest that organizations’ capability to retain older workers is hindered by line managers’ high workload, inflexible HR strategies, ab- sence of elaborated age- management strategies, and lack of coordination within the or- ganization. These are barriers to organizations’ capability to manage interpersonal varia- tion in abilities and preferences among older workers. Because a chain is no stronger than 55 its weakest link, implement ing age-management policies requires th at different organiza- tional functions be coordinated, in t erms of both prioritizing (i.e., preferences) and allo- cating resources (i.e., adequate resources to achieve goals). A notable theoretical and em- pirical finding was that an organization’s implementation of age management is not ex- clusively determined by its preferences or resources, as claimed by Kadefors et al. (2020) . Inst ead, the implementation of age-management strategies also requires that different lev- els within the organization be coordinated. Drawing on Kadefors et al. (2020) typology of organizational capability, the examined health care organizations could be categorized as both passive with respect to the strategic HR function and chained with respect to their line managers’ limited resources. The lack of coordination demonstrates that organizational capability is the product of the combined operations extending from top management to the street level, where line managers perform da ily work. The potential consequence of the lack of coordination and the absence of “ age-management ” strategies is that the re- tention of older healthcare workers is suboptimal. Here, I put age management within quotation marks because none of the participants interviewed for Article IV used the term “ age management” or described such measures being applied where they worked. This points to a potential discrepancy between the academic and societal debates about age management and its practical application in the labor market. Evidence of this discrepancy has indirectly been manifested by the fact that employers seem to take a passive or re- stricted approach to such measures (Conen et al., 2012; Fuertes et al., 2 013; Jensen, 2020; Principi et al., 2020; Vickerstaff et al., 2003) . The Swedish health care sector’ s lack of age awareness is a potential push factor at the individual level that affects both workers’ abil- ities and motivation to continue employment. Also, it counteracts the key national goal of the Swedish Association of Local Authorities and Regions of retaining older workers (SALAR, 2020) . The absence of organizational policies has led to the responsibility for retaining older workers being transferred to lower managers on the line. Earlier research in Sweden has found similar developments in public welfare organizations in which line managers encounter goal conflict and are torn between the organization’ s goals and what the employees or the operation require (Corin and Björk, 2 016; Kankkunen et al., 2014) . What could enhance the Swedish healthcare sector’ s capability to retain older workers, making it shift from a passive to active approach? Here, the strategic HR function would need to articulate the necessity and priority to higher decision-makers so that sufficient resources are allocated and line managers are provided with appropriate tools, instruc- tions, and authority to take the necessary actions. However, higher managers in the healthcare sector often face complex decisions pertaining to balancing budgets, maintain- ing operations, and accomplishing contradictory organizational goals. Therefore, it might be necessary for strategic HR to provide decision guid ance material that discloses the potential costs and benefits of implementing age-management practices. One illustrative example of such decision guidance material is the cost– benefit analysis carried out by Malmqvist (2007 ) , who evaluated two approaches to age management and identified some practical applications. A strategic organizational approach to the older workforce will 56 likely produce more sustainable working conditions and counteract negative attitudes. B e- fore investing in age-management policies, a starting point is for top management to en- lighten and signal to the managers and their employees that prolonged labor market par- ticipation is a prioritized and desirable goal. Also, line managers must have sufficient time for day-to -day conversations with their employees to identify personal or general prob- lems in the work environment. It is also critical that the implementation of age -manage- ment policies be supported by co-workers and be considered fair and legitimate. This appears crucial in public welfare organizations in which line managers and HR personnel describe how the strong collective value of equal treatment can sometimes be a problem when managing interpersonal variation (Article IV). To sum up, findings of Article IV were developed based on the experience of line managers and their HR partners in the Swedish healthcare sector. However, the mechanisms that keep organizations from ad- justing to internal or external change , i.e., their adaptability to developments such as work- force aging and delayed retirement, are likely to apply to organizations outside this context as well. This is so, I argue , because the bureaucratic organizational principles on which the Swedish healthcare sector is partially based are among the most common organiza- tional principles in society. On the inter action between individual and organizational capabilities The challenges to prolong ing working lives become particularly apparent when both in- dividuals’ and organizations’ capabilities are deficient. Two older workers with similar abilities and preferences may have different employment opportunities depending on whether their employers are passive or proactive in implementing age-management strat- egies. As evident in Article IV, lack of resources, coordination , and age-awareness in pol- icies and practices seems to push older public welfare employees out of the workforce. In contrast, opportunities to negotiate idiosyncratic deals regarding tasks and responsibilities help older public welfare employees stick to their later preferred retirement age (Article III). Several conditions at the workplace level seem to interact in these relationships. For instance, high physical workload , low job satisfaction (Article I), and lack of de cision lat- itude when experiencing poor health push older workers toward early retirement (Article II) . Taken together, the afor ementioned results emphasize that the preconditions for a pro- longed working life are partially shaped through the interaction between conditions at the individual and organizational levels. However, what occupational groups are more likely to be employed in a passive organization and simultaneously experience low personal ca- pability? This thesis provides some answers to this ques tion. As previously discussed, the public welfare organizations examined in Article IV were characterized as passive or chained in relation to implementing age-management practices. At the same time, several 57 occupational groups within public welfare organizations generally experience high job de- mands (Aronsson et al., 2021; Cerdas et al., 2019; FORTE, 2015; SWEA, 2013) and are overrepresented in sick- leave statistics (SSIA, 2015; SSIA, 2020) . Also, older public wel- fare employees’ opportunities to negotiate individualized work arrangements were found to be unevenly distributed based on sex, age, and socioeconomic position (Article III) . Therefore, the overall preconditions for extending working life require a broader p erspec- tive that includes both individual capabilities ( i.e., willingness and ability) and organiza- tional capabilities ( i.e., resources and preferences). T o provide some practical examples, we will return to the fictive people from section 3 to demonstrate how individual and organizational capability interact. Figure 7. Preconditions for a prolonged working life: organizational capability to implement age-management practices, combined with a typology of individual abilities and willingness. Practical application of the mode l using Katrin, Johnny , Helena, and Olov Helena, Katrin, Johnny , and Olov have experienced various forms of mismatches between personal abilities/needs and demands/supplies at their job s, or a combination of both. In Katrin ’ s ( unwilling/able ) situation, the employer could, for example, reduce job demands or negotiate to set tasks and work responsibilities that are closer to her preferences. How- ever, although Katrin ’ s employer currently faces staffing shortfalls and expresses willing- ness to reorganize Katrin ’ s job, it currently lacks the resources to do so ( chained). For Johnny ( unwilling/unable ) , the employer could, for example, ergonomically adapt the work environment, reduce job demands, or offer new work tasks. Because the private enterprise that employs Johnny is undergoing financial struggles, the HR department has prohibited all such interventions ( passive). In a proactive or reactive organizat ion, Katrin ’ s and Johnny’ s situations could have been different. In exit interviews with older employees, the HR staff at the eldercare center where Helena ( willing/unable ) works identified the 56 likely produce more sustainable working conditions and counteract negative attitudes. B e- fore investing in age- management policies, a starting point is for top management to en- lighten and signal to the managers and their employees that prolonged labor market par- ticipation is a prioritized and desirable goal. Also, line managers must have sufficient time for day- to- day conversations with their employees to identify personal or general prob- lems in the work environment. It is also critical that the implementation of age - manage- ment policies be supported by co- workers and be considered fair and legitimate. This appears crucial in public welfare organizations in which line managers and HR personnel describe how the strong collective value of equal treatment can sometimes be a problem when managing interpersonal variation (Article IV) . To sum up, findings of Article IV were developed based on the experience of line managers and their HR partners in the Swedish healthcare sector. However, the mechanisms that keep organizations from ad- justing to internal or external change , i.e., their adaptability to developments such as work- force aging and delayed retirement, are likely to apply to organizations outside this context as well. This is so, I argue , because the bureaucratic organizational principles on which the Swedish healthcare sector is partially based are among the most common organiza- tional principles in society. On the interaction between individual and organizational capabilities The challenges to prolong ing working lives become particularly apparent when both in- dividuals’ and organizations’ capabilities are deficient. Two older workers with similar abilities and preferences may have different employment opportunities depending on whether their employers are passive or proactive in implementing age- management strat- egies. As evident in Article IV, lack of resources, coordination , and age- awareness in pol- icies and practices seems to push older public welfare employees out of the workforce. In contrast, opportunities to negotiate idiosyncratic deals regarding tasks and responsibilities help older public welfare employees stick to their later preferred retirement age (Article III). Several conditions at the workplace level seem to interact in these relationships. For instance, high physical workload , low job satisfaction (Article I), and lack of de cision lat- itude when experiencing poor health push older workers toward early retirement (Article II) . Taken together, the afor ementioned results emphasize that the preconditions for a pro- longed working life are partially shaped through the interaction between conditions at the individual and organizational levels. However, what occupational groups are more likely to be employed in a passive organization and simultaneously experience low personal ca- pability? This thesis provides some answers to this ques tion. As previously discussed, the public welfare organizations examined in Article IV were characterized as passive or chained in relation to implementing age- management practices. At the same time, several 57 occupational groups within public welfare organizations generally experience high job de- mands (Aronsson et al., 2021; Cerdas et al., 2019; FORTE, 2015; SWEA, 2013) and are overrepresented in sick-leave statistics (SSIA, 2015; SSIA, 2020) . Also, older public wel- fare employees’ opportunities to negotiate individualized work arrangements were found to be unevenly distributed based on sex, age, and socioeconomic position (Article III) . Therefore, the overall preconditions for extending working life require a broader p erspec- tive that includes both individual capabilities ( i.e., willingness and ability) and organiza- tional capabilities ( i.e., resources and preferences). T o provide some practical examples, we will return to the fictive people from section 3 to demonstrate how individual and organizational capability interact. Figure 7. Preconditions for a prolonged working life: organizational capability to implement age-management practices, combined with a typology of individual abilities and willingness. Practical application of the mode l using Katrin, Johnny , Helena, and Olov Helena, Katrin, Johnny , and Olov have experienced various forms of mismatches between personal abilities/needs and demands/supplies at their job s, or a combination of both. In Katrin ’ s ( unwilling/able ) situation, the employer could, for example, reduce job demands or negotiate to set tasks and work responsibilities that are closer to her preferences. How- ever, although Katrin ’ s employer currently faces staffing shortfalls and expresses willing- ness to reorganize Katrin ’ s job, it currently lacks the resources to do so ( chained). For Johnny ( unwilling/unable ) , the employer could, for example, ergonomically adapt the work environment, reduce job demands, or offer new work tasks. Because the private enterprise that employs Johnny is undergoing financial struggles, the HR department has prohibited all such interventions ( passive). In a proactive or reactive organizat ion, Katrin’ s and Johnny’ s situations could have been different. In exit interviews with older employees, the HR staff at the eldercare center where Helena ( willing/unable ) works identified the 58 problems and extended the age at which the employment contract would be terminated ( proactive). In Helena ’ s situation, the organization’ s capability to adapt enabled her to continue working. For Olov ( willing/able ) , the employer could offer additional monetary incentives, higher work flexibility, or competence development. Although Olov ’ s em- ployer, a private sheet metal firm, has sound finances and staffing, underdeveloped per- sonnel policies and the low engagement of lower managers resulted in failure to recognize Olov ’ s needs. However, after talking with his immediate manager about retiring as soon as financially feasible, the company reorganized Olov ’ s working hours and provided fi- nancial incentives to retain Olov a s an employee ( reactive). These hands -on examples il- lustrate how older workers’ preconditions to continue working are formed in the interac- tion between personal and work-related conditions and organizations ’ capabilities to re- tain older workers. Reflections on working life prolongation: for whom and for what purpose? Definitions of societal problems are seldom neutral. Certain actors with power and re- sources can dominate, select, and have the “ privilege of formulating the problem.” In the case of prolonged working life , there is a remarkable political consensus among most of the elected political parties in Sweden. Simultaneously, the resistance from social partners and other interest groups is relatively weak, despite this being one of the most important policy changes in decades. The research questions stressed in this thesis can be said, at least to some degree, to align with the dominant narratives of the political “ flank” supporting prolonged working life policies. For instance, many of the articles set out to explore under what personal and work-related circumstances older workers ’ labor market participation is shortened or extended and what conditions at the organizational level hinder older workers from continuing to work. However, it could be argued that ways to promote better working conditions and the voluntary prolongation of working lives can be found in answers to these research questions, regardless of whether or not the questions are politically biased. For instanc e, the scientific endeavor to find ways to create circumstances that enhance older workers’ capabilities may reduce involuntary labor market participation and strengthen work motivation. For labor market sectors experiencing labor shortages, such as Swedish welfare organizations, the ability to recruit and retain older workers might depend on their capability to provide stimulating, attractive, and decent work environ- ments. In addition , because higher retirement ages have already been legislated in many countries, my belief is that policymakers and employers need to pay attention to job qual- ity, equality, and sustainabili ty in contemporary working life. In previous research, retirement preferences and subsequent retirement behavior have often been assessed using retirement age as an outcome. It is relatively common for the authors, in line with the political narrative described above, to perceive attaining higher 59 retirement ages as a “ success” — intentionally or unintentionally. Hasselhorn (2020) ar- gued that, by doing so, “ retirement research is at risk of mirroring dominating societal interpretations, ” and that “ such conceptions may overlook potential costs of extension of working lives.” (Hasselhorn, 2020, p. 21) . The costs Hasselhorn (2020) referred to are the consequences for individuals (e.g., health and wellbeing), organizations (e.g., productivity), and society (e.g., intergenerational solidarity). Reflecting on this topic, I see that some of the emphases and perspectives in the individual constituent articles of this thesis could be accused of depicting higher retirement ages and their predictors as positive. T his has not been my intention. Instead, based on the findings of this thesis and earlier research, it is clear that delayed retirement is not necessarily positive for everyone. Considering scholars ’ concerns about PWL policies triggering inequalities in employment participation, health, social participation, and life chances (Foster and Walker, 2014; Hasselhorn, 2020; Ni Leime and Street, 2019; Nilsson et al., 2017) , it is not an understatement to claim that many of these problems can be derived from peoples’ working conditions and social po- sition in the labor market. C ertain demographic groups are more likely to experience an involuntary transition to retirement, either preceded by a financial need to continue work- ing or being pushed out of the labor market due to poor health or low work ability (Hofäcker and Radl, 2016) . Previous research shows that an involuntary transition to re- tirement is associated with adverse health and well- being (Hyde et al., 2015; König et al., 2018) . Therefore, insights into the interaction between individuals’ and organizations’ ca- pabilities could reduce involuntary employment participation or exit at the end of people’ s working lives. Methodological considerations In this section , I start by describing some of the limitations and strengths of the quantita- tive studies, and after that, I discuss the qualitative study. The objective of this thesis was to explore and explain individuals’ and organizations’ capabilities in the context of pro- longed working lives. The generalization of the present results merits further exploration within both the Swedish and international contexts given the considerable contextual dif- ferences between countries in terms of the configurations of their social security systems and cultures. Moreover , most relevant phenomena and conditions at the individual level are cumulative and relate to previous work histories. That is to say, the present findings to some extent reflect people’ s working life experiences from the past. This is worth con- sidering, given that much early retirement is related to poor health, which often originates from prior work exposures. The participants ’ working life participation spanned several decades, most ly during the second half of the 20th century. Participants in PSAE were born between 1938 and 1947 (median year of birth , 1944 ) and in Hearts – Lexliv were born between 1946 and 1964 (median year of birth , 1959). Given the many factors that affect 58 problems and extended the age at which the employment contract would be terminated ( proactive). In Helena ’ s situation, the organization’ s capability to adapt enabled her to continue working. For Olov ( willing/able ) , the employer could offer additional monetary incentives, higher work flexibility, or competence development. Although Olov ’ s em- ployer, a private sheet metal firm, has sound finances and staffing, underdeveloped per- sonnel policies and the low engagement of lower managers resulted in failure to recognize Olov ’ s needs. However, after talking with his immediate manager about retiring as soon as financially feasible, the company reorganized Olov ’ s working hours and provided fi- nancial incentives to retain Olov a s an employee ( reactive). These hands - on examples il- lustrate how older workers’ preconditions to continue working are formed in the interac- tion between personal and work- related conditions and organizations’ capabilities to re- tain older workers. Reflections on working life prolongation: for whom and for what purpose? Definitions of societal problems are seldom neutral. Certain actors with power and re- sources can dominate, select, and have the “ privilege of formulating the problem.” In the case of prolonged working life , there is a remarkable political consensus among most of the elected political parties in Sweden. Simultaneously, the resistance from social partners and other interest groups is relatively weak, despite this being one of the most important policy changes in decades. The research questions stressed in this thesis can be said, at least to some degree, to align with the dominant narratives of the political “ flank” supporting prolonged working life policies. For instance, many of the articles set out to explore under what personal and work- related circumstances older workers’ labor market participation is shortened or extended and what conditions at the organizational level hinder older workers from continuing to work. However, it could be argued that ways to promote better working conditions and the voluntary prolongation of working lives can be found in answers to these research questions, regardless of whether or not the questions are politically biased. For instanc e, the scientific endeavor to find ways to create circumstances that enhance older workers’ capabilities may reduce involuntary labor market participation and strengthen work motivation. For labor market sectors experiencing labor shortages, such as Swedish welfare organizations, the ability to recruit and retain older workers might depend on their capability to provide stimulating, attractive, and decent work environ- ments. In addition , because higher retirement ages have already been legislated in many countries, my belief is that policymakers and employers need to pay attention to job qual- ity, equality, and sustainabili ty in contemporary working life. In previous research, retirement preferences and subsequent retirement behavior have often been assessed using retirement age as an outcome. It is relatively common for the authors, in line with the political narrative described above, to perceive attaining higher 59 retirement ages as a “ success” —intentionally or unintentionally. Hasselhorn (2020) ar- gued that, by doing so, “ retirement research is at risk of mirroring dominating societal interpretations, ” and that “ such conceptions may overlook potential costs of extension of working lives.” (Hasselhorn, 2020, p. 21) . The costs Hasselhorn (2020) referred to are the consequences for individuals (e.g., health and wellbeing), organizations (e.g., productivity), and society (e.g., intergenerational solidarity). Reflecting on this topic, I see that some of the emphases and perspectives in the individual constituent articles of this thesis could be accused of depicting higher retirement ages and their predictors as positive. T his has not been my intention. Instead, based on the findings of this thesis and earlier research, it is clear that delayed retirement is not necessarily positive for everyone. Considering scholars ’ concerns about PWL policies triggering inequalities in employment participation, health, social participation, and life chances (Foster and Walker, 2014; Hasselhorn, 2020; Ni Leime and Street, 2019; Nilsson et al., 2017) , it is not an understatement to claim that many of these problems can be derived from peoples’ working conditions and social po- sition in the labor market. C ertain demographic groups are more likely to experience an involuntary transition to retirement, either preceded by a financial need to continue work- ing or being pushed out of the labor market due to poor health or low work ability (Hofäcker and Radl, 2016) . Previous research shows that an involuntary transition to re- tirement is associated with adverse health and well-being (Hyde et al., 2015; König et al., 2018) . Therefore, insights into the interaction between individuals’ and organizations’ ca- pabilities could reduce involuntary employment participation or exit at the end of people’ s working lives. Methodological considerations In this section , I start by describing some of the limitations and strengths of the quantita- tive studies, and after that, I discuss the qualitative study. The objective of this thesis was to explore and explain individuals’ and organizations’ capabilities in the context of pro- longed working lives. The generalization of the present results merits further exploration within both the Swedish and international contexts given the considerable contextual dif- ferences between countries in terms of the configurations of their social security systems and cultures. Moreover , most relevant phenomena and conditions at the individual level are cumulative and relate to previous work histories. That is to say, the present findings to some extent reflect people’ s working life experiences from the past. This is worth con- sidering, given that much early retirement is related to poor health, which often originates from prior work exposures. The participants ’ working life participation spanned several decades, most ly during the second half of the 20th century. Participants in PSAE were born between 1938 and 1947 (median year of birth , 1944) and in Hearts – Lexliv were born between 1946 and 1964 (median year of birth , 1959). Given the many factors that affect 60 older workers’ abilities and motivation to continue working, it is worth noting that this thesis covers only a limited number of independent factors. As far as the present author is aware, no response analyses were conducted for the PSAE dataset at the time of collection. However, a report from Stati stics Sweden reveals that the non-response was approximately 25 % for the ULF survey of living conditions (Statistics Sweden, 2003) . The present quantitative studies focus ed on individuals em- ployed in the labor market (Article s I and II) and in public welfare organizations (Article III), thereby excluding certain groups such as the unemployed, students, farmers, and the self-employed. Because all participants are employed, there is a risk of selection bias, often referred to as the “ healthy worker effect” (McMichael, 1976) , in which the sample consists of a stronger group of “ survivors.” H ere, I argue it is reasonable to assume that the selec- tion bias underestimates rather than overestimates the effect of the relationship between an independent variable and the dependent variable retirement. For example, in the rela- tionship between health and retirement, the “ unhealthy workers” would probably con- tribute to the effect. At the same time, the stronger “ survivors” are more resilient and capable of working despite having poor health and would not contribute to the effect. Error may also arise from how surveys are designed, questions asked, and data are col- lected. In Article s I and II, the personal and work -related determinants of retirement were assessed using a cross-sectional survey with self -reported questions at baseline. Because many participants were under 60 years old at the time of data collection, most of them continued working for many additional years. Respondents could have answered differ- ently before or after receiving the survey due to, for instance, leaving their job s for new employers or occupations or because of worsening personal health. The large time gap between the baseline survey and the actual exit is a limitation in Article s I and II , where a longitudinal approach with repeated measures would have been beneficial. In addition, several of the indicators in Article s I and II were measured using only single item s, which may fail to capture complex phenomena such as health, social support , and job satisfac- tion. For example, a prior study concerning job satisfaction by Oshagbemi (1999) exam- ined single versus multiple items and found that using single items overestimated positive attitudes and underestimated negative attitudes. The use of single items therefore poten- tially limits measurement validity. The post hoc face validity analysis in Article III indi- cated a two-fold interpretation of I -deals : either the respondents interpreted the questions as an explicit negotiation, which was the intention, or they interpreted them as conditions that were part of an employment or workplace contract in general. Although the post hoc analysis was restricted to a small number of participants, it potentially has implications for measurement validity and needs to be considered in future studies. Because several different measures for predicting retirement and its determinants can be found in earlier research, it is important to highlight some of their strengths and limita- tions. The limitations of subjective retirement measures, such as self -reported retirement 61 preferences (also plans, thoughts, and intentions), include recall bias, social norm effects, and variations in wording across survey studies. Subjective assessments are also more sen- sitive to respondents’ knowledge of, for example, pension systems , and overlook unfore- seen future changes in their private lives (e.g., health problems and di vorce) and working lives (e.g., career change and job loss). Prospective and subjective retirement measures may also fail to capture business- cycle fluctuations, evolving management ideas in organ- izations, and policy changes over time that may affect older workers’ motivation, ability, and need to remain employed. Also, it is important to note that actual retirement cannot be exactly defined by a drop in work- related income. Retirement is a complex process, and many individuals move in and out of different income protections schemes before they leave the labor market to take their pensions on a part- or full- time basis. Therefore, the discrimination criteria and ability to separate people as retired or active should not be considered exact. However, similar register- based definitions and measurements have been applied in recent peer- reviewed studies (see, for example, Eyjólfsdóttir et al., 2019b; König et al., 2021) . Although there are some limitations to defining retirement by a drop in work- related income, it still can be considered a strength. The high- quality data u sed in Article s I and II originate from a representative population survey combined with longi- tudinal register data and provide opportunities to follow individuals over 14 years, from baseline in 2002– 2003 to the end of the data collection in 2015. Three of the articles in this thesis focused on retirement age— preferred or actual— as the outcome. Because the existing data restrict us from explaining whether or not people’ s transition to retirement is voluntary, an important component of the capability model is missing, i. e., insights into whether retirement was selected and realized by the individual. Instead, the empirical studies focused on factors assumed to increase older workers’ ca- pabilities to select and achieve valuable goals, such as freedom at work, ability to in fluence work tasks, and options to negotiate individual ized work arrangements with employers. Another problem is that measuring working life duration per se, i.e., lower or higher re- tirement age, can be a blunt tool because it omits information about people’ s needs, abil- ities, and motivation s. For instance, unskilled manual workers and higher managers could potentially share the same average retirement age, although their motives for the retire- ment transition may be completely different. Therefore , future studies aiming to identify associations may include, for example, information about the degree of voluntariness in the transition to retirement. Here, I believe that the capability concept can help elucidate how an individual’ s decision to remain in work or leave the labor market aligns with their abilities and willingness, i.e., capabilities . Interesting contributions to the capability set needed to achieve sustainable employment have been proposed in theory by Van der Klink et al. (2016) and in practice by Abma et al. (2016) . Such an approach requires the gathering of considerable information and broadens our understanding of what individu- als can, want to, and actually do, i.e., not only what they do . Doing this makes it possible to assess what the older workforce request s in order to maintain its abilities and motivation 60 older workers’ abilities and motivation to continue working, it is worth noting that this thesis covers only a limited number of independent factors. As far as the present author is aware, no response analyses were conducted for the PSAE dataset at the time of collection. However, a report from Stati stics Sweden reveals that the non- response was approximately 25 % for the ULF survey of living conditions (Statistics Sweden, 2003) . The present quantitative studies focus ed on individuals em- ployed in the labor market (Article s I and II) and in public welfare organizations (Article III), thereby excluding certain groups such as the unemployed, students, farmers, and the self- employed. Because all participants are employed, there is a risk of selection bias, often referred to as the “ healthy worker effect” (McMichael, 1976) , in which the sample consists of a stronger group of “ survivors.” H ere, I argue it is reasonable to assume that the selec- tion bias underestimates rather than overestimates the effect of the relationship between an independent variable and the dependent variable retirement. For example, in the rela- tionship between health and retirement, the “ unhealthy workers” would probably con- tribute to the effect. At the same time, the stronger “ survivors” are more resilient and capable of working despite having poor health and would not contribute to the effect. Error may also arise from how surveys are designed, questions asked, and data are col- lected. In Article s I and II, the personal and work - related determinants of retirement were assessed using a cross- sectional survey with self- reported questions at baseline. Because many participants were under 60 years old at the time of data collection, most of them continued working for many additional years. Respondents could have answered differ- ently before or after receiving the survey due to, for instance, leaving their job s for new employers or occupations or because of worsening personal health. The large time gap between the baseline survey and the actual exit is a limitation in Article s I and II , where a longitudinal approach with repeated measures would have been beneficial. In addition, several of the indicators in Article s I and II were measured using only single item s, which may fail to capture complex phenomena such as health, social support , and job satisfac- tion. For example, a prior study concerning job satisfaction by Oshagbemi (1999) exam- ined single versus multiple items and found that using single items overestimated positive attitudes and underestimated negative attitudes. The use of single items therefore poten- tially limits measurement validity. The post hoc face validity analysis in Article III indi- cated a two- fold interpretation of I - deals: either the respondents interpreted the questions as an explicit negotiation, which was the intention, or they interpreted them as conditions that were part of an employment or workplace contract in general. Although the post hoc analysis was restricted to a small number of participants, it potentially has implications for measurement validity and needs to be considered in future studies. Because several different measures for predicting retirement and its determinants can be found in earlier research, it is important to highlight some of their strengths and limita- tions. The limitations of subjective retirement measures, such as self - reported retirement 61 preferences (also plans, thoughts, and intentions), include recall bias, social norm effects, and variations in wording across survey studies. Subjective assessments are also more sen- sitive to respondents’ knowledge of, for example, pension systems , and overlook unfore- seen future changes in their private lives (e.g., health problems and di vorce) and working lives (e.g., career change and job loss). Prospective and subjective retirement measures may also fail to capture business-cycle fluctuations, evolving management ideas in organ- izations, and policy changes over time that may affect older workers’ motivation, ability, and need to remain employed. Also, it is important to note that actual retirement cannot be exactly defined by a drop in work-related income . Retirement is a complex process, and many individuals move in and out of different income protections schemes before they leave the labor market to take their pensions on a part- or full-time basis. Therefore, the discrimination criteria and ability to separate people as retired or active should not be considered exact. However, similar register-based definitions and measurements have been applied in recent peer-reviewed studies (see, for example, Eyjólfsdóttir et al., 2019b; König et al., 2021) . Although there are some limitations to defining retirement by a drop in work-related income , it still can be considered a strength. The high-quality data u sed in Article s I and II originate from a representative population survey combined with longi- tudinal register data and provide opportunities to follow individuals over 14 years, from baseline in 2002– 2003 to the end of the data collection in 2015. Three of the articles in this thesis focused on retirement age— preferred or actual—as the outcome. Because the existing data restrict us from explaining whether or not people’ s transition to retirement is voluntary, an important component of the capability model is missing, i. e., insights into whether retirement was selected and realized by the individual. Instead, the empirical studies focused on factors assumed to increase older workers’ ca- pabilities to select and achieve valuable goals, such as freedom at work, ability to in fluence work tasks, and options to negotiate individual ized work arrangements with employers. Another problem is that measuring working life duration per se, i.e., lower or higher re- tirement age, can be a blunt tool because it omits information about people’ s needs, abil- ities, and motivation s. For instance, unskilled manual workers and higher managers could potentially share the same average retirement age, although their motives for the retire- ment transition may be completely different. Therefore , future studies aiming to identify associations may include, for example, information about the degree of voluntariness in the transition to retirement. Here, I believe that the capability concept can help elucidate how an individual’ s decision to remain in work or leave the labor market aligns with their abilities and willingness, i.e., capabilities . Interesting contributions to the capability set needed to achieve sustainable employment have been proposed in theory by Van der Klink et al. (2016) and in practice by Abma et al. (2016) . Such an approach requires the gathering of considerable information and broadens our understanding of what individu- als can, want to, and actually do, i.e., not only what they do . Doing this makes it possible to assess what the older workforce request s in order to maintain its abilities and motivation 62 to prolong the working life. Also, u sing this information allows organizations and policy- makers to tailor their policies and practices so that the older workforce has access to the needed resources in their workplaces (e.g., idiosyncratic deals and decision latitude) or the labor market (e.g., by regulation or laws) . Finally, the individual – organizational model pre- sented in this thesis provides a tool with which to identify vulnerable groups or positions in the labor market. The line managers and HR personnel participating in the qualitative study were recruited from two distinct regions in Sweden and were selected using a non-random sampling procedure. During the sampling process, a number of organizations declined to partici- pate in the study for various reasons, such as lack of time, staff shortages, and lack of interest. The organizations that declined to participate in the intervention program might have been worse off than those that participated. If this were the case, the different or- ganizational barriers described in A rticle IV might have been even more pronounced. It is also uncertain whether the organizations participating in the study reported in Article IV were actually representative of Sweden’ s whole healthcare sector. Therefore, a broader empirical approach covering a larger geographical area would be beneficial, in order to capture the differences in organizational and local contexts. As described earlier, because the participants first attended an age-management intervention, their attitudes to ward older workers were likely to change for the better, making it more challenging to identify negative attitudes or structural ageism and resulting in the overlooking of important fac- tors. 63 8 Conclusion Based on the results of the constituent studies of this thesis, both published (Jonsson et al., 2019; Jonsson et al., 2021; Jonsson et al., 2020) and, to date, unpublished (Article I), this thesis ends with the following conclusions: (1) Higher age, being female, and lower socioeconomic position were associated with preferences for earlier retirement and with earlier actual retirement behavior. In addition, these demographic groups were also shown to have fewer opportunities to negotiate in- dividualized work arrangements. (2) Poor health ( i.e., disease, illness, and sickness) i ncreased the likelihood of earlier retire- ment. This relationship appeared stronger when combined with an individual’ s absence of decision latitude, indicating a potential interaction effect. (3) Higher work ability and job satisfaction kept older workers in employment, while phys- ical exposure pushed older workers toward preferences for earlier retirement and earlier actual exit from the labor market. ( 4 ) Individualized work arrangements concerning the match between employees’ compe- tence, experience, and growth opportunities and their job s were related to later preferred retirement age among public welfare employees. ( 5 ) Multiple challenges related to the high workload of line managers, inflexible HR strat- egies, the absence of elaborated age- management strategies, and lack of coordination within the organization reinforced the public welfare employers’ inability to retain older workers. Practical implications, contributions, and relevance to society This thesis offers some guidance to legislators and organizations dealing with an aging workforce. It contributes to the debate on sustainability in the labor market and on PWL policies’ consequences for individuals’ capability to continue working. B ased on this the- sis’ s empirical results, policymakers and organizations can, for example, counteract phys- ical exposure with new equipment or technology, combat poor health among older work- ers by implementing workplace health promotion programs, plan interventions to increase 62 to prolong the working life. Also, u sing this information allows organizations and policy- makers to tailor their policies and practices so that the older workforce has access to the needed resources in their workplaces (e.g., idiosyncratic deals and decision latitude) or the labor market (e.g., by regulation or laws) . Finally, the individual – organizational model pre- sented in this thesis provides a tool with which to identify vulnerable groups or positions in the labor market. The line managers and HR personnel participating in the qualitative study were recruited from two distinct regions in Sweden and were selected using a non- random sampling procedure. During the sampling process, a number of organizations declined to partici- pate in the study for various reasons, such as lack of time, staff shortages, and lack of interest. The organizations that declined to participate in the intervention program might have been worse off than those that participated. If this were the case, the different or- ganizational barriers described in A rticle IV might have been even more pronounced. It is also uncertain whether the organizations participating in the study reported in Article IV were actually representative of Sweden’ s whole healthcare sector. Therefore, a broader empirical approach covering a larger geographical area would be beneficial, in order to capture the differences in organizational and local contexts. As described earlier, because the participants first attended an age- management intervention, their attitudes to ward older workers were likely to change for the better, making it more challenging to identify negative attitudes or structural ageism and resulting in the overlooking of important fac- tors. 63 8 Conclusion Based on the results of the constituent studies of this thesis, both published (Jonsson et al., 2019; Jonsson et al., 2021; Jonsson et al., 2020) and, to date, unpublished (Article I), this thesis ends with the following conclusions: (1) Higher age, being female, and lower socioeconomic position were associated with preferences for earlier retirement and with earlier actual retirement behavior. In addition, these demographic groups were also shown to have fewer opportunities to negotiate in- dividualized work arrangements. (2) Poor health ( i.e., disease, illness, and sickness) i ncreased the likelihood of earlier retire- ment. This relationship appeared stronger when combined with an individual’ s absence of decision latitude, indicating a potential interaction effect. (3) Higher work ability and job satisfaction kept older workers in employment, while phys- ical exposure pushed older workers toward preferences for earlier retirement and earlier actual exit from the labor market. ( 4 ) Individualized work arrangements concerning the match between employees’ compe- tence, experience, and growth opportunities and their job s were related to later preferred retirement age among public welfare employees. ( 5 ) Multiple challenges related to the high workload of line managers, inflexible HR strat- egies, the absence of elaborated age-management strategie s, and lack of coordination within the organization reinforced the public welfare employers’ inability to retain older workers. Practical implications, contributions, and relevance to society This thesis offers some guidance to legislators and organizations dealing with an aging workforce. It contributes to the debate on sustainability in the labor market and on PWL policies’ consequences for individuals’ capability to continue working. B ased on this the- sis’ s empirical results, policymakers and organizations can, for example, counteract phys- ical exposure with new equipment or technology, combat poor health among older work- ers by implementing workplace health promotion programs, plan interventions to increase 64 work ability, provide flexibility through offering higher decision latitude, improve skill matching, and offer growth opportunities and work tasks outside employees’ ordinary area of responsibilities. In practical terms, such measures are likely to strengthen older workers’ capability to continue working and are in line with conclusions drawn in past research (e.g., Bal et al., 2015; Hasselhorn and Apt, 2015; Ilmarinen, 2006; Naegele and Walker, 2006; Nilsson, 2016) . Researchers and policymakers can use the individual– or- ganizational model proposed here to identify vulnerable groups and understand how or- ganizational practices affect certain occupational groups in the workplace or labor market. It appears rational for policymakers and organizations to identify individual- specific and/or occupational -specific needs so that resources are allocated to those groups that need them the most. The examples of practical measures presented do not essentially diverge from measures mandated by existing regulations to prevent accidents and ill-health (SWEA, 2015) , which are assumed to be materialized by systematic work environment management (SAM) in Swedish workplaces. SAM requires that employers monitor older employees’ capabilities by conducting follow-up interviews and employee surveys. The close link between differ- ent forms of poor health and early retirement found in the present studies and earlier research sheds light on the importance of decent working conditions and opportunities to adapt work at the workplace. This is particularly relevant in several female - dominated occupations such as health, social care, and education, which have experienced high sick - leave rates (SSIA, 2015; SSIA, 2020) and have seen an increase in job demands over time (Aronsson et al., 2021; Cerdas et al., 2019) . Although this thesis did not focus on or contain any empirical data on OHS providers’ role in retaining older workers, it is likely that the poor organization and use of OHS providers in Swedish public welfare organiza- tions demonstrated by Schmidt et al. (2017) may reinforce the various challenges de- scribed earlier. Finally, in more general terms, it is essential to recognize that the lack of resources and challenges identified in the Swedish healthcare organizations in Article IV cannot be isolated from the broader political debate about the marketization, structural rearrangement, management (governed by philosophies such as NPM ) , and financing of public welfare as a whole. Therefore, several structural conditions need to be acknowl- edged to gain a complete picture of the Swedish healthcare sector’ s ability to create decent working conditions and provide high-quality services. Future perspectives In the section below, I suggest four areas where there is space for more research: (a) intervention and longitudinal studies, (a) the individual heterogeneity– universalism di- lemma, (c) technolog ical advances and older workers, and , finally, d) the impact of Covid - 19 on future working life. 65 a) Because most retirement determinants have been deduced from cross- sectional popu- lation- based studies, more workplace interventions to elucidate conditions that strengthen older workers’ employability, work ability, and health are needed. Also, longitudinal stud- ies to assess the cumulative nature of work exposure and how it relates to retirement behavior could be a potential avenue for future research. b) Given the aging and increasing heterogeneity of the workforce, we need more knowledge of how different universal approaches to work and working life affect the im- plementation of individual solutions and accommodations. Universal approaches refer to collective agreements, universal HRM practices targeting aggregated groups, and a collec- tive mindset. c) The significant advances in information technology, digitalization, automation, auton- omous vehicles, artificial intelligence, and robots will affect both contemporary and future working life. Although the consequences are difficult to predict, some occupations will likely disappear, be reorganized, or be replaced with new forms of jobs (for a Nordic overview, see, Dølvik and Steen, 2018) . Also, changing job requirements may increase the need for continuous competence development, often referred to as lifelong learning. For instance, ongoing discussions concerning policy adjustment in the Swedish Employment Protection Act include strengthening people’ s opportunities for competence development at their workplaces. The proposed changes in the Act also include public support for continuing education financed by the state (Government offices of Sweden, 2021) . How- ever, the notion of life- long learning requires workers to be mot ivated and continue learn- ing over their life course. Although robots and automation are likely to reduce physical work exposure, an increasing share of the workforce is carrying out work tasks in front of computer screens and mobile phones. Research indicates that use of such technology can increase the risks of new occupational hazards such as sleep disturbance, stress , and depression (Thomée et al., 2007) as well as neck pain and upper back and shoulder/arm symptoms (Lindegård Andersson, 2007) . These issues could suggest potential future re- search agendas. d) This thesis was partly written during a pandemic (the “ Covid - 19 ” virus) that spread from the city of Wuhan in China to Sweden in early 2020. To reduce the spread of the virus, the Swedish government asked public and private employers to reorganize their routines and permit their employees to work from home when possible. From early 2020 until the end of September 2021, public restrictions were imposed in offices, stores, the- atres, and sports venues. O ffice workers, academics, teachers, and students used video conferencing for communication, meetings, and lectures. A possible consequence of the Covid - 19 cri sis is that specific work organization features, such as flexibility (e.g., of sched- ule, location, and planning), ha ve taken a significant step forward. The experience will 64 work ability, provide flexibility through offering higher decision latitude, improve skill matching, and offer growth opportunities and work tasks outside employees’ ordinary area of responsibilities. In practical terms, such measures are likely to strengthen older workers’ capability to continue working and are in line with conclusions drawn in past research (e.g., Bal et al., 2015; Hasselhorn and Apt, 2015; Ilmarinen, 2006; Naegele and Walker, 2006; Nilsson, 2016) . Researchers and policymakers can use the individual– or- ganizational model proposed here to identify vulnerable groups and understand how or- ganizational practices affect certain occupational groups in the workplace or labor market. It appears rational for policymakers and organizations to identify individual- specific and/or occupational - specific needs so that resources are allocated to those groups that need them the most. The examples of practical measures presented do not essentially diverge from measures mandated by existing regulations to prevent accidents and ill- health (SWEA, 2015) , which are assumed to be materialized by systematic work environment management (SAM) in Swedish workplaces. SAM requires that employers monitor older employees’ capabilities by conducting follow- up interviews and employee surveys. The close link between differ- ent forms of poor health and early retirement found in the present studies and earlier research sheds light on the importance of decent working conditions and opportunities to adapt work at the workplace. This is particularly relevant in several female - dominated occupations such as health, social care, and education, which have experienced high sick - leave rates (SSIA, 2015; SSIA, 2020) and have seen an increase in job demands over time (Aronsson et al., 2021; Cerdas et al., 2019) . Although this thesis did not focus on or contain any empirical data on OHS providers’ role in retaining older workers, it is likely that the poor organization and use of OHS providers in Swedish public welfare organiza- tions demonstrated by Schmidt et al. (2017) may reinforce the various challenges de- scribed earlier. Finally, in more general terms, it is essential to recognize that the lack of resources and challenges identified in the Swedish healthcare organizations in Article IV cannot be isolated from the broader political debate about the marketization, structural rearrangement, management (governed by philosophies such as NPM ) , and financing of public welfare as a whole. Therefore, several structural conditions need to be acknowl- edged to gain a complete picture of the Swedish healthcare sector’ s ability to create decent working conditions and provide high- quality services. Future perspectives In the section below, I suggest four areas where there is space for more research: (a) intervention and longitudinal studies, (a) the individual heterogeneity– universalism di- lemma, (c) technolog ical advances and older workers, and , finally, d) the impact of Covid - 19 on future working life. 65 a) Because most retirement determinants have been deduced from cross-sectional popu- lation- based studies, more workplace interventions to elucidate conditions that strengthen older workers’ employability, work ability, and health are needed. Also, longitudinal stud- ies to assess the cumulative nature of work exposure and how it relates to retirement behavior could be a potential avenue for future research. b) Given the aging and increasing heterogeneity of the workforce, we need more knowledge of how different universal approaches to work and working life affect the im- plementation of individual solutions and accommodations. Universal approaches refer to collective agreements, universal HRM practices targeting aggregated groups, and a collec- tive mindset. c) The significant advances in information technology, digitalization, automation, auton- omous vehicles, artificial intelligence, and robots will affect both contemporary and future working life. Although the consequences are difficult to predict, some occupations will likely disappear, be reorganized, or be replaced with new forms of jobs (for a Nordic overview, see, Dølvik and Steen, 2018) . Also, changing job requirements may increase the need for continuous competence development, often referred to as lifelong learning. For instance, ongoing discussions concerning policy adjustment in the Swedish Employment Protection Act include strengthening people’ s opportunities for competence development at their workplaces. The proposed changes in the Act also include public support for continuing education financed by the state (Government offices of Sweden, 2021) . How- ever, the notion of life-long learning requires workers to be mot ivated and continue learn- ing over their life course. Although robots and automation are likely to reduce physical work exposure, an increasing share of the workforce is carrying out work tasks in front of computer screens and mobile phones. Research indicates that use of such technology can increase the risks of new occupational hazards such as sleep disturbance, stress , and depression (Thomée et al., 2007) as well as neck pain and upper back and shoulder/arm symptoms (Lindegård Andersson, 2007) . These issues could suggest potential future re- search agendas. d) This thesis was partly written during a pandemic (the “ Covid -19” virus) that spread from the city of Wuhan in China to Sweden in early 2020. To reduce the spread of the virus, the Swedish government asked public and private employers to reorganize their routines and permit their employees to work from home when possible. From early 2020 until the end of September 2021, public restrictions were imposed in offices, stores, the- atres, and sports venues. O ffice workers, academics, teachers, and students used video conferencing for communication, meetings, and lectures. A possible consequence of the Covid -19 cri sis is that specific work organization features, such as flexibility (e.g., of sched- ule, location, and planning), ha ve taken a significant step forward. The experience will 66 probably also change how people assess and value flexibility and how they view the work- place as an arena for social relationships and cohesion. The pandemic did not have an impact on the data collection for this thesis, but it will likely influence people’ s perception of work flexibility and organization in the future. However, this has not been the case for occupations with lower location flexibility, such as store clerks, chefs, truck drivers, and healthcare personnel. Instead, the pandemic has led to higher strain and job demands for some of these occupations, especially in the healthcare sector and welfare organizations. The pandemic put extraordinary pressure on the healthcare sector and increased public awareness of current staffing and equipment shortages. At the same time, the pandemic placed the Swedish health sector at the center of public debate and strengthened the sec- tor’ s role and status in society; in addition, many retire es, such as nurses, were re-employed in the national vaccination program. The implication s of the pandemic will open many potential research agendas in the future. 67 Svensk sammanfattning [summ ary in Swedish] Att behålla den åldrande arbetskraften: Studier av individu- ell och organisatorisk kapabilitet i kontexten av ett förlängt ar- betsliv Bakgrund Mellan arbetslivets två noder – inträde och utträde – har arbetets utformning och innehåll betydelse för våra kroppar och livschanser. Pension är vanligtvis ett normalt avslut efter ett långt arbetsliv, men vägen dit kan se olika ut beroende på yrke, kön, hälsa, levnadsför- hållanden och etnisk bakgrund. Varje dag går ungefär 317 svenskar i pension, de flesta permanent. Alltmedan 1900 - talet kantades av politiska reformer för att förbättra levnads- förhållandena för pensionärer och höja pensionssystemets kompensationsgrad i förhål- lande till slutlönen, har 2000 - talet kommit att handla om att mobilisera arbetskraften till ett längre arbetsliv. I den politiska debatten har förhållandet mellan antalet personer utan- för arbetsmarknaden i relation till antalet sysselsatta varit huvudargumentet till det för- längda arbetslive t i syfte att säkerställa pensionssystemets finansiering. Sedan 1990 - talets slut fram till idag har därför en rad politiska beslut fattats som syftat till att höja de eko- nomiska incitamenten för arbete, begränsa möjligheten till förtidspensionering samt hö ja ålderströsklar i pensionssystemet. Frågan har dock inte enbart berört det statsfinansiella perspektivet. I de sektorer på arbetsmarknaden som upplevt problem med personalför- sörjningen har äldre medarbetare lyfts fram som en potentiell arbetskraftsresurs att behålla och rekrytera, särskilt bland arbetsgivare inom välfärden. En höjning av pensionsåldern är emellertid inte oproblematisk eftersom människors för- måga och vilja att arbeta påverkas av deras ackumulerade erfarenheter från privat - och arbetsliv. Sent i arbetslivet upplever många, särskilt inom vissa yrkesgrupper, att deras arbetsförmåga försämras och att motivation till att fortsätta arbeta saknas. I strävan efter att förlänga arbetslivet har mycket forskning ägnats åt att identifiera faktorer som möjlig- gör och hindrar äldre arbetstagare från att fortsätta arbeta på individuell, organisatorisk och samhällelig nivå. En mängd faktorer såsom arbetstrivsel, hälsa, förhållanden på ar- betsplatsen samt social- och pensionssystem har visat sig betydelsefulla för äldre arbetsta- gares pensionsbeslut (Browne et al., 2019; Edge et al., 2017; Fisher et al., 2016) . De statliga målen om en hållbar förlängning av arbetslivet förutsätter alltså att individer bibe- håller sin arbetsförmåga och vilja att fortsätta arbeta, samt att organisationer uppfattar den äldre arbetskraften som önskvärd och har förmågan at t hantera dess särskilda behov och önskemål på arbetsplatsen. Med andra ord bör anställdas minskande arbetsförmåga eller 66 probably also change how people assess and value flexibility and how they view the work- place as an arena for social relationships and cohesion. The pandemic did not have an impact on the data collection for this thesis, but it will likely influence people’ s perception of work flexibility and organization in the future. However, this has not been the case for occupations with lower location flexibility, such as store clerks, chefs, truck drivers, and healthcare personnel. Instead, the pandemic has led to higher strain and job demands for some of these occupations, especially in the healthcare sector and welfare organizations. The pandemic put extraordinary pressure on the healthcare sector and increased public awareness of current staffing and equipment shortages. At the same time, the pandemic placed the Swedish health sector at the center of public debate and strengthened the sec- tor’ s role and status in society; in addition, many retire es, such as nurses, were re- employed in the national vaccination program. The implication s of the pandemic will open many potential research agendas in the future. 67 Svensk sammanfattning [summ ary in Swedish] Att behålla den åldrande arbetskraften: Studier av individu- ell och organisatorisk kapabilitet i kontexten av ett förlängt ar- betsliv Bakgrund Mellan arbetslivets två noder – inträde och utträde – har arbetets utformning och innehåll betydelse för våra kroppar och livschanser. Pension är vanligtvis ett normalt avslut efter ett långt arbetsliv, men vägen dit kan se olika ut beroende på yrke, kön, hälsa, levnadsför- hållanden och etnisk bakgrund. Varje dag går ungefär 317 svenskar i pension, de flesta permanent. Alltmedan 1900 - talet kantades av politiska reformer för att förbättra levnads- förhållandena för pensionärer och höja pensionssystemets kompensationsgrad i förhål- lande till slutlönen, har 2000 -tale t kommit att handla om att mobilisera arbetskraften till ett längre arbetsliv. I den politiska debatten har förhållandet mellan antalet personer utan- för arbetsmarknaden i relation till antalet sysselsatta varit huvudargumentet till det för- längda arbetslive t i syfte att säkerställa pensionssystemets finansiering. Sedan 1990 -talets slut fram till idag har därför en rad politiska beslut fattats som syftat till att höja de eko- nomiska incitamenten för arbete, begränsa möjligheten till förtidspensionering samt hö ja ålderströsklar i pensionssystemet. Frågan har dock inte enbart berört det statsfinansiella perspektivet. I de sektorer på arbetsmarknaden som upplevt problem med personalför- sörjningen har äldre medarbetare lyfts fram som en potentiell arbetskraftsresurs att behålla och rekrytera, särskilt bland arbetsgivare inom välfärden. En höjning av pensionsåldern är emellertid inte oproblematisk eftersom människors för- måga och vilja att arbeta påverkas av deras ackumulerade erfarenheter från privat - och arbetsliv. Sent i arbetslivet upplever många, särskilt inom vissa yrkesgrupper, att deras arbetsförmåga försämras och att motivation till att fortsätta arbeta saknas. I strävan efter att förlänga arbetslivet har mycket forskning ägnats åt att identifiera faktorer som möjlig- gör och hindrar äldre arbetstagare från att fortsätta arbeta på individuell, organisatorisk och samhällelig nivå. En mängd faktorer såsom arbetstrivsel, hälsa, förhållanden på ar- betsplatsen samt social- och pensionssystem har visat sig betydelsefulla för äldre arbetsta- gares pensionsbeslut (Browne et al., 2019; Edge et al., 2017; Fisher et al., 2016) . De statliga målen om en hållbar förlängning av arbetslivet förutsätter alltså att individer bibe- håller sin arbetsförmåga och vilja att fortsätta arbeta, samt att organisationer uppfattar den äldre arbetskraften som önskvärd och har förmågan at t hantera dess särskilda behov och önskemål på arbetsplatsen. Med andra ord bör anställdas minskande arbetsförmåga eller 68 ovilja att fortsätta arbeta inte enbart förstås utifrån ett individperspektiv, utan måste be- traktas i sitt sammanhang och i relation till den arbetsplats och det arbete individen utför. För att öka förståelsen för samspelet mellan påverkansfaktorer på individ - respektive or- ganisationsnivå används i denna avhandling begreppet kapabilitet som övergripande teo- retisk utgångspunkt (Sen, 2009) . Kapabilitet belyser människors och organisationers möj- ligheter att omvandla existerande resurser i syfte att nå uppsatta mål. På individnivå avser kapabilitet den enskilda arbetstagarens förutsättningar att kunna fortsätta arbeta, exem- pelvis dennes hälsa och/eller ekonomiska situation (Sen, 2009) . På organisationsnivå avser begreppet arbetsgivares vilja och förmåga att anpassa arbetet på ett sådant sätt att äldre anställda väljer att fortsätta arbeta längre (Kadefors et al., 2020). Med en ökad andel äldre på arbetsmarknaden har arbetsgivarnas roll lyfts fram som central i såväl forskning och statliga utredningar eftersom de kan erbjuda arbets förhållanden som bättre passar en åldrande arbetskraft. Tidigare forskning har exempelvis visat att tidig pensionering många gånger orsakas av att arbetsuppgifter inte är tillräckligt anpassade ef- ter arbetstagaren. Ofta används den engelska termen age management för att beskriva olika former av insatser avseende arbetsmiljö och personalpolicys som arbetsgivare vidtar för att säkerställa goda, hållbara och utvecklande arbetsförhållanden för äldre medarbetare. Med utgångspunkt i idén om att äldre är en grupp med särskilda behov, vars önskemål arbetsgivare kan tillgodose med hjälp av (pro)aktiva insatser, kan arbetet modifieras ge- nom exempelvis ökad flexibilitet eller förbättrade utvecklingsmöjligheter (Eppler -Hattab et al., 2019) . I förlängningen antas detta kunna främja äldre arbetstagares förmåga, såväl som vilja, att arbeta längre. Såda na arbetsanpassningar kan vara generella, det vill säga gälla alla anställda, men också individuella i den bemärkelsen att arbetsuppgifterna aktivt an- passas efter den enskilda medarbetarens förmåga och preferenser. Gemensamt för alla typer av anpassningar är dock att implementeringen villkoras av organisationens kapabili- tet , det vill säga av arbetsplatsens kontext, de aktuella arbetsuppgifternas natur samt vilka resurser och vilja arbetsgivaren har att exempelvis genomföra arbetsanpassningar för äldre arbetstagare. I sin tur betonar detta vikten av att studera samspelet mellan individuella och organisatoriska förutsättningar samt hur det påverkar arbetsmarknadsdeltagande bland den äldre arbetskraften. Denna avhandling till att öka förståelsen för hur förhåll anden på organisationsnivå påverkar arbetstagares förutsättningar att förlänga sitt arbetsliv. Likaså bidrar avhandlingen metodologiskt genom att empiriskt studera hur betydelsen av ett antal möjliga påverkansfaktorer på såväl individ - som organisationsnivå skiljer sig åt beroende på hur pension operationaliseras. Syfte Syftet med avhandlingen är att undersöka hur individuella och organisatoriska faktorer påverkar äldre arbetstagares möjligheter och motivation att fortsätta arbeta i ljuset av ett längre arbetsliv. Följande tre frågeställningar undersöks: 69 (1) Vilka personliga och arbetsrelaterade faktorer främjar e ller hindrar äldre arbetsta- gare förmåga eller vilja fortsätta arbeta? (2) Vilka hinder och möjligheter påverkar offentliga välfärdsorganisationers kapabilitet att behålla äldre arbetare? (3) Hur samspelar individers kapabilitet med organisationers kapabilitet i kontexten av ett längre arbetsliv? Datamaterial Den empiriska analysen är baserad på tre datamängder. Den första datamängden består av en befolkningsrepresentativ enkätundersökning från Panel Survey Aging and the El- derly (PSAE) med 1143 individer mellan 55 och 64 år. Deltagarna i PSAE matchades med inkomstinformation från registerdata (Longitudinal Integration Data Base for Health In- surance and Labor Market Studies, LISA) . PSAE användes i Artikel 1 och Artikel 2. Den andra datamängden användes i Artikel III och består av en tvärsnittsstudie med 4499 medarbetare mellan 55 och 74 år i Göteborgs stad (” Hearts - Lexliv ”). Deltagarna i Hearts - Lexliv är verksamma inom olika avdelningar i Göteborg stad, exempelvis omsorg, teknisk förvaltning och skola. Den tredje datamängden användes i Artikel IV och består av en intervjustudie med 8 enhetschefer och deras 11 HR partner inom HR i sjukvård, vård och omsorg i Göteborg samt Skåne (”ArbetsKraftsprojektet”). Resultat Artikel I: Personliga och arbetsrelaterade faktorer för pension I Artikel 1 undersöktes om pensionspreferenser predikterar faktiskt utträde. Dessutom jämfördes hur ett antal personliga och arbetsrelaterade faktorer påverkar pensionsprefe- renser respektive faktiskt pensionsutträde. Resultaten tyder på att deltagarna tenderade att gå i pension senare än vad de ursprungligen uppgav i enkäten, vilket troligen kan förklaras med förändringar i pensionssystemet och förbättrad hälsa över datainsamlingsperioden. Faktorer som påverkar både tidigare pensionspreferenser och faktiskt arbetsmark nadsut- träde är dålig hälsa, låg arbetsförmåga, fysiskt tunga arbeten och låg arbetstrivsel. De iden- tifierade faktorerna för pensionsutträde har god förankring i tidigare forskningslitteratur. Faktorer som rör sociala aspekter på arbetsplatsen, exempelvis s ocialt stöd och erkän- nande visade sig enbart ha betydelse för pensionspreferenser men inte faktiskt arbets- marknadsutträde. Resultatet visade också att fler personliga och arbetsrelaterade faktorer har ett signifikant samband med pensionspreferens jämfört m ed faktiskt utträde. På vilket sätt pension operationaliseras kan alltså ha betydelse för vilka samband som identifieras och därmed vilka övergripande slutsatser som dras. Det innebär att det finns en risk att 68 ovilja att fortsätta arbeta inte enbart förstås utifrån ett individperspektiv, utan måste be- traktas i sitt sammanhang och i relation till den arbetsplats och det arbete individen utför. För att öka förståelsen för samspelet mellan påverkansfaktorer på individ - respektive or- ganisationsnivå används i denna avhandling begreppet kapabilitet som övergripande teo- retisk utgångspunkt (Sen, 2009) . Kapabilitet belyser människors och organisationers möj- ligheter att omvandla existerande resurser i syfte att nå uppsatta mål. På individnivå avser kapabilitet den enskilda arbetstagarens förutsättningar att kunna fortsätta arbeta, exem- pelvis dennes hälsa och/eller ekonomiska situation (Sen, 2009) . På organisationsnivå avser begreppet arbetsgivares vilja och förmåga att anpassa arbetet på ett sådant sätt att äldre anställda väljer att fortsätta arbeta längre (Kadefors et al., 2020). Med en ökad andel äldre på arbetsmarknaden har arbetsgivarnas roll lyfts fram som central i såväl forskning och statliga utredningar eftersom de kan erbjuda arbets förhållanden som bättre passar en åldrande arbetskraft. Tidigare forskning har exempelvis visat att tidig pensionering många gånger orsakas av att arbetsuppgifter inte är tillräckligt anpassade ef- ter arbetstagaren. Ofta används den engelska termen age management för att beskriva olika former av insatser avseende arbetsmiljö och personalpolicys som arbetsgivare vidtar för att säkerställa goda, hållbara och utvecklande arbetsförhållanden för äldre medarbetare. Med utgångspunkt i idén om att äldre är en grupp med särskilda behov, vars önskemål arbetsgivare kan tillgodose med hjälp av (pro)aktiva insatser, kan arbetet modifieras ge- nom exempelvis ökad flexibilitet eller förbättrade utvecklingsmöjligheter (Eppler - Hattab et al., 2019) . I förlängningen antas detta kunna främja äldre arbetstagares förmåga, såväl som vilja, att arbeta längre. Såda na arbetsanpassningar kan vara generella, det vill säga gälla alla anställda, men också individuella i den bemärkelsen att arbetsuppgifterna aktivt an- passas efter den enskilda medarbetarens förmåga och preferenser. Gemensamt för alla typer av anpassningar är dock att implementeringen villkoras av organisationens kapabili- tet , det vill säga av arbetsplatsens kontext, de aktuella arbetsuppgifternas natur samt vilka resurser och vilja arbetsgivaren har att exempelvis genomföra arbetsanpassningar för äldre arbetstagare. I sin tur betonar detta vikten av att studera samspelet mellan individuella och organisatoriska förutsättningar samt hur det påverkar arbetsmarknadsdeltagande bland den äldre arbetskraften. Denna avhandling till att öka förståelsen för hur förhåll anden på organisationsnivå påverkar arbetstagares förutsättningar att förlänga sitt arbetsliv. Likaså bidrar avhandlingen metodologiskt genom att empiriskt studera hur betydelsen av ett antal möjliga påverkansfaktorer på såväl individ - som organisationsnivå skiljer sig åt beroende på hur pension operationaliseras. Syfte Syftet med avhandlingen är att undersöka hur individuella och organisatoriska faktorer påverkar äldre arbetstagares möjligheter och motivation att fortsätta arbeta i ljuset av ett längre arbetsliv. Följande tre frågeställningar undersöks: 69 (1) Vilka personliga och arbetsrelaterade faktorer främjar e ller hindrar äldre arbetsta- gare förmåga eller vilja fortsätta arbeta? (2) Vilka hinder och möjligheter påverkar offentliga välfärdsorganisationers kapabilitet att behålla äldre arbetare? (3) Hur samspelar individers kapabilitet med organisationers kapabilitet i kontexten av ett längre arbetsliv? Datamaterial Den empiriska analysen är baserad på tre datamängder. Den första datamängden består av en befolkningsrepresentativ enkätundersökning från Panel Survey Aging and the El- derly (PSAE) med 1143 individer mellan 55 och 64 år. Deltagarna i PSAE matchades med inkomstinformation från registerdata (Longitudinal Integration Data Base for Health In- surance and Labor Market Studies, LISA) . PSAE användes i Artikel 1 och Artikel 2. Den andra datamängden användes i Artikel III och består av en tvärsnittsstudie med 4499 medarbetare mellan 55 och 74 år i Göteborgs stad (” Hearts -Lexliv ”). Deltagarna i Hearts - Lexliv är verksamma inom olika avdelningar i Göteborg stad, exempelvis omsorg, teknisk förvaltning och skola. Den tredje datamängden användes i Artikel IV och består av en intervjustudie med 8 enhetschefer och deras 11 HR partner inom HR i sjukvård, vård och omsorg i Göteborg samt Skåne (”ArbetsKraftsprojektet”). Resultat Artikel I: Personliga och arbetsrelaterade faktorer för pension I Artikel 1 undersöktes om pensionspreferenser predikterar faktiskt utträde. Dessutom jämfördes hur ett antal personliga och arbetsrelaterade faktorer påverkar pensionsprefe- renser respektive faktiskt pensionsutträde. Resultaten tyder på att deltagarna tenderade att gå i pension senare än vad de ursprungligen uppgav i enkäten, vilket troligen kan förklaras med förändringar i pensionssystemet och förbättrad hälsa över datainsamlingsperioden. Faktorer som påverkar både tidigare pensionspreferenser och faktiskt arbetsmark nadsut- träde är dålig hälsa, låg arbetsförmåga, fysiskt tunga arbeten och låg arbetstrivsel. De iden- tifierade faktorerna för pensionsutträde har god förankring i tidigare forskningslitteratur. Faktorer som rör sociala aspekter på arbetsplatsen, exempelvis s ocialt stöd och erkän- nande visade sig enbart ha betydelse för pensionspreferenser men inte faktiskt arbets- marknadsutträde. Resultatet visade också att fler personliga och arbetsrelaterade faktorer har ett signifikant samband med pensionspreferens jämfört m ed faktiskt utträde. På vilket sätt pension operationaliseras kan alltså ha betydelse för vilka samband som identifieras och därmed vilka övergripande slutsatser som dras. Det innebär att det finns en risk att 70 forskare ger felaktiga policyrekommendationer om pensionspreferenser används som ut- fall istället för faktiskt utträde. Artikel II: Arbetsanpassningar för att stödja äldre arbetstagare med dålig hälsa I Artikel II undersöktes om dålig hälsa i form av tidigare sjukfrånvaro, låg självskattad hälsa eller långvarig sjukdom har ett samband med en lägre arbetsförmåga och tidigare utträde från arbetslivet. Dålig hälsa visade sig ha en negativ effekt på fysisk och psykisk arbetsförmåga samt ökar risken för tidigare arbetsmarknadsutträde. I artikeln undersöktes även antagandet att en hög grad av handlingsutrymme ökar arbetstagarens möjligheter att anpassa arbetet efter förmåga och önskemål. Med handlingsutrymme avsågs i detta fall frihet att bestämma över sin arbetstid, arbetstakt och planering, samt möjligheter till in- flytande över när, var och hur arbetet utförs. Resultaten visar att dålig hälsa i form av långvarig sjukdom, sjukfrånvaro eller låg självskattad hälsa, är viktiga faktorer som tidiga- relägger arbetstagares arbetsmarknadsutträde. Vidare visar resulta tet att ha dålig hälsa i form av långvarig sjukdom och sjukfrånvaro i kombination med lågt handlingsutrymme ökar risken för tidig pension vilket kan indikera en interaktionseffekt. Samma resultat för självskattad hälsa var däremot inte lika konsistent. Med andra ord kan möjlighet till ar- betsanpassning vid ohälsa öka äldre arbetstagares möjligheter att fortsätta arbeta. Dessa samband bör emellertid studeras vidare med mer validerade mätinstrument. Artikel III : Betydelsen av kompetensmatchning för pensionspr eferenser I Artikel III undersöktes betydelsen av individuella arbetsöverenskommelser mellan an- ställd och arbetsgivare. Individuella arbetsöverenskommelser avsåg bland annat arbets- uppgifter och ansvar, platsflexibilitet, minskad arbetsbelastning, finansiel la incitament och schemaflexibilitet för äldre arbetstagares pensionspreferenser. Resultaten pekar på att ar- betsöverenskommelser avseende arbetsuppgifter och ansvar senarelägger anställdas öns- kemål om pension. Intressant nog visade resultaten även på ett m otsatt förhållande för arbetsöverenskommelser rörande mindre arbetsbelastning. Troligtvis beror det på att de som har fått reducerad arbetsbelastning är samma arbetstagare som lider av sämre hälsa och därmed även har tidigare pensionspreferenser. Resultate t indikerar också att det finns utrymme för ytterligare arbetsanpassningar utöver nedsatt arbetsbelastning. Studien visar även att omfattningen av individuella överenskommelser är begränsad , särskilt bland kvin- nor, äldre, arbetstagare med dålig hälsa och i lägre socioekonomisk position. Efter datain- samling och analys genomfördes en valideringsanalys av mätinstrumentet som indikerade att individuella arbetsöverenskommelser kan förväxlas med arbetsanpassningar och flexi- bilitet som redan existerar på arbetsplatsen eller erbjuds av arbetsgivaren. Den svenska kontexten med starka kollektivavtal, kollektivt tankesätt och personalpolicys riktade till aggregerade grupper, särskilt i välfärdsorganisat ioner, är några förklaringar till den begrän- sade omfattningen av arbetsöverenskommelser men också de delvis svårtolkade resultaten i studien. Slutsatserna från artikeln är att arbetsöverenskommelser som berör arbetsupp- gifter och ansvar har betydelse för ar betstagarnas önskemål om tidigare eller senarelägg- 71 ning av pension. Detta innebär att en potentiell policyåtgärd är att säkerställa god kompe- tensmatchning. Frågan om arbetsöverenskommelsernas roll och betydelse i svensk kon- text behöver emellertid djupare te oretisk och empirisk analys i framtida forskning. Artikel IV: Organisatoriska hinder att behålla äldre anställda I Artikel IV , intervjuades linjechefer och deras HR stöd om hinder och möjligheter att behålla äldre anställda inom vård - och sjukvård. Genere llt uppvisade både linjechefer och HR personal en positiv attityd till äldre medarbetare och framhävde deras betydelse som mentorer för yngre. Dock var implementeringen av age- management- strategier ytterst be- gränsad och begreppet age management användes in te i de undersökta organisationerna. De organisatoriska hinder för att organisationen skulle behålla äldre anställda identifiera- des som linjechefernas höga arbetsbelastning, frånvaro av flexibla personalpolicys, under- utvecklade age- management strategier och dålig koordination inom organisationen. Dessa organisatoriska hinder tycks förstärka varandra i en ond cirkel, exempelvis innebar de stora personalbehoven och det administrativa arbete att linjecheferna tvingades fokusera på den dagliga driften vilket to g tid från strategiskt utvecklingsarbete med arbetsmetoder och personalpolicys. Avsaknaden av nya arbetsmetoder eller policys resulterade i frånvaro av verktyg och incitament för linjechefer i arbetet med att behålla äldre arbetstagare. Ef- tersom linjechefe rna redan befinner sig i en ansträngd arbetssituation leder det till att frå- gan om att bibehålla äldre anställda riskerar att bli beroende av individuella linjechefens vilja och attityder till äldre medarbetare och därmed godtycklig ur ett strategiskt pers pek- tiv. Linjechefer och HR personal beskrev också svårigheterna med att hantera individuella önskemål eller behov utifrån existerande personalpolicys och avtal, men också på grund av starka likabehandlingsideal, exempelvis uttrycktes svårigheter att indivi danpassa sche- maläggning eller arbetsuppgifter samt begränsade möjligheter att omplacera medarbetare efter rehabilitering. Sammantaget leder dessa förhållanden till att vård - och omsorgsorga- nisationer inte lyckas tillvarata den potentiella arbetskraftsresurs som äldre medarbetare utgör. För att kunna behålla den arbetskraftsresurs som äldre anställda utgör måste orga- nisationerna bli mer proaktiva och koordinerade på både strategisk - och operativ nivå. Detta innebär att högre ledning måste lyfta fram betydels en av att behålla äldre medarbe- tare och allokera tillräckliga resurser så att linjechefer får en balanserad arbetssituation för att kunna prioritera frågan om äldre medarbetare. Diskussion Avhandlingen undersökte hur äldre medarbetares arbetsmarknadsdelt agande formas via ett samspel mellan å ena sidan personliga förmågor och preferenser och å andra sidan förutsättningar på arbetsplats - och organisationsnivå. Centralt i denna diskussion är ka- pabilitetsbegreppet , vilket innefattar både individers kapabilite t att fortsätta arbeta och orga- nisationers kapabilitet att realisera olika former av age management praktiker, samt sam- spelet dem emellan. 70 forskare ger felaktiga policyrekommendationer om pensionspreferenser används som ut- fall istället för faktiskt utträde. Artikel II: Arbetsanpassningar för att stödja äldre arbetstagare med dålig hälsa I Artikel II undersöktes om dålig hälsa i form av tidigare sjukfrånvaro, låg självskattad hälsa eller långvarig sjukdom har ett samband med en lägre arbetsförmåga och tidigare utträde från arbetslivet. Dålig hälsa visade sig ha en negativ effekt på fysisk och psykisk arbetsförmåga samt ökar risken för tidigare arbetsmarknadsutträde. I artikeln undersöktes även antagandet att en hög grad av handlingsutrymme ökar arbetstagarens möjligheter att anpassa arbetet efter förmåga och önskemål. Med handlingsutrymme avsågs i detta fall frihet att bestämma över sin arbetstid, arbetstakt och planering, samt möjligheter till in- flytande över när, var och hur arbetet utförs. Resultaten visar att dålig hälsa i form av långvarig sjukdom, sjukfrånvaro eller låg självskattad hälsa, är viktiga faktorer som tidiga- relägger arbetstagares arbetsmarknadsutträde. Vidare visar resulta tet att ha dålig hälsa i form av långvarig sjukdom och sjukfrånvaro i kombination med lågt handlingsutrymme ökar risken för tidig pension vilket kan indikera en interaktionseffekt. Samma resultat för självskattad hälsa var däremot inte lika konsistent. Med andra ord kan möjlighet till ar- betsanpassning vid ohälsa öka äldre arbetstagares möjligheter att fortsätta arbeta. Dessa samband bör emellertid studeras vidare med mer validerade mätinstrument. Artikel III : Betydelsen av kompetensmatchning för pensionspr eferenser I Artikel III undersöktes betydelsen av individuella arbetsöverenskommelser mellan an- ställd och arbetsgivare. Individuella arbetsöverenskommelser avsåg bland annat arbets- uppgifter och ansvar, platsflexibilitet, minskad arbetsbelastning, finansiel la incitament och schemaflexibilitet för äldre arbetstagares pensionspreferenser. Resultaten pekar på att ar- betsöverenskommelser avseende arbetsuppgifter och ansvar senarelägger anställdas öns- kemål om pension. Intressant nog visade resultaten även på ett m otsatt förhållande för arbetsöverenskommelser rörande mindre arbetsbelastning. Troligtvis beror det på att de som har fått reducerad arbetsbelastning är samma arbetstagare som lider av sämre hälsa och därmed även har tidigare pensionspreferenser. Resultate t indikerar också att det finns utrymme för ytterligare arbetsanpassningar utöver nedsatt arbetsbelastning. Studien visar även att omfattningen av individuella överenskommelser är begränsad , särskilt bland kvin- nor, äldre, arbetstagare med dålig hälsa och i lägre socioekonomisk position. Efter datain- samling och analys genomfördes en valideringsanalys av mätinstrumentet som indikerade att individuella arbetsöverenskommelser kan förväxlas med arbetsanpassningar och flexi- bilitet som redan existerar på arbetsplatsen eller erbjuds av arbetsgivaren. Den svenska kontexten med starka kollektivavtal, kollektivt tankesätt och personalpolicys riktade till aggregerade grupper, särskilt i välfärdsorganisat ioner, är några förklaringar till den begrän- sade omfattningen av arbetsöverenskommelser men också de delvis svårtolkade resultaten i studien. Slutsatserna från artikeln är att arbetsöverenskommelser som berör arbetsupp- gifter och ansvar har betydelse för ar betstagarnas önskemål om tidigare eller senarelägg- 71 ning av pension. Detta innebär att en potentiell policyåtgärd är att säkerställa god kompe- tensmatchning. Frågan om arbetsöverenskommelsernas roll och betydelse i svensk kon- text behöver emellertid djupare te oretisk och empirisk analys i framtida forskning. Artikel IV: Organisatoriska hinder att behålla äldre anställda I Artikel IV, intervjuades linjechefer och deras HR stöd om hinder och möjligheter att behålla äldre anställda inom vård - och sjukvård. Genere llt uppvisade både linjechefer och HR personal en positiv attityd till äldre medarbetare och framhävde deras betydelse som mentorer för yngre. Dock var implementeringen av age-management -strategier ytterst be- gränsad och begreppet age management användes in te i de undersökta organisationerna. De organisatoriska hinder för att organisationen skulle behålla äldre anställda identifiera- des som linjechefernas höga arbetsbelastning, frånvaro av flexibla personalpolicys, under- utvecklade age-management strategier oc h dålig koordination inom organisationen. Dessa organisatoriska hinder tycks förstärka varandra i en ond cirkel, exempelvis innebar de stora personalbehoven och det administrativa arbete att linjecheferna tvingades fokusera på den dagliga driften vilket to g tid från strategiskt utvecklingsarbete med arbetsmetoder och personalpolicys. Avsaknaden av nya arbetsmetoder eller policys resulterade i frånvaro av verktyg och incitament för linjechefer i arbetet med att behålla äldre arbetstagare. Ef- tersom linjechefe rna redan befinner sig i en ansträngd arbetssituation leder det till att frå- gan om att bibehålla äldre anställda riskerar att bli beroende av individuella linjechefens vilja och attityder till äldre medarbetare och därmed godtycklig ur ett strategiskt pers pek- tiv. Linjechefer och HR personal beskrev också svårigheterna med att hantera individuella önskemål eller behov utifrån existerande personalpolicys och avtal, men också på grund av starka likabehandlingsideal, exempelvis uttrycktes svårigheter att indivi danpassa sche- maläggning eller arbetsuppgifter samt begränsade möjligheter att omplacera medarbetare efter rehabilitering. Sammantaget leder dessa förhållanden till att vård - och omsorgsorga- nisationer inte lyckas tillvarata den potentiella arbetskraftsresurs som äldre medarbetare utgör. För att kunna behålla den arbetskraftsresurs som äldre anställda utgör måste orga- nisationerna bli mer proaktiva och koordinerade på både strategisk - och operativ nivå. Detta innebär att högre ledning måste lyfta fram betydels en av att behålla äldre medarbe- tare och allokera tillräckliga resurser så att linjechefer får en balanserad arbetssituation för att kunna prioritera frågan om äldre medarbetare. Diskussion Avhandlingen undersökte hur äldre medarbetares arbetsmarknadsdelt agande formas via ett samspel mellan å ena sidan personliga förmågor och preferenser och å andra sidan förutsättningar på arbetsplats - och organisationsnivå. Centralt i denna diskussion är ka- pabilitetsbegreppet , vilket innefattar både individers kapabilite t att fortsätta arbeta och orga- nisationers kapabilitet att realisera olika former av age management praktiker, samt sam- spelet dem emellan. 72 Vilka personliga och arbetsrelaterade faktorer främjar eller hindrar äldre arbetstagare från att kunna och vilja for tsätta arbeta ? I de kvantitativa studierna visar resultatet att det finns ett antal grundläggande faktorer som har betydelse för äldre arbetstagares tidigareläggning av pens- ion; dålig hälsa, låg arbetstrivsel, hög fysisk arbetsbelastning och lägre arbetsfö rmåga. Dessa samband kan indikera en obalans mellan personers förmåga och önskemål i relation till kraven och resurser i deras nuvarande arbetssituation. Resultaten indikerar också, i linje med annan svensk statistik (Johansson et al., 2018) , att det skett en förskjutning i pens- ionsålder de senaste två decennierna. Individuella arbetsöverenskommelser mellan arbets- tagare och arbetsgivare rörande arbetsuppgifter och ansvar har betydelse för välfärdsan- ställdas pensionspreferenser. Dessutom pekar resultatet på att individer med låg självskat- tad hälsa, långvarig sjukdom eller lång sjukskrivning har högre risk för tidigare pensions- utträde vid frånvaro av handlingsutrymme i arbetet. Förutsättningarna för ett längre ar- betsliv är sämre för de äldre arbetstagare som har låg arbetsförmåga och som arbetar i yrkesgrupper med lägre handlingsutrymme (Artikel II) , har lägre möjligheter att förhandla arbetsöverenskommelser (Artikel III ) eller arbetar för en arbetsgivare med fr ånvaro av personalpolicys riktade till äldre anställda (Artikel IV). Detta gäller även för de arbetstagare som i sitt arbete exponeras för hög fysisk arbetsbelastning eller saknar arbetstillfredsstäl- lelse i sitt nuvarande arbete (Artikel I). Avhandlingens resultat visar samtidigt att hand- lingsutrymme i arbete och individuella överenskommelser varierar mellan olika demo- grafiska grupper, exempelvis var tillgången till individuella arbetsöverenskommelser inom offentliga välfärdsorganisationer lägre bland kvinn or, arbetstagare i lägre socioekonomisk position, högre ålder, med dålig hälsa och kortare anställning inom organisationen. Dess- utom tycks graden av handlingsutrymme vara lägre bland kvinnor och de arbetstagarna med lägre socioekonomisk position (se tabell i engelska resultatdelen för Artikel II) . Mot bakgrund av dessa resultat visar avhandlingen att en ojämn fördelning av arbetsrelaterade krav såsom fysisk exponering men också arbetsresurser såsom individuella arbetsöverens- kommelser eller handlingsutrymme kan ge olika möjligheter att förlänga arbetslivet. Vilka hinder och möjligheter påverkar offentliga välfärdsorganisationers kapabilitet att behålla äldre medarbetare? Resultat från avhandlingens kvalitativa studie visade att de undersökta offent- liga välf ärdsorganisationerna saknade särskilda åtgärder eller policys riktade till äldre an- ställda. Frånvaron av arbetsgivares engagemang i åldersrelaterade personalpolicys har in- direkt manifesteras i tidigare forskning vilket kan indikera en diskrepans mellan forskning och praktik. Trots att linjechefer och deras HR -stöd uppgav att den äldre arbetskraften var en tillgång, saknades de organisatoriska förutsättningarna för att kunna arbeta proak- tivt med att behålla de äldre medarbetarna. På en organisatorisk nivå sa knades dessutom koordination mellan nivåer och faktiska policys vilket innebar att ansvaret för att hantera och behålla den äldre arbetskraften förflyttades till lägre linjechefer vars arbetssituation redan var ansträngd. Linjechefernas situation kunde lik nas vid en ond cirkel där hög ar- betsbelastning, frånvaro av strategisk planering, universella policys och praktiker samt otillräckliga incitament hade en synergieffekt vilket skapade en situation där arbetskraft lämnade organisationerna utan någon insats. Eftersom en kedja aldrig är starkare än sin 73 svagaste länk, förutsätter implementering av åldersmedvetenhet i policys att organisation- ens olika funktioner är koordinerade både i avseende på prioritet (preferenser) men också allokering av resurser (tillräckl iga resurser för att nå mål). Utifrån Kadefors et al. (2020) typlogi om organisatorisk kapabilitet kan därför de undersökta organisationerna både ka- tegoriseras som passiva med avseende på den strategiska personalfunktionen (HR) men också bakbunden (engelska: chained) i relation till linjechefernas arbetssituation. Detta pekar samtidigt på att organisationers kapabilitet är ett resultat av dess sammansatta ope- rativa verksamhet. För att bryta denna trend skulle de offentliga välfärdsorganisationer som deltog i delstudien behöva att den högre ledningen sanktionerar en strategi för att behålla äldre medarbetare som är överenskommen och koordinerad mellan olika organi- satoriska nivåer. För att motivera den högre ledningen, som har en komplex verksamhets- beskrivning och begränsat budgetutrymme, kan företrädare för HR eller HR avdelni ngen behöva erbjuda ett beslutsunderlag i form av en kostnads - nyttoanalys och praktiska ex- empel på implementering i andra verksamheter eller företag (se exempelvis, Malmqvist, 2007) . S å hur samspelar individer och organisationers k apabilitet i kontexten av ett längre arbetsliv? Den teoretiska modell som presenterades i avhandlingen erbjuder ett verktyg för att identifiera särskilt starka och svaga positioner eller grupper på arbetsmarknaden. I modellen innebär det att de individer som har låg arbetsförmåga och låg vilja att fortsätta arbeta antas ha ett bättre utgångsläge i en organisation med tillgängliga resurser och vilja att behålla äldre anställda än vice versa. Empiriskt styrks detta av avhandlingens delstudier som visar att individers förutsättningar för fortsatt arbetsdeltagande skapas i samspelet mellan organi- satorisk och individuell kapabilitet. Individers förutsättningar för ett längre arbetsliv är särskilt tydliga när både individuell och organisatorisk kapabilitet är låg. Så vilka yrkes- grupper är mer troliga att vara anställda i en passiv organisation och samtidigt ha en låg arbetsförmåga? För att exemplifiera är gruppen undersköterskor som är en relativt stor yrkesgrupp på den svenska arbetsmarknaden och som till stor del utgörs av kvinnor och anställda inom offentliga välfärdsorganisationer som undersöktes i Artikel IV och till viss del i Artikel III . Arbetet består inte sällan av tunga lyft, långa skift och pressade scheman där utförandet av vården regleras i en tydlig b eslutshierarki inom organisationen. Under- sköterskegruppen är överrepresenterade i sjukfrånvaro och uppvisar tidigare pensions- mönster än många andra yrkesgrupper. Utifrån resultatet kategoriserades de offentliga välfärdsorganisationerna som passiva när det kommer till policys och praktiker för den äldre arbetskraften, vilket innebär begränsade möjligheter att erbjuda flexibilitet och ar- betsanpassningar efter äldre medarbetares behov och önskemål enligt age management imperativet. Utöver de organisatoriska hi ndren finns goda skäl att anta att gruppen under- sköterskor samtidigt har lägre handlingsutrymme i deras arbete och mindre möjligheter till individuella arbetsöverenskommelser vilka försvagar möjligheterna till att anpassa ar- betet efter individens förmåga o ch önskemål. Sammantaget belyser detta att individers kapabilitet är kontextberoende och bestäms i relation till arbetsplatsspecifika förhållanden, 72 Vilka personliga och arbetsrelaterade faktorer främjar eller hindrar äldre arbetstagare från att kunna och vilja for tsätta arbeta ? I de kvantitativa studierna visar resultatet att det finns ett antal grundläggande faktorer som har betydelse för äldre arbetstagares tidigareläggning av pens- ion; dålig hälsa, låg arbetstrivsel, hög fysisk arbetsbelastning och lägre arbetsfö rmåga. Dessa samband kan indikera en obalans mellan personers förmåga och önskemål i relation till kraven och resurser i deras nuvarande arbetssituation. Resultaten indikerar också, i linje med annan svensk statistik (Johansson et al., 2018) , att det skett en förskjutning i pens- ionsålder de senaste två decennierna. Individuella arbetsöverenskommelser mellan arbets- tagare och arbetsgivare rörande arbetsuppgifter och ansvar har betydelse för välfärdsan- ställdas pensionspreferenser. Dessutom pekar resultatet på att individer med låg självskat- tad hälsa, långvarig sjukdom eller lång sjukskrivning har högre risk för tidigare pensions- utträde vid frånvaro av handlingsutrymme i arbetet. Förutsättningarna för ett längre ar- betsliv är sämre för de äldre arbetstagare som har låg arbetsförmåga och som arbetar i yrkesgrupper med lägre handlingsutrymme (Artikel II) , har lägre möjligheter att förhandla arbetsöverenskommelser (Artikel III ) eller arbetar för en arbetsgivare med fr ånvaro av personalpolicys riktade till äldre anställda (Artikel IV). Detta gäller även för de arbetstagare som i sitt arbete exponeras för hög fysisk arbetsbelastning eller saknar arbetstillfredsstäl- lelse i sitt nuvarande arbete (Artikel I). Avhandlingens resultat visar samtidigt att hand- lingsutrymme i arbete och individuella överenskommelser varierar mellan olika demo- grafiska grupper, exempelvis var tillgången till individuella arbetsöverenskommelser inom offentliga välfärdsorganisationer lägre bland kvinn or, arbetstagare i lägre socioekonomisk position, högre ålder, med dålig hälsa och kortare anställning inom organisationen. Dess- utom tycks graden av handlingsutrymme vara lägre bland kvinnor och de arbetstagarna med lägre socioekonomisk position (se tabell i engelska resultatdelen för Artikel II) . Mot bakgrund av dessa resultat visar avhandlingen att en ojämn fördelning av arbetsrelaterade krav såsom fysisk exponering men också arbetsresurser såsom individuella arbetsöverens- kommelser eller handlingsutrymme kan ge olika möjligheter att förlänga arbetslivet. Vilka hinder och möjligheter påverkar offentliga välfärdsorganisationers kapabilitet att behålla äldre medarbetare? Resultat från avhandlingens kvalitativa studie visade att de undersökta offent- liga välf ärdsorganisationerna saknade särskilda åtgärder eller policys riktade till äldre an- ställda. Frånvaron av arbetsgivares engagemang i åldersrelaterade personalpolicys har in- direkt manifesteras i tidigare forskning vilket kan indikera en diskrepans mellan forskning och praktik. Trots att linjechefer och deras HR - stöd uppgav att den äldre arbetskraften var en tillgång, saknades de organisatoriska förutsättningarna för att kunna arbeta proak- tivt med att behålla de äldre medarbetarna. På en organisatorisk nivå sa knades dessutom koordination mellan nivåer och faktiska policys vilket innebar att ansvaret för att hantera och behålla den äldre arbetskraften förflyttades till lägre linjechefer vars arbetssituation redan var ansträngd. Linjechefernas situation kunde lik nas vid en ond cirkel där hög ar- betsbelastning, frånvaro av strategisk planering, universella policys och praktiker samt otillräckliga incitament hade en synergieffekt vilket skapade en situation där arbetskraft lämnade organisationerna utan någon insats. Eftersom en kedja aldrig är starkare än sin 73 svagaste länk, förutsätter implementering av åldersmedvetenhet i policys att organisation- ens olika funktioner är koordinerade både i avseende på prioritet (preferenser) men också allokering av resurser (tillräckl iga resurser för att nå mål). Utifrån Kadefors et al. (2020) typlogi om organisatorisk kapabilitet kan därför de undersökta organisationerna både ka- tegoriseras som passiva med avseende på den strategiska personalfunktionen (HR) men också bakbunden (engelska: chained) i relation till linjechefernas arbetssituation. Detta pekar samtidigt på att organisationers kapabilitet är ett resultat av dess sammansatta ope- rativa verksamhet. För att bryta denna trend skulle de offentliga välfärdsorganisationer som deltog i delstudien behöva att den högre ledningen sanktionerar en strategi för att behålla äldre medarbetare som är överenskommen och koordinerad mellan olika organi- satoriska nivåer. För att motivera den högre ledningen, som har en komplex verksamhets- beskrivning och begränsat budgetutrymme, kan företrädare för HR eller HR avdelni ngen behöva erbjuda ett beslutsunderlag i form av en kostnads - nyttoanalys och praktiska ex- empel på implementering i andra verksamheter eller företag (se exempelvis, Malmqvist, 2007) . S å hur samspelar individer och organisationers k apabilitet i kontexten av ett längre arbetsliv? Den teoretiska modell som presenterades i avhandlingen erbjuder ett verktyg för att identifiera särskilt starka och svaga positioner eller grupper på arbetsmarknaden. I modellen innebär det att de individer som har låg arbetsförmåga och låg vilja att fortsätta arbeta antas ha ett bättre utgångsläge i en organisation med tillgängliga resurser och vilja att behålla äldre anställda än vice versa. Empiriskt styrks detta av avhandlingens delstudier som visar att individers förutsättningar för fortsatt arbetsdeltagande skapas i samspelet mellan organi- satorisk och individuell kapabilitet. Individers förutsättningar för ett längre arbetsliv är särskilt tydliga när både individuell och organisatorisk kapabilitet är låg. Så vilka yrkes- grupper är mer troliga att vara anställda i en passiv organisation och samtidigt ha en låg arbetsförmåga? För att exemplifiera är gruppen undersköterskor som är en relativt stor yrkesgrupp på den svenska arbetsmarknaden och som till stor del utgörs av kvinnor och anställda inom offentliga välfärdsorganisationer som undersöktes i Artikel IV och till viss del i Artikel III . Arbetet består inte sällan av tunga lyft, långa skift och pressade scheman där utförandet av vården regleras i en tydlig b eslutshierarki inom organisationen. Under- sköterskegruppen är överrepresenterade i sjukfrånvaro och uppvisar tidigare pensions- mönster än många andra yrkesgrupper. Utifrån resultatet kategoriserades de offentliga välfärdsorganisationerna som passiva när det kommer till policys och praktiker för den äldre arbetskraften, vilket innebär begränsade möjligheter att erbjuda flexibilitet och ar- betsanpassningar efter äldre medarbetares behov och önskemål enligt age management imperativet. Utöver de organisatoriska hi ndren finns goda skäl att anta att gruppen under- sköterskor samtidigt har lägre handlingsutrymme i deras arbete och mindre möjligheter till individuella arbetsöverenskommelser vilka försvagar möjligheterna till att anpassa ar- betet efter individens förmåga o ch önskemål. Sammantaget belyser detta att individers kapabilitet är kontextberoende och bestäms i relation till arbetsplatsspecifika förhållanden, 74 det vill säga organisatorisk kapabilitet. Avhandlingens resultat visar att det krävs ett bre- dare perspektiv för att fullt ut kunna förstå individers förutsättningar att förlänga arbets- livet. Med andra ord bör såväl individers kapabilitet (förmåga, vilja), organisationers ka- pabilitet (resurser, preferenser) , och samspele t dem emellan, tas i beaktande. Avslutningsvis belyser avhandlingen betydelsen av arbetskvalité, hållbarhet och huruvida möjligheter att förlänga arbetet är jämlikt mellan olika grupper. Arbetsgivare kan under- stödja hälsa och välbefinnande, höja kvalitén i arbete, främja arbetsförmåga ge nom ar- betsanpassningar, kompetensmatchning och handlingsutrymme i arbetet. Organisatorer kan bevaka äldre anställdas arbetsförmåga och motivation i exempelvis uppföljningsinter- vjuer samt sätta in åtgärder där behoven är som störst så att den enskildes förm åga och önskemål matchar arbetets krav och resurser. Ur ett arbetsmiljöperspektiv behöver de åtgärder som är riktade till äldre inte vara väsensskilda från systematiskt arbetsmiljöarbetet för att förebygga ohälsosamma arbetsmiljöer och skydda anställdas hä lsa och välbefin- nande (SAM). Exempelvis kan insatser som berör anställdas arbetsförmåga och motivat- ion inkorporeras i existerande policys. Det förutsätter emellertid samtidigt att organisat- ionens funktioner är koordinerade och att linjechefer, HR avdelning samt företagshälso- vård har tillräckliga resurser och stöd från högre ledning att prioritera ett förlängt arbetsliv för den äldre arbetskraften. 75 Appendix A Interview guide (t ranslated from Swedish) ARBETSKRAFT I SKÅNE OCH GOTHENB URG Introduction: During the autumn and spring, you participated in the ArbetsK raft project in Skåne or the city of Gothenburg. The participants in the project gathered and participated in meetings ( “ intervention” ) between small business owners, managers, and HR staff and listened to some experts on various topics. The goal of the project w as to investigate how organiza- tions create the conditions for sustainable workplaces for all ages. Inform about: formal consent, anonymity, recording, voluntary participation, take time to think and pause, thanks for participating Warm - up questions: Where do you work? Can you describe what you do at your job? What is your title ( e.g., l ine manager or HR manager) ? Organization: What are the reasons, incentives , or motivations for you to continue to work with measures that enable older employees to remain in the workplace?  How do you get support from your organization or management in how to adapt work for older employees? (… ↠) Can you describe a little more about what you think could support and facilitate your efforts to adapt work for older employees (e.g., knowledge, financial condi- tions, and practical tools) ?  Have you felt or experienced any understanding and awareness in your organization of changes occurring in employees with increasing age and how they affect work and private life? Can you cite an example? Has this awareness been affected by your participation in Arbets K raft? If so, h ow?  What active measures are being implemented in your organization to help employees be able and willing to work until age 67 years or longer? H as your participation in Ar- betsK raft affected your commitment in any way? How ha s ArbetsK raft in practice been useful in the organization?  What do you think are the most important factors in enabling older people to remain in your organization? Feel free to reflect on how you think your organization differs from so- ciety in general.  What are the opportunities for older employees to have special accommodations to cope with tasks or to motivate them? ( e.g., in terms of time, location , or working pace) 74 det vill säga organisatorisk kapabilitet. Avhandlingens resultat visar att det krävs ett bre- dare perspektiv för att fullt ut kunna förstå individers förutsättningar att förlänga arbets- livet. Med andra ord bör såväl individers kapabilitet (förmåga, vilja), organisationers ka- pabilitet (resurser, preferenser) , och samspele t dem emellan, tas i beaktande. Avslutningsvis belyser avhandlingen betydelsen av arbetskvalité, hållbarhet och huruvida möjligheter att förlänga arbetet är jämlikt mellan olika grupper. Arbetsgivare kan under- stödja hälsa och välbefinnande, höja kvalitén i arbete, främja arbetsförmåga ge nom ar- betsanpassningar, kompetensmatchning och handlingsutrymme i arbetet. Organisatorer kan bevaka äldre anställdas arbetsförmåga och motivation i exempelvis uppföljningsinter- vjuer samt sätta in åtgärder där behoven är som störst så att den enskildes förm åga och önskemål matchar arbetets krav och resurser. Ur ett arbetsmiljöperspektiv behöver de åtgärder som är riktade till äldre inte vara väsensskilda från systematiskt arbetsmiljöarbetet för att förebygga ohälsosamma arbetsmiljöer och skydda anställdas hä lsa och välbefin- nande (SAM). Exempelvis kan insatser som berör anställdas arbetsförmåga och motivat- ion inkorporeras i existerande policys. Det förutsätter emellertid samtidigt att organisat- ionens funktioner är koordinerade och att linjechefer, HR avdelning samt företagshälso- vård har tillräckliga resurser och stöd från högre ledning att prioritera ett förlängt arbetsliv för den äldre arbetskraften. 75 Appendix A Interview guide (t ranslated from Swedish) ARBETSKRAFT I SKÅNE OCH GOTHENB URG Introduction: During the autumn and spring, you participated in the ArbetsK raft project in Skåne or the city of Gothenburg. The participants in the project gathered and participated in meetings ( “ intervention” ) between small business owners, managers, and HR staff and listened to some experts on various topics. The goal of the project w as to investigate how organiza- tions create the conditions for sustainable workplaces for all ages. Inform about: formal consent, anonymity, recording, voluntary participation, take time to think and pause, thanks for participating Warm -up questions: Where do you work? Can you describe what you do at your job? What is your title ( e.g., l ine manager or HR manager) ? Organization: What are the reasons, incentives , or motivations for you to continue to work with measures that enable older employees to remain in the workplace?  How do you get support from your organization or management in how to adapt work for older employees? (… ↠) Can you describe a little more about what you think could support and facilitate your efforts to adapt work for older employees (e.g., knowledge, financial condi- tions, and practical tools) ?  Have you felt or experienced any understanding and awareness in your organization of changes occurring in employees with increasing age and how they affect work and private life? Can you cite an example? Has this awareness been affected by your participation in Arbets K raft? If so, h ow?  What active measures are being implemented in your organization to help employees be able and willing to work until age 67 years or longer? Has your participation in Ar- betsK raft affected your commitment in any way? How ha s ArbetsK raft in practice been useful in the organization?  What do you think are the most important factors in enabling older people to remain in your organization? Feel free to reflect on how you think your organization differs from so- ciety in general.  What are the opportunities for older employees to have special accommodations to cope with tasks or to motivate them? ( e.g., in terms of time, location , or working pace) 76 What types of physical work environment measures have been taken in the organiza- tion to make it easier for older employees? What types of psychosocial work environment measures are available in the organization to help older employees?  Are there any obstacles in your organization that make it difficult for you to work with age- related work accommodations? If so, w hat and how?  What focus on older employees did your organization have before ArbetsK raft be- gan? Has it developed during or after you participated in ArbetsK raft?  How do you intend to proceed with older people employees’ participation in your organiza- tion? Is there any current issue or problem in the organization that you want to continue working on?  How has your organization spread the knowledge from Arbets K raft ( e.g., via family, friends, coffee rooms, lectures, entrepreneurs, HR networks, management networks, and other formal and informal knowledge transfer situations)?  Reflect on and say whether you believe that ArbetsK raft may have contributed to sus- tainable working life for all ages during the (1) project pe riod and in (2) the long term , for example, via new knowledge and attitude changes? Society and the individual  How do you think that ArbetsKraft can contribute to sustainable working life for all ages in society? (a) During the project period and (b) in the long term through the new knowledge and attitude changes, for instance?  What did you as an individual get out of participating in Arbets K raft (at work and lei- sure)? What did you get at an abstract level out of participation in Arbets K raft, i.e. , the experience of participating? What did you get in more concrete/practical terms from participating— for ex- ample, have you implemented any actions/ changes to which Arbets K raft could have contributed via “ new thinking” ?  Have you shared the knowledge gained from Arbets K raft (in your own assessment of your personal situation, family, friends, coffee room, lecture s, APT, entrepreneurs, HR network manager’ s network, etc., i.e., formal and informal) ?  Do you have any other comments or thoughts you would like to share? 77 References Abma, F. I., Brouwer, S., de Vries, H. J., Arends, I., Robroek, S. J., Cuijpers, M. P., van der Wilt, G. J., Bültmann, U. & van der Klink, J. J. (2016). The capability set for work: development and validation of a new questionnaire. Scandinavian Journal of Work, Environment & Health, 42 , 34 - 42. Ahlberg, O., Almérus, A., Engström, J., Gustavsson, H., Hauer, E., Nyberg, P. & Tingvall, P. (2019). Var finns jo bben? Bedömning för 2019 och på fem års sikt [Where are the jobs? Assessment for 2019 and in five years' time]. Stockholm Arbetsförmedlingen. Ahmed, A. M. , Andersson, L. & Hammarstedt, M. (2012). Does age matter for employability? A field experiment on age ism in the Swedish labour market. Applied Economics Letters, 19 , 403 - 406. Aronsson, G., Marklund, S., Leineweber, C. & Helgesson, M. (2021). The changing nature of work– Job strain, job support and sickness absence among care workers and in other occupations in Sweden 1991 – 2013. S S M -Population Health, 15 , 100893. Axelrad, H. & Mahoney, K. J. (2017). Increasing the Pensionable Age: What Changes Are OECD Countries Making? What Considerations Are Driving Policy? Open Journal of Social Sciences, 5 , 56. Bal, P. M ., De Jong, S. B., Jansen, P. G. & Bakker, A. B. (2012). Motivating employees to work beyond retirement: A multi ‐level study of the role of I ‐deals and unit climate. Journal of Management Studies, 49 , 306 - 331. Bal, P. M., Kooij, D. T. & Rousseau, D. M. (20 15). Aging workers and the employee -employer relationship, New York, Springer. Baltes, P. B. & Baltes, M. M. (1993). Successful aging: Perspectives from the behavioral sciences, New York, Cambridge University Press. Becker, G. S. (1965). A Theory of the Al location of Time. The economic journal, 75 , 493 - 517. Bengtsson, H. & Melke, A. (2014). V år offentliga förvaltning. Samverkan i välfärdspolitiken [Our public administration. Collaboration in welfare policy], Malmö, Gleerups Utbildning. Bengtsson, M. & Flisb äck, M. (2017). On leaving work as a calling: retirement as an existential imperative. International Journal of Ageing and Later Life, 11 , 37 - 67. Bengtsson, T. & Scott, K. (2011). Population aging and the future of the welfare state: The example of Sweden. Population and Development Review, 37 , 158 - 170. Berglund, T., Seldén, D. & Halleröd, B. (2017). Factors affecting prolonged working life for the older workforce: the Swedish case. Nordic Journal of Working Life Studies, 7 , 19 – 36. Blau, P. (1964). Exchange and power in social life, New York, Routledge. Blauner, R. (1964). Alienation and freedom: The factory worker and his industry, Oxford, England, Chicago University Press. Boissonneault, M. & de Beer, J. (2018). Work ability trajectories and retire ment pathways: a longitudinal analysis of older American workers. Journal of Occupational and Environmental Medicine, 60 , e343 - e348. Box - Steffensmeier, J. M. & Jones, B. S. (2004). Event history modeling: A guide for social scientists, New York, Cambridge University Press. 76 What types of physical work environment measures have been taken in the organiza- tion to make it easier for older employees? What types of psychosocial work environment measures are available in the organization to help older employees?  Are there any obstacles in your organization that make it difficult for you to work with age- related work accommodations? If so, w hat and how?  What focus on older employees did your organization have before ArbetsK raft be- gan? Has it developed during or after you participated in ArbetsK raft?  How do you intend to proceed with older people employees’ participation in your organiza- tion? Is there any current issue or problem in the organization that you want to continue working on?  How has your organization spread the knowledge from Arbets K raft ( e.g., via family, friends, coffee rooms, lectures, entrepreneurs, HR networks, management networks, and other formal and informal knowledge transfer situations)?  Reflect on and say whether you believe that ArbetsK raft may have contributed to sus- tainable working life for all ages during the (1) project pe riod and in (2) the long term , for example, via new knowledge and attitude changes? Society and the individual  How do you think that ArbetsKraft can contribute to sustainable working life for all ages in society? (a) During the project period and (b) in the long term through the new knowledge and attitude changes, for instance?  What did you as an individual get out of participating in Arbets K raft (at work and lei- sure)? What did you get at an abstract level out of participation in Arbets K raft, i.e. , the experience of participating? What did you get in more concrete/practical terms from participating— for ex- ample, have you implemented any actions/ changes to which Arbets K raft could have contributed via “ new thinking” ?  Have you shared the knowledge gained from Arbets K raft (in your own assessment of your personal situation, family, friends, coffee room, lecture s, APT, entrepreneurs, HR network manager’ s network, etc., i.e., formal and informal) ?  Do you have any other comments or thoughts you would like to share? 77 References Abma, F. I., Brouwer, S., de Vries, H. J., Arends, I., Robroek, S. J., Cuijpers, M. P., van der Wilt, G. J., Bültmann, U. & van der Klink, J. J. (2016). The capability set for work: development and validation of a new questionnaire. Scandinavian Journal of Work, Environment & Health, 42 , 34-42. Ahlberg, O., Almérus, A., Engström, J., Gustavsson, H., Hauer, E., Nyberg, P. & Tingvall, P. (2019). Var finns jo bben? Bedömning för 2019 och på fem års sikt [Where are the jobs? Assessment for 2019 and in five years' time]. Stockholm Arbetsförmedlingen. Ahmed, A. M. , Andersson, L. & Hammarstedt, M. (2012). Does age matter for employability? A field experiment on age ism in the Swedish labour market. Applied Economics Letters, 19 , 403 -406. Aronsson, G., Marklund, S., Leineweber, C. & Helgesson, M. (2021). The changing nature of work– Job strain, job support and sickness absence among care workers and in other occupations in Sweden 1991 – 2013. S S M -Population Health, 15 , 100893. Axelrad, H. & Mahoney, K. J. (2017). Increasing the Pensionable Age: What Changes Are OECD Countries Making? What Considerations Are Driving Policy? Open Journal of Social Sciences, 5 , 56. Bal, P. M ., De Jong, S. B., Jansen, P. G. & Bakker, A. B. (2012). Motivating employees to work beyond retirement: A multi ‐level study of the role of I ‐deals and unit climate. Journal of Management Studies, 49 , 306-331. Bal, P. M., Kooij, D. T. & Rousseau, D. M. (20 15). Aging workers and the employee -employer relationship, New York, Springer. Baltes, P. B. & Baltes, M. M. (1993). Successful aging: Perspectives from the behavioral sciences, New York, Cambridge University Press. Becker, G. S. (1965). A Theory of the Al location of Time. The economic journal, 75 , 493-517. Bengtsson, H. & Melke, A. (2014). V år offentliga förvaltning. Samverkan i välfärdspolitiken [Our public administration. Collaboration in welfare policy], Malmö, Gleerups Utbildning. Bengtsson, M. & Flisb äck, M. (2017). On leaving work as a calling: retirement as an existential imperative. International Journal of Ageing and Later Life, 11 , 37-67. Bengtsson, T. & Scott, K. (2011). Population aging and the future of the welfare state: The example of Sweden. Population and Development Review, 37 , 158 -170. Berglund, T., Seldén, D. & Halleröd, B. (2017). Factors affecting prolonged working life for the older workforce: the Swedish case. Nordic Journal of Working Life Studies, 7, 19– 36. Blau, P. (1964). Exchange and power in social life, New York, Routledge. Blauner, R. (1964). Alienation and freedom: The factory worker and his industry, Oxford, England, Chicago University Press. Boissonneault, M. & de Beer, J. (2018). Work ability trajectories and retire ment pathways: a longitudinal analysis of older American workers. Journal of Occupational and Environmental Medicine, 60 , e343 - e348. Box -Steffensmeier, J. M. & Jones, B. S. (2004). Event history modeling: A guide for social scientists, New York, Cambridge University Press. 78 Braun, V. & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3 , 77-101. Braverman, H. (1998) . Labor and monopoly capital: The degradation of work in the twentieth century, New York, NYU Press. Browne, P., Carr, E., Fleischmann, M., Xue, B. & Stansfeld, S. A. (2019). The relationship between workplace psychosocial environment and retirement intentions and actual retirement: a systematic review. European Journal of Ageing, 16 , 73-82. Brunello, G. & Wruuck, P. (2019). Skill shortages and skill mismatch in Europe: a review of the literature. Bonn: IZA institute of labor economics. Byrne, B. M. (2013). Structural equation modeling with Mplus: Basic concepts, applications, and programming, New York, Ro utledge. Carr, E., Hagger -Johnson, G., Head, J., Shelton, N., Stafford, M., Stansfeld, S. & Zaninotto, P. (2016). Working conditions as predictors of retirement intentions and exit from paid employment: a 10 -year follow -up of the English Longitudinal Study of Ageing. European Journal of Ageing, 13 , 39- 48. Carstensen, L. L. (1995). Evidence for a life -span theory of socioemotional selectivity. Current directions in Psychological science, 4 , 151-156. Cerdas, S., Härenstam, A., Johansson, G. & Nyberg, A. (2019 ). Development of job demands, decision authority and social support in industries with different gender composition– Sweden, 1991 – 2013. B MC Public Health, 19 , 1-16. Cloostermans, L., Bekkers, M. B., Uiters, E. & Proper, K. I. (2015). The effectiveness of interventions for ageing workers on (early) retirement, work ability and productivity: a systematic review. International Archives of Occupational and Environmental Health, 88 , 521 -532. Conen, W., Henkens, K. & Schippers, J. (2012). Employers’ attitudes and actions towards the extension of working lives in Europe. International Journal of Manpower, 33 , 648-66 5 . Conen, W., Henkens, K. & Schippers, J. (2014). Ageing organisations and the extension of working lives: A case study approach. Journal of Social Poli cy, 43 , 773-792. Conen, W. S., Henkens, K. & Schippers, J. J. (2011). Are employers changing their behavior toward older workers? An analysis of employers' surveys 2000 – 2009. Journal of Aging & Social Policy, 23 , 141-158. Corin, L. & Björk, L. (2016). Job demands and job resources in human service managerial work an external assessment through work content analysis. Nordic Journal of Working Life Studies, 6 , 3 -28. Corin, L. , Pousette, A., Berglund, T., Dellve, L., Hensing, G. & Bj örk, L. (2021). Occupational trajectories of working conditions in Sweden: Development trends in the workforce, 1997 – 2015. Scandinavian Journal of Work, Environment & Health, 47 , 335 -34 8 . Dahlberg, J. (2013). Pensionsbluffen: tryggheten som gick upp i rök [The pension scam: the security that went up in smoke], Stockholm, Ordfront. Damman, M. & Henkens, K. (2018). Gender differences in perceived workplace flexibility among older workers in the Netherlands: A brief report. Journal of Applied Gerontolog y, 39 , 915– 921. 79 Dellve, L. & Eriksson, A. (2016). Hållbart ledarskap i vardag och förändring, Styrning, Organisering och Ledning [Sustainable leadership in everyday life and in change, Governance, Organization and Management]. Högskolan i Borås. Dellve, L. , Jonsson, R., Flisbäck, M. & Bengtsson, M. (2021). System and life- course perspectives on capability to work and capability through work In: Falk Erhag, H., Nilsson Lagerlöf, U., Rydberg Sterner, T. & Skoog, I. (eds.) A Multidisciplinary Approach to Capab ility in Age and Ageing. Switzerland: Springer. Dellve, L., Karlberg, C., Allebeck, P., Herloff, B. & Hagberg, M. (2006). Macro - organizational factors, the incidence of work disability, and work ability among the total workforce of home care workers in Sweden. Scandinavian Journal of Public Health, 34 , 17 - 25. Dellve, L., Lagerström, M. & Hagberg, M. (2003). Work - system risk factors for permanent work disability among home- care workers: a case - control study. International Archives of Occupational and Environ mental Health, 76 , 216 - 224. Denton, F. T. & Spencer, B. G. (2009). What is retirement? A review and assessment of alternative concepts and measures. Canadian Journal on Aging, 28 , 63 - 76. Docherty, P., Kira, M. & Shani, A. R. (2008). Creating sustainable wo rk systems: Developing social sustainability, New York, Routledge. Domschke, P., Kadefors, R. & Wallin, M. (2014). Competitiveness and Employment for the Future: Good Examples of Age Management in the Baltic Sea Region. Hamburg. Ds (2013:19). Future challe nges for Sweden. Final report of the commission on the future of Sweden. Stockholm: Fritzes Ds (2019:2). Höjda åldersgränser i pensionssystemet och i andra trygghetssystem [Increased age thresholds in the pension system and other social security systems]. Stockholm: Fritzes Dølvik, J. E. & Steen, J. R. (2018). The Nordic future of work: Drivers, institutions, and politics (TemaNord 2018:555). Denmark: Nordic Council of Ministers. Edge, C. E., Cooper, A. M. & Coffey, M. (2017). Barriers and facilitators to extended working lives in Europe: a gender focus. Public Health Reviews, 38 , 2. Edwards, J. R. (1991). Person -job fit: A conceptual integration, literature review, and methodological critique, New York, John Wiley & Sons. Edwards, J. R., Caplan, R. D. & Van Harrison, R. (1998) . Person - environment fit theory. Theories of Organizational Stress, 28 , 67 - 94. Edwards, R. (1982). Contested terrain: The transformation of the workplace in the twentieth century. Science and Society, 46 , 237 - 240. Eismann, M., Henkens, K. & Kalmijn, M. (2019). Why Singles Prefer to Retire Later. Research on Aging, 41 , 936 - 960. Elovainio, M., Kivimäki, M., Vahtera, J., Ojanlatva, A., Korkeila, K., Suominen, S., Helenius, H. & Koskenvuo, M. (2003). Social support, early retirement, and a retirement preference: a study of 10,489 Finnish adults. Journal of Occupational and Environme ntal Medicine, 45 , 433 - 43 9 . Eppler - Hattab, R., Meshoulam, I. & Doron, I. (2019). Conceptualizing Age - friendliness in Workplaces: Proposing a New Multidimensional Model. The Gerontologist, 60 , 12- 21. 78 Braun, V. & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3 , 77 - 101. Braverman, H. (1998) . Labor and monopoly capital: The degradation of work in the twentieth century, New York, NYU Press. Browne, P., Carr, E., Fleischmann, M., Xue, B. & Stansfeld, S. A. (2019). The relationship between workplace psychosocial environment and retirement intentions and actual retirement: a systematic review. European Journal of Ageing, 16 , 73 - 82. Brunello, G. & Wruuck, P. (2019). Skill shortages and skill mismatch in Europe: a review of the literature. Bonn: IZA institute of labor economics. Byrne, B. M. (2013). Structural equation modeling with Mplus: Basic concepts, applications, and programming, New York, Ro utledge. Carr, E., Hagger - Johnson, G., Head, J., Shelton, N., Stafford, M., Stansfeld, S. & Zaninotto, P. (2016). Working conditions as predictors of retirement intentions and exit from paid employment: a 10 - year follow- up of the English Longitudinal Study of Ageing. European Journal of Ageing, 13 , 39 - 48. Carstensen, L. L. (1995). Evidence for a life - span theory of socioemotional selectivity. Current directions in Psychological science, 4 , 151 - 156. Cerdas, S., Härenstam, A., Johansson, G. & Nyberg, A. (2019 ). Development of job demands, decision authority and social support in industries with different gender composition– Sweden, 1991 – 2013. B MC Public Health, 19 , 1- 16. Cloostermans, L., Bekkers, M. B., Uiters, E. & Proper, K. I. (2015). The effectiveness of interventions for ageing workers on (early) retirement, work ability and productivity: a systematic review. International Archives of Occupational and Environmental Health, 88 , 521 - 532. Conen, W., Henkens, K. & Schippers, J. (2012). Employers’ attitudes and actions towards the extension of working lives in Europe. International Journal of Manpower, 33 , 648 - 66 5 . Conen, W., Henkens, K. & Schippers, J. (2014). Ageing organisations and the extension of working lives: A case study approach. Journal of Social Poli cy, 43 , 773 - 792. Conen, W. S., Henkens, K. & Schippers, J. J. (2011). Are employers changing their behavior toward older workers? An analysis of employers' surveys 2000 – 2009. Journal of Aging & Social Policy, 23 , 141 - 158. Corin, L. & Björk, L. (2016). Job demands and job resources in human service managerial work an external assessment through work content analysis. Nordic Journal of Working Life Studies, 6 , 3 - 28. Corin, L. , Pousette, A., Berglund, T., Dellve, L., Hensing, G. & Bj örk, L. (2021). Occupational trajectories of working conditions in Sweden: Development trends in the workforce, 1997 – 2015. Scandinavian Journal of Work, Environment & Health, 47 , 335 - 34 8 . Dahlberg, J. (2013). Pensionsbluffen: tryggheten som gick upp i rök [The pension scam: the security that went up in smoke], Stockholm, Ordfront. Damman, M. & Henkens, K. (2018). Gender differences in perceived workplace flexibility among older workers in the Netherlands: A brief report. Journal of Applied Gerontolog y, 39 , 915 – 921. 79 Dellve, L. & Eriksson, A. (2016). Hållbart ledarskap i vardag och förändring, Styrning, Organisering och Ledning [Sustainable leadership in everyday life and in change, Governance, Organization and Management]. Högskolan i Borås. Dellve, L. , Jonsson, R., Flisbäck, M. & Bengtsson, M. (2021). System and life-course perspectives on capability to work and capability through work In: Falk Erhag, H., Nilsson Lagerlöf, U., Rydberg Sterner, T. & Skoog, I. (eds.) A Multidisciplinary Approach to Capab ility in Age and Ageing. Switzerland: Springer. Dellve, L., Karlberg, C., Allebeck, P., Herloff, B. & Hagberg, M. (2006). Macro- organizational factors, the incidence of work disability, and work ability among the total workforce of home care workers in Sweden. Scandinavian Journal of Public Health, 34 , 17-25. Dellve, L., Lagerström, M. & Hagberg, M. (2003). Work -system risk factors for permanent work disability among home- care workers: a case -control study. International Archives of Occupational and Environ mental Health, 76 , 216 -224. Denton, F. T. & Spencer, B. G. (2009). What is retirement? A review and assessment of alternative concepts and measures. Canadian Journal on Aging, 28 , 63-76. Docherty, P., Kira, M. & Shani, A. R. (2008). Creating sustainable wo rk systems: Developing social sustainability, New York, Routledge. Domschke, P., Kadefors, R. & Wallin, M. (2014). Competitiveness and Employment for the Future: Good Examples of Age Management in the Baltic Sea Region. Hamburg. Ds (2013:19). Future challe nges for Sweden. Final report of the commission on the future of Sweden. Stockholm: Fritzes Ds (2019:2). Höjda åldersgränser i pensionssystemet och i andra trygghetssystem [Increased age thresholds in the pension system and other social security systems]. Stockholm: Fritzes Dølvik, J. E. & Steen, J. R. (2018). The Nordic future of work: Drivers, institutions, and politics (TemaNord 2018:555). Denmark: Nordic Council of Ministers. Edge, C. E., Cooper, A. M. & Coffey, M. (2017). Barriers and facilitators to extended working lives in Europe: a gender focus. Public Health Reviews, 38 , 2. Edwards, J. R. (1991). Person -job fit: A conceptual integration, literature review, and methodological critique, New York, John Wiley & Sons. Edwards, J. R., Caplan, R. D. & Van Harrison, R. (1998) . Person -environment fit theory. Theories of Organizational Stress, 28 , 67-94. Edwards, R. (1982). Contested terrain: The transformation of the workplace in the twentieth century. Science and Society, 46 , 237-240. Eismann, M., Henkens, K. & Kalmijn, M. (2019). Why Singles Prefer to Retire Later. Research on Aging, 41 , 936-960. Elovainio, M., Kivimäki, M., Vahtera, J., Ojanlatva, A., Korkeila, K., Suominen, S., Helenius, H. & Koskenvuo, M. (2003). Social support, early retirement, and a retirement preference: a study of 10,489 Finnish adults. Journal of Occupational and Environme ntal Medicine, 45 , 433- 43 9 . Eppler -Hattab, R., Meshoulam, I. & Doron, I. (2019). Conceptualizing Age -friendliness in Workplaces: Proposing a New Multidimensional Model. The Gerontologist, 60 , 12-21. 80 Erikson, R. & Goldthorpe, J. H. (1992). The constant flux : A study of class mobility in industrial societies, Oxford, Clarendon. ESV (2021). Utfallet för statens budget - en del av ESV:s underlag till årsredovisning för staten 2020 [The outcome for the central government budget - part of the ESV's basis for the annual report for the central government 2020]. Stockholm: Swedish National Financial Management Authority. European Council (2010). Council conclusion on active ageing. 3019th Employment, Social Policy, Health and Consumer Affairs Council meeting. Luxembourg: Council of the European Union. Eurostat (2018). Smarter, greener, more inclusive?: indicators to support the Europe 2020 strategy: 2018 edition. Luxembourg: Publications Office of the European Union. Eurostat. (2021a). Duration of working life, EU28, Country of citizenship by sex, age group 55 -64 [ Online]. Eurostat Database: Eurostat. Available: https://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=lfsi_dwl_a &lang= en [Accessed]. Eurostat. (2021b). Employment rates, EU28, Country of citizenship by sex, age group 55 -64 [ Online]. Eurostat Database: Eurostat. Available: https://ec.europa.eu/eurostat/databrowser/view/t2020_10/default/table?lan g=en [Accessed]. Eurostat. (2021c). Long -standing illness or health problem, EU28, Country of citizenship by sex, age group 55 -64 [ Online]. Eurostat Database: Eurostat. Available: https://ec.europa.eu/eurostat/databrowser/view/HLTH_SIL C_26__custom_ 1484843 /default/table?lang=en [Accessed 2021 -10 -29]. Eurostat. (2021d). Self -perceived health, EU28, Country of citizenship by sex, age group 55 -64 [ Online]. Eurostat Database: Eurostat. Available: https://ec.europa.eu/eurostat/databrowser/view/HLTH_SIL C_24__custom_ 1484697 /default/table?lang=en [Accessed 2021 -10 -29]. Eyjólfsdóttir, H. S., Baumann, I., Agahi, N. , Fritzell, J. & Lennartsson, C. (2019a). Prolongation of working life and its effect on mo rtality and health in older adults: propensity score matching. Social Science & Medicine, 226 , 77-86 . Eyjólfsdóttir, H. S., Baumann, I., Agahi, N. & Lennartsson, C. (2019b). How to Measure Retirement Age? A Comparison of Survey and Register Data. Journal o f Population Ageing, 14 , 143– 161. Fisher, G. G., Chaffee, D. S. & Sonnega, A. (2016). Retirement timing: A review and recommendations for future research. Work, Aging and Retirement, 2 , 230- 261. Foa, E. B. & Foa, U. G. (1980). Resource theory. Social excha nge. Boston: Springer. FORTE (2015). En kunskapsöversikt: Psykisk ohälsa, arbetsliv och sjukfrånvaro [An overview of knowledge: Mental illness, working life and sick leave]. In: Vingård, E. (ed.). Stockholm: Swedish Research Council for Health, Working Lif e (FORTE). Foster, L. & Walker, A. (2014). Active and successful aging: A European policy perspective. The Gerontologist, 55 , 83-90. Foster, L. & Walker, A. (2015). Active and successful aging: A European policy perspective. The Gerontologist, 55 , 83-90. 81 F uertes, V., Egdell, V. & McQuaid, R. (2013). Extending working lives: age management in SMEs. Employee Relations, 35 , 272 - 293. Furunes, T. & Mykletun, R. (2011). Managers’ decision latitude for age management: Do managers and employees have the same (impli cit) understanding? In: Ennals, R. & Salomon, R. H. (eds.) Older Workers in a Sustainable Society, Labor, Education & Society, Peter Lang Verlag, Frankfurt. Frankfurt: Peter Lang Verlag. Furunes, T. & Mykletun, R. J. (2007). Why diversity management fails: Metaphor analyses unveil manager attitudes. International Journal of Hospitality Management, 26 , 974 - 990. Furunes, T., Mykletun, R. J. & Solem, P. E. (2011). Age management in the public sector in Norway: exploring managers' decision latitude. The Interna tional Journal of Human Resource Management, 22 , 1232 - 1247. Government offices of Sweden. (2021). Nu tas nästa steg mot genomförandet av Januariavtalet med flexibilitet, omställningsförmåga och trygghet på arbetsmarknaden [The next step is now being taken towards the implementation of the January agreement with flexibility, adaptability and security in the labor market.] [ Online]. Available: https://www.regeringen.se/informationsmaterial/2021/06/flexibilitet - omstallningsformaga- och- trygghet- pa- arbetsmarknaden/ [Accessed]. Hagen, J. (2013). A history of the Swedish pension system. Uppsala: Uppsala University. Halleröd, B. & Gustafsson, J. - E. (2011). A longitudinal analysis of the relationship between changes in socio- economic status and changes in health. Social Science & Medicine, 72 , 116 - 123. Halleröd, B., Örestig, J. & Stattin, M. (2013). Leaving the labour market: the impact of exit routes from employment to retirement on health and wellbeing in old age. European Journal of Ageing, 10 , 25 - 35. Harris, K., Krygsman, S., Waschenko, J. & Laliberte Rudman, D. (2017). Ageism and the older worker: A scoping review. The Gerontologist, 58 , e1- e14. Hasselhorn, H. M. (2020). Social inequality in the transition from work to retirement. In: Theorell, T. (ed.) Handbook of Socioeconomic Determinants of Occupational Health: From Macro -level to Micro -level Evidence. New York: Springer, Cham. Hasselhorn, H. M. & Apt, W. (2015). Understanding employment participation of older workers: Creating a knowledge base for future labour market challenges, Berlin, Federal Ministry of Labour and Social Affairs. Hilsen, A. I. & Midtsundstad, T. (2015). Domain: Human resource management and interventions. In: Hasselhorn, H. M. & Apt, W. (eds.) Understanding employment participation of older workers: Creating a knowledge base for future labour market challenges. BMA S / B AuA, Berlin. Berlin: Federal Ministry of Labour and Social Affairs. Hinrichs, K. (2021). Recent pension reforms in Europe: More challenges, new directions. An overview. Social Policy & Administration, 55 , 409 - 422. Hofäcker, D., Hess, M. & König, S. (201 6). Delaying retirement, London, Palgrave Macmillan Hofäcker, D. & Radl, J. (2016). Retirement transitions in times of institutional change: Theoretical concept. In: Hofäcker, D., Hess, M. & König, S. (eds.) Delaying Retirement. London: Springer. 80 Erikson, R. & Goldthorpe, J. H. (1992). The constant flux : A study of class mobility in industrial societies, Oxford, Clarendon. ESV (2021). Utfallet för statens budget - en del av ESV:s underlag till årsredovisning för staten 2020 [The outcome for the central government budget - part of the ESV's basis for the annual report for the central government 2020]. Stockholm: Swedish National Financial Management Authority. European Council (2010). Council conclusion on active ageing. 3019th Employment, Social Policy, Health and Consumer Affairs Council meeting. Luxembo urg: Council of the European Union. Eurostat (2018). Smarter, greener, more inclusive?: indicators to support the Europe 2020 strategy: 2018 edition. Luxembourg: Publications Office of the European Union. Eurostat. (2021a). Duration of working life, EU28, Country of citizenship by sex, age group 55 -64 [ Online]. Eurostat Database: Eurostat. Available: https://appsso.eurostat.ec.europa.eu/nui/show.do?dataset=lfsi_dwl_a &lang= en [Accessed]. Eurostat. (2021b). Employment rates, EU28, Country of citizenship by sex, age group 55 -64 [ Online]. Eurostat Database: Eurostat. Available: https://ec.europa.eu/eurostat/databrowser/view/t2020_10/default/table?lan g=en [Accessed]. Eurostat. (2021c). Long -standing illness or health problem, EU28, Country of citizenship by sex, age group 55 -64 [ Online]. Eurostat Database: Eurostat. Available: https://ec.europa.eu/eurostat/databrowser/view/HLTH_SIL C_26__custom_ 1484843 /default/table?lang=en [Accessed 2021 - 10- 29]. Eurostat. (2021d). Self -perceived health, EU28, Country of citizenship by sex, age group 55 -64 [ Online]. Eurostat Database: Eurostat. Available: https://ec.europa.eu/eurostat/databrowser/view/HLTH_SIL C_24__custom_ 1484697 /default/table?lang=en [Accessed 2021 - 10- 29]. Eyjólfsdóttir, H. S., Baumann, I., Agahi, N. , Fritzell, J. & Lennartsson, C. (2019a). Prolongation of working life and its effect on mo rtality and health in older adults: propensity score matching. Social Science & Medicine, 226 , 77 - 86 . Eyjólfsdóttir, H. S., Baumann, I., Agahi, N. & Lennartsson, C. (2019b). How to Measure Retirement Age? A Comparison of Survey and Register Data. Journal o f Population Ageing, 14 , 143 – 161. Fisher, G. G., Chaffee, D. S. & Sonnega, A. (2016). Retirement timing: A review and recommendations for future research. Work, Aging and Retirement, 2 , 230 - 261. Foa, E. B. & Foa, U. G. (1980). Resource theory. Social excha nge. Boston: Springer. FORTE (2015). En kunskapsöversikt: Psykisk ohälsa, arbetsliv och sjukfrånvaro [An overview of knowledge: Mental illness, working life and sick leave]. In: Vingård, E. (ed.). Stockholm: Swedish Research Council for Health, Working Lif e (FORTE). Foster, L. & Walker, A. (2014). Active and successful aging: A European policy perspective. The Gerontologist, 55 , 83 - 90. Foster, L. & Walker, A. (2015). Active and successful aging: A European policy perspective. The Gerontologist, 55 , 83 - 90. 81 F uertes, V., Egdell, V. & McQuaid, R. (2013). Extending working lives: age management in SMEs. Employee Relations, 35 , 272-293. Furunes, T. & Mykletun, R. (2011). Managers’ decision latitude for age management: Do managers and employees have the same (impli cit) understanding? In: Ennals, R. & Salomon, R. H. (eds.) Older Workers in a Sustainable Society, Labor, Education & Society, Peter Lang Verlag, Frankfurt. Frankfurt: Peter Lang Verlag. Furunes, T. & Mykletun, R. J. (2007). Why diversity management fails: Metaphor analyses unveil manager attitudes. International Journal of Hospitality Management, 26 , 974 - 990. Furunes, T., Mykletun, R. J. & Solem, P. E. (2011). Age management in the public sector in Norway: exploring managers' decision latitude. The Interna tional Journal of Human Resource Management, 22 , 1232-1247. Government offices of Sweden. (2021). Nu tas nästa steg mot genomförandet av Januariavtalet med flexibilitet, omställningsförmåga och trygghet på arbetsmarknaden [The next step is now being taken towards the implementation of the January agreement with flexibility, adaptability and security in the labor market.] [ Online]. Available: https://www.regeringen.se/informationsmaterial/2021/06/flexibilitet - omstallningsformaga-och -trygghet -pa -arbetsmarknaden/ [Accessed]. Hagen, J. (2013). A history of the Swedish pension system. Uppsala: Uppsala University. Halleröd, B. & Gustafsson, J. -E. (2011). A longitudinal analysis of the relationship between changes in socio-economic status and changes in health. Social Science & Medicine, 72 , 116-123. Halleröd, B., Örestig, J. & Stattin, M. (2013). Leaving the labour market: the impact of exit routes from employment to retirement on health and wellbeing in old age. European Journal of Ageing, 10 , 25 -35. Harris, K., Krygsman, S., Waschenko, J. & Laliberte Rudman, D. (2017). Ageism and the older worker: A scoping review. The Gerontologist, 58 , e1-e14. Hasselhorn, H. M. (2020). Social inequality in the transition from work to retirement. In: Theorell, T. (ed.) Handbook of Socioeconomic Determinants of Occupational Health: From Macro- level to Micro -level Evidence. New York: Springer, Cham. Hasselhorn, H. M. & Apt, W. (2015). Understanding employment participation of older workers: Creating a knowledge base for future labour market challenges, Berlin, Federal Ministry of Labour and Social Affairs. Hilsen, A. I. & Midtsundstad, T. (2015). Domain: Human resource management and interventions. In: Hasselhorn, H. M. & Apt, W. (eds.) Understanding employment participation of older workers: Creating a knowledge base for future labour market challenges. BMA S / B AuA, Berlin. Berlin: Federal Ministry of Labour and Social Affairs. Hinrichs, K. (2021). Recent pension reforms in Europe: More challenges, new directions. An overview. Social Policy & Administration, 55 , 409-422. Hofäcker, D., Hess, M. & König, S. (201 6). Delaying retirement, London, Palgrave Macmillan Hofäcker, D. & Radl, J. (2016). Retirement transitions in times of institutional change: Theoretical concept. In: Hofäcker, D., Hess, M. & König, S. (eds.) Delaying Retirement. London: Springer. 82 Hornung , S., Rousseau, D. M. & Glaser, J. (2008). Creating flexible work arrangements through idiosyncratic deals. Journal of Applied Psychology, 93 , 655. Hornung, S., Rousseau, D. M. & Glaser, J. (2009). Why supervisors make idiosyncratic deals: Antecedents and outcomes of i‐deals from a managerial perspective. Journal of Managerial Psychology, 24 , 738– 76 4 . Hornung, S., Rousseau, D. M., Glaser, J., Angerer, P. & Weigl, M. (2010). Beyond top ‐ down and bottom‐up work redesign: Customizing job content through idiosyncratic deals. Journal of Organizational Behavior, 31 , 187-215. Hyde, M., Hanson, L. M., Chungkham, H. S., Leineweber, C. & Westerlund, H. (2015). The impact of involuntary exit from employment in lat er life on the risk of major depression and being prescribed anti-depressant medication. Aging & Mental Health, 19 , 381-389. Ilmarinen, J. (2006). Towards a longer working life, Helsinki, Finnish Institute of Occupational Health. Ilmarinen, J. & Rantanen, J. (1999). Promotion of work ability during ageing. American journal of industrial medicine, 36 , 21-23. ILO (2012). International Standard Classification of Occupations 200 8 (IS CO -08) : Structure, group definitions and correspondence tables, Geneva, Interna tional Labour Office. Jahoda, M. (1981). Work, employment, and unemployment: Values, theories, and approaches in social research. American Psychologist, 36 , 184-191. Jensen, P. H. (2020). Seniorarbejdsliv, Frederiksberg, Frydenlund Academic. Johansson, G. & Lundberg, I. (2004). Adjustment latitude and attendance requirements as determinants of sickness absence or attendance. Empirical tests of the illness flexibility model. Social science & medicine, 58 , 1857- 1868. Johansson, P., Laun, L. & Palme, M. (2015) . Kan vi jobba tills vi blir 75 ? Vad säger mikrodata om hälsa och arbetskapacitet bland de äldre i arbetskraften [Can we work up to 75? - What does micro data on health and work capacity in the older workforce say]. Uppsala: Institute for evaluation of lab our market and education policy. Johansson, P., Laun, L., Palme, M. & Stensöta, H. O. (2018). Drivkrafter och möjligheter till ett förlängt arbetsliv [Driving forces and opportunities for an extended working life], Stockholm, SNS Förlag. Jonsson, R., Dellv e, L. & Halleröd, B. (2019). Work despite poor health? A 14 -year follow - up of how individual work accommodations are extending the time to retirement for workers with poor health conditions. Social Science & Medicine (SSM) Population Health, 9 , 100514. Jonsson, R., Hasselgren, C., Dellve, L., Seldén, D., Larsson, D. & Stattin, M. (2021). Matching the Pieces: The Presence of Idiosyncratic Deals and Their Impact on Retirement Preferences Among Older Workers. Work, Aging and Retirement, 7 , 240– 255. Jonsson, R., Lidwall, U. & Holmgren, K. (2013). Does unbalanced gender composition in the workplace influence the association between psychosocial working conditions and sickness absence? Work, 46 , 59-66. 83 Jonsson, R., Lindegård, A., Björk, L. & Nilsson, K. (2020). Organizational Hindrances to the Retention of Older Healthcare Workers. Nordic Journal of Working Life Studies, 10. Jääskeläinen, A., Kausto, J., Seitsamo, J., Ojajarvi, A., Nygård, C. - H., Arjas, E. & Leino - Arjas, P. (2016). Work ability index and perceiv ed work ability as predictors of disability pension: a prospective study among Finnish municipal employees. Scandinavian Journal of Work, Environment & Health, 42 , 490 - 499. Kadefors, R. & Hanse, J. J. (2012). Employers’ Attitudes Toward Older Workers and O bstacles and Opportunities for the Older Unemployed to Reenter Working Life. Nordic Journal of Working Life Studies, 2 , 29 – 47. Kadefors, R., Nilsson, K., Rylander, L., Östergren, P. - O. & Albin, M. (2017). Work life length in different occupations: A Swedis h population study. Ageing & Society, 38 , 1- 18. Kadefors, R., Nilsson, K., Rylander, L., Östergren, P. - O. & Albin, M. (2018a). Occupation, gender and work - life exits: a Swedish population study. Ageing & Society, 38 , 1- 18. Kadefors, R., Nilsson, K., Österg ren, P. - O., Rylander, L. & Albin, M. (2018b). Social inequality in working life expectancy in Sweden. Zeitschrift für Gerontologie und Geriatrie, 52 , 52 - 61. Kadefors, R., Thorin, A. & Öhman, A. (2014). Vad hände i offentlig sektor när LAS - gränsen höjdes: V ilka fortsatte att jobba till 67? [What happened in the public sector when the LAS was raised: Who continued to work until 67?] . Arbetsmarknad & Arbetsliv, 20 , 25 - 37. Kadefors, R., Wikström, E. & Arman, R. (2020). The capability of organizations to manage delayed retirement. Journal of Organizational Effectiveness: People and Performance, 7 , 38 - 51. Kalleberg, A. L., Knoke, D., Marsden, P. V. & Spaeth, J. L. (199 6). Organizations in America: Analysing their structures and human resource practices, London, Sa ge. Kanfer, R. & Ackerman, P. L. (2004). Aging, adult development, and work motivation. Academy of Management Review, 29 , 440 - 458. Kankkunen, T. F., Bejerot, E., Björk, L. & Härenstam, A. (2014). New Public Management i kommunal praktik: En studie om chefers möjlighet att hantera styrning inom verksamheterna Vatten, Gymnasium och Äldreomsorg [New Public Management in municipal practice: A study of managers' ability to manage governance within Water, high school and Elderly care operations]. Gothenburg: Institute for Stress Medicine. Karasek, R. A. (1979) . Job demands, job decision latitude, and mental strain: Implications for job redesign. Administrative Science Quarterly, 24 , 285 - 308. Kautonen, T., Hytti, U., Bögenhold, D. & Heinonen, J. (2012). Job s atisfaction and retirement age intentions in Finland: self - employed versus salary earners. International Journal of Manpower, 33 , 424 - 440. Kira, M. (2003). From good work to sustainable development -Human resources consumption and regeneration in the post-bureaucratic working life. Royal Institute of Technology 82 Hornung , S., Rousseau, D. M. & Glaser, J. (2008). Creating flexible work arrangements through idiosyncratic deals. Journal of Applied Psychology, 93 , 655. Hornung, S., Rousseau, D. M. & Glaser, J. (2009). Why supervisors make idiosyncratic deals: Antecedents and outcomes of i‐deals from a managerial perspective. Journal of Managerial Psychology, 24 , 738 – 76 4 . Hornung, S., Rousseau, D. M., Glaser, J., Angerer, P. & Weigl, M. (2010). Beyond top ‐ down and bottom‐up work redesign: Customizing job content through idiosyncratic deals. Journal of Organizational Behavior, 31 , 187 - 215. Hyde, M., Hanson, L. M., Chungkham, H. S., Leineweber, C. & Westerlund, H. (2015). The impact of involuntary exit from employment in lat er life on the risk of major depression and being prescribed anti- depressant medication. Aging & Mental Health, 19 , 381 - 389. Ilmarinen, J. (2006). Towards a longer working life, Helsinki, Finnish Institute of Occupational Health. Ilmarinen, J. & Rantanen, J. (1999). Promotion of work ability during ageing. American journal of industrial medicine, 36 , 21- 23. ILO (2012). International Standard Classification of Occupations 200 8 (IS CO -08) : Structure, group definitions and correspondence tables, Geneva, Interna tional Labour Office. Jahoda, M. (1981). Work, employment, and unemployment: Values, theories, and approaches in social research. American Psychologist, 36 , 184 - 191. Jensen, P. H. (2020). Seniorarbejdsliv, Frederiksberg, Frydenlund Academic. Johansson, G. & Lundberg, I. (2004). Adjustment latitude and attendance requirements as determinants of sickness absence or attendance. Empirical tests of the illness flexibility model. Social science & medicine, 58 , 1857 - 1868. Johansson, P., Laun, L. & Palme, M. (2015) . Kan vi jobba tills vi blir 75 ? Vad säger mikrodata om hälsa och arbetskapacitet bland de äldre i arbetskraften [Can we work up to 75? - What does micro data on health and work capacity in the older workforce say]. Uppsala: Institute for evaluation of lab our market and education policy. Johansson, P., Laun, L., Palme, M. & Stensöta, H. O. (2018). Drivkrafter och möjligheter till ett förlängt arbetsliv [Driving forces and opportunities for an extended working life], Stockholm, SNS Förlag. Jonsson, R., Dellv e, L. & Halleröd, B. (2019). Work despite poor health? A 14 - year follow- up of how individual work accommodations are extending the time to retirement for workers with poor health conditions. Social Science & Medicine (SSM) Population Health, 9 , 100514. Jonsson, R., Hasselgren, C., Dellve, L., Seldén, D., Larsson, D. & Stattin, M. (2021). Matching the Pieces: The Presence of Idiosyncratic Deals and Their Impact on Retirement Preferences Among Older Workers. Work, Aging and Retirement, 7 , 240 – 255. Jonsson, R., Lidwall, U. & Holmgren, K. (2013). Does unbalanced gender composition in the workplace influence the association between psychosocial working conditions and sickness absence? Work, 46 , 59 - 66. 83 Jonsson, R., Lindegård, A., Björk, L. & Nilsson, K. (2020). Organizational Hindrances to the Retention of Older Healthcare Workers. Nordic Journal of Working Life Studies, 10. Jääskeläinen, A., Kausto, J., Seitsamo, J., Ojajarvi, A., Nygård, C. -H., Arjas, E. & Leino - Arjas, P. (2016). Work ability index and perceiv ed work ability as predictors of disability pension: a prospective study among Finnish municipal employees. Scandinavian Journal of Work, Environment & Health, 42 , 490-499. Kadefors, R. & Hanse, J. J. (2012). Employers’ Attitudes Toward Older Workers and O bstacles and Opportunities for the Older Unemployed to Reenter Working Life. Nordic Journal of Working Life Studies, 2 , 29– 47. Kadefors, R., Nilsson, K., Rylander, L., Östergren, P. -O. & Albin, M. (2017). Work life length in different occupations: A Swedis h population study. Ageing & Society, 38 , 1-18. Kadefors, R., Nilsson, K., Rylander, L., Östergren, P. -O. & Albin, M. (2018a). Occupation, gender and work - life exits: a Swedish population study. Ageing & Society, 38 , 1-18. Kadefors, R., Nilsson, K., Österg ren, P. -O., Rylander, L. & Albin, M. (2018b). Social inequality in working life expectancy in Sweden. Zeitschrift für Gerontologie und Geriatrie, 52 , 52-61. Kadefors, R., Thorin, A. & Öhman, A. (2014). Vad hände i offentlig sektor när LAS - gränsen höjdes: V ilka fortsatte att jobba till 67? [What happened in the public sector when the LAS was raised: Who continued to work until 67?] . Arbetsmarknad & Arbetsliv, 20 , 25-37. Kadefors, R., Wikström, E. & Arman, R. (2020). The capability of organizations to manage delayed retirement. Journal of Organizational Effectiveness: People and Performance, 7 , 38-51. Kalleberg, A. L., Knoke, D., Marsden, P. V. & Spaeth, J. L. (199 6). Organizations in America: Analysing their structures and human resource practices, London, Sa ge. Kanfer, R. & Ackerman, P. L. (2004). Aging, adult development, and work motivation. Academy of Management Review, 29 , 440-458. Kankkunen, T. F., Bejerot, E., Björk, L. & Härenstam, A. (2014). New Public Management i kommunal praktik: En studie om chefers möjlighet att hantera styrning inom verksamheterna Vatten, Gymnasium och Äldreomsorg [New Public Management in municipal practice: A study of managers' ability to manage governance within Water, high school and Elderly care operations]. Gothenburg: Institute for Stress Medicine. Karasek, R. A. (1979) . Job demands, job decision latitude, and mental strain: Implications for job redesign. Administrative Science Quarterly, 24 , 285 -308. Kautonen, T., Hytti, U., Bögenhold, D. & Heinonen, J. (2012). Job s atisfaction and retirement age intentions in Finland: self -employed versus salary earners. International Journal of Manpower, 33 , 424-440. Kira, M. (2003). From good work to sustainable development -Human resources consumption and regeneration in the post-bureaucratic working life. Royal Institute of Technology 84 Kooij, D. T., Jansen, P. G., Dikkers, J. S. & de Lange, A. H. (2014). Managing aging workers: A mixed methods study on bundles of HR practices for aging workers. The International Journal of Human Re source Management, 25 , 2192- 2212. Kooij, D. T., Nijssen, H., Bal, P. M. & van der Kruijssen, D. T. (2020). Crafting an interesting job: Stimulating an active role of older workers in enhancing their daily work engagement and job performance. Work, Aging an d Retirement, 6 , 165– 174. Krekula, C., Engström, L. -G. & Alvinius, A. (2017). Sweden: An extended working life policy that overlooks gender considerations. In: Ní Léime, Á., Street, D., Vickerstaff, S., Krekula, C. & Loretto, W. (eds.) Gender, ageing and e xtended working life: Cross -national perspectives. Kridahl, L. & Silverstein, M. (2019). Retirement and Aging Parents in the Swedish Population. Journal of Population Ageing, 13 , 81– 112. Kruse, A. (2010). A stable pension system: The eighth wonder. In: Ben gtsson, T. (ed.) Population Ageing -A Threat to the Welfare State? Switzerland: Springer. Kunst, A. E., Bos, V., Lahelma, E., Bartley, M., Lissau, I. , Regidor, E., Mielck, A., Cardano, M., Dalstra, J. A. & Geurts, J. J. (2005). Trends in socioeconomic inequ alities in self-assessed health in 10 European countries. International Journal of Epidemiology, 34 , 295-305. König, S., Gustafsson, P., Anxo, D. & Ericson, T. (2020). Tidig och sen pensionering: en analys av pensionsålder, pensionsmotiv och förutsättninga r för ett förlängt arbetsliv [Early and late retirement: an analysis of retirement age, pension motives and conditions for an extended working life]. Gothenburg: The Swedish Social Insurance Inspectorate (ISF). König, S., Kelfve, S., Motel - Klingebiel, A. & Wetzel, M. (2021). Development of healthcare use across contemporary retirement pathways: results from a register based cohort study. Scandinavian Journal of Public Health, 8 , 1– 8. König, S., Lindwall, M. & Johansson, B. (2018). Involuntary and Delayed Re tirement as a Possible Health Risk for Lower Educated Retirees. Journal of Population Ageing , 1- 15. Leisink, P. L. & Knies, E. (2011). Line managers' support for older workers. The International Journal of Human Resource Management, 22 , 1902-1917. Lindegård Andersson, A. (2007). Working technique during computer work Associations with biomechanical and psychological strain, neck and upper extremity musculoskeletal symptoms. The Sahlgrenska Academy, University of Gothenburg. Little, T. D. (2013). Lon gitudinal structural equation modeling, New York, Guilford press. Little, T. D., Lindenberger, U. & Nesselroade, J. R. (1999). On selecting indicators for multivariate measurement and modeling with latent variables: When" good" indicators are bad and" bad" indicators are good. Psychological methods, 4 , 192 – 211. Litwin, H. & Tur -Sinai, A. (2015). The role of the social network in early retirement among older Europeans. Work, Aging and Retirement, 1 , 340 -349. Mackenbach, J. P., Stirbu, I. , Roskam, A. -J. R., S chaap, M. M., Menvielle, G., Leinsalu, M. & Kunst, A. E. (2008). Socioeconomic inequalities in health in 22 European countries. New England Journal of Medicine, 358 , 2468 -2481. 85 Magnusson Hanson, L. L., Theorell, T., Oxenstierna, G., Hyde, M. & Westerlund, H. (2008). Demand, control and social climate as predictors of emotional exhaustion symptoms in working Swedish men and women. Scandinavian Journal of Public Health, 36 , 737 - 74 3 . Malmqvist, C. (2007). Företagsekonomiska konsekvenser av aktiv Age management [The financial consequences of active age management in businesses]. Report to the Swedish government. Stockholm: National Institute for Working Life. Marinker, M. (1975 ). Why make people patients? Journal of Medical Ethics, 1 , 81 - 84. Marques, S., Mariano , J., Mendonça, J., De Tavernier, W., Hess, M., Naegele, L., Peixeiro, F. & Martins, D. (2020). Determinants of Ageism against Older Adults: A Systematic Review. International Journal of Environmental Research and Public Health, 17 , 2560. McAllister, A., B odin, T., Brønnum - Hansen, H., Harber - Aschan, L., Barr, B., Bentley, L., Liao, Q., Koitzsch Jensen, N., Andersen, I. & Chen, W. - H. (2020). Inequalities in extending working lives beyond age 60 in Canada, Denmark, Sweden and England — By gender, level of educa tion and health. PloS one, 15 , e0234900. McMichael, A. J. (1976) . Standardized mortality ratios and the" healthy worker effect": Scratching beneath the surface. Journal of occupational medicine.: official publication of the Industrial Medical Association, 18 , 165 - 168. Mårtensson, M. (2018). 231 500 ”fattigpensionärer” – och snart kan de bli ännu fler" [231 500 poor pensioners - and soon they can become more]. Aftonbladet , 2018 - 05 - 05. Mårtensson, M. (2020). Siv, 81, har låg pension trots 40 års arbete [Siv, 81, has a low pension despite 40 years of working experience]. Aftonbladet , 2021 - 06 - 26. Naegele, G. & Walker, A. (2006). A guide to good practice in age management. Luxembourg: European Foundation for the Improvement of Living and Working Conditions. Ni Leime, A. & Street, D. (2019). Gender, ageing and extended working life: Cross -national perspectives, Bristol, Policy Press. Nilsson, K. (2016). Conceptualisation of ageing in rel ation to factors of importance for extending working life – a review. Scandinavian Journal of Public Health, 44 , 490 - 505. Nilsson, K. (2020). A sustainable working life for all ages – The swAge - model. Applied Ergonomics, 86 , 103082. Nilsson, K., Hydbom, A. R . & Rylander, L. (2011). Factors influencing the decision to extend working life or retire. Scandinavian Journal of Work, Environment & Health , 473 - 480. Nilsson, K., Johansson, B., Östergren, P. - O., Kadefors, R., Aronsson, G., Torgén, M., Gellerstedt, S., Abrahamsson, L., Nordander, C. & Österberg, K. (2017). 54 forskare: Inte alla klarar höjd pensions - ålder [54 researchers: Not everyone is capable of prolonging their working careers]. Svenska Dagbladet , 2017 - 12- 20. Norbäck, L. - E., Olsson, L. - E. & Odenrick, P. (2006). Homo Pracademicus -Om att lära om -om sig själv och sin verksamhet , Bilda förlag. 84 Kooij, D. T., Jansen, P. G., Dikkers, J. S. & de Lange, A. H. (2014). Managing aging workers: A mixed methods study on bundles of HR practices for aging workers. The International Journal of Human Re source Management, 25 , 2192 - 2212. Kooij, D. T., Nijssen, H., Bal, P. M. & van der Kruijssen, D. T. (2020). Crafting an interesting job: Stimulating an active role of older workers in enhancing their daily work engagement and job performance. Work, Aging an d Retirement, 6 , 165 – 174. Krekula, C., Engström, L. - G. & Alvinius, A. (2017). Sweden: An extended working life policy that overlooks gender considerations. In: Ní Léime, Á., Street, D., Vickerstaff, S., Krekula, C. & Loretto, W. (eds.) Gender, ageing and e xtended working life: Cross -national perspectives. Kridahl, L. & Silverstein, M. (2019). Retirement and Aging Parents in the Swedish Population. Journal of Population Ageing, 13 , 81 – 112. Kruse, A. (2010). A stable pension system: The eighth wonder. In: Ben gtsson, T. (ed.) Population Ageing -A Threat to the Welfare State? Switzerland: Springer. Kunst, A. E., Bos, V., Lahelma, E., Bartley, M., Lissau, I. , Regidor, E., Mielck, A., Cardano, M., Dalstra, J. A. & Geurts, J. J. (2005). Trends in socioeconomic inequ alities in self- assessed health in 10 European countries. International Journal of Epidemiology, 34 , 295 - 305. König, S., Gustafsson, P., Anxo, D. & Ericson, T. (2020). Tidig och sen pensionering: en analys av pensionsålder, pensionsmotiv och förutsättninga r för ett förlängt arbetsliv [Early and late retirement: an analysis of retirement age, pension motives and conditions for an extended working life]. Gothenburg: The Swedish Social Insurance Inspectorate (ISF). König, S., Kelfve, S., Motel - Klingebiel, A. & Wetzel, M. (2021). Development of healthcare use across contemporary retirement pathways: results from a register based cohort study. Scandinavian Journal of Public Health, 8 , 1– 8. König, S., Lindwall, M. & Johansson, B. (2018). Involuntary and Delayed Re tirement as a Possible Health Risk for Lower Educated Retirees. Journal of Population Ageing , 1- 15. Leisink, P. L. & Knies, E. (2011). Line managers' support for older workers. The International Journal of Human Resource Management, 22 , 1902 - 1917. Lindegård Andersson, A. (2007). Working technique during computer work Associations with biomechanical and psychological strain, neck and upper extremity musculoskeletal symptoms. The Sahlgrenska Academy, University of Gothenburg. Little, T. D. (2013). Lon gitudinal structural equation modeling, New York, Guilford press. Little, T. D., Lindenberger, U. & Nesselroade, J. R. (1999). On selecting indicators for multivariate measurement and modeling with latent variables: When" good" indicators are bad and" bad" indicators are good. Psychological methods, 4 , 192 – 211. Litwin, H. & Tur - Sinai, A. (2015). The role of the social network in early retirement among older Europeans. Work, Aging and Retirement, 1 , 340 - 349. Mackenbach, J. P., Stirbu, I. , Roskam, A. - J. R., S chaap, M. M., Menvielle, G., Leinsalu, M. & Kunst, A. E. (2008). Socioeconomic inequalities in health in 22 European countries. New England Journal of Medicine, 358 , 2468 - 2481. 85 Magnusson Hanson, L. L., Theorell, T., Oxenstierna, G., Hyde, M. & Westerlund, H. (2008). Demand, control and social climate as predictors of emotional exhaustion symptoms in working Swedish men and women. Scandinavian Journal of Public Health, 36 , 737-74 3 . Malmqvist, C. (2007). Företagsekonomiska konsekvenser av aktiv Age management [The financial consequences of active age management in businesses]. Report to the Swedish government. Stockholm: National Institute for Working Life. Marinker, M. (1975 ). Why make people patients? Journal of Medical Ethics, 1 , 81-84. Marques, S., Mariano , J., Mendonça, J., De Tavernier, W., Hess, M., Naegele, L., Peixeiro, F. & Martins, D. (2020). Determinants of Ageism against Older Adults: A Systematic Review. International Journal of Environmental Research and Public Health, 17 , 2560. McAllister, A., B odin, T., Brønnum -Hansen, H., Harber -Aschan, L., Barr, B., Bentley, L., Liao, Q., Koitzsch Jensen, N., Andersen, I. & Chen, W. -H. (2020). Inequalities in extending working lives beyond age 60 in Canada, Denmark, Sweden and England —By gender, level of educa tion and health. PloS one, 15 , e0234900. McMichael, A. J. (1976) . Standardized mortality ratios and the" healthy worker effect": Scratching beneath the surface. Journal of occupational medicine.: official publication of the Industrial Medical Association, 18 , 165-168. Mårtensson, M. (2018). 231 500 ”fattigpensionärer” – och snart kan de bli ännu fler" [231 500 poor pensioners - and soon they can become more]. Aftonbladet , 2018 -05 - 05. Mårtensson, M. (2020). Siv, 81, har låg pension trots 40 års arbete [Siv, 81, has a low pension despite 40 years of working experience]. Aftonbladet , 2021 -06-26. Naegele, G. & Walker, A. (2006). A guide to good practice in age management. Luxembourg: European Foundation for the Improvement of Living and Working Conditions. Ni Leime, A. & Street, D. (2019). Gender, ageing and extended working life: Cross -national perspectives, Bristol, Policy Press. Nilsson, K. (2016). Conceptualisation of ageing in rel ation to factors of importance for extending working life – a review. Scandinavian Journal of Public Health, 44 , 490 - 505. Nilsson, K. (2020). A sustainable working life for all ages – The swAge -model. Applied Ergonomics, 86 , 103082. Nilsson, K., Hydbom, A. R . & Rylander, L. (2011). Factors influencing the decision to extend working life or retire. Scandinavian Journal of Work, Environment & Health , 473 -480. Nilsson, K., Johansson, B., Östergren, P. -O., Kadefors, R., Aronsson, G., Torgén, M., Gellerstedt, S., Abrahamsson, L., Nordander, C. & Österberg, K. (2017). 54 forskare: Inte alla klarar höjd pensions -ålder [54 researchers: Not everyone is capable of prolonging their working careers]. Svenska Dagbladet , 2017 -12-20. Norbäck, L. - E., Olsson, L. - E. & Odenrick, P. (2006). Homo Pracademicus -Om att lära om -om sig själv och sin verksamhet , Bilda förlag. 86 Nowell, L. S., Norris, J. M., White, D. E. & Moules, N. J. (2017). Thematic Analysis: Striving to Meet the Trustworthiness Criteria. International Journal of Qualit ative Methods, 16 , 1– 13. Nussbaum, M. C. (1997 ). Capabilities and human rights. Fordham L. Rev., 66 , 273. Odeberg, E. (2020). Trygghet på arbetsmarknaden. En rapport om arbetsrätt och arbetsvillkor [Security in the labor market. A report on labor law and w orking conditions]. Stockholm: The Swedish Municipal Workers’ Union. OECD (2006). Live Longer, Work Longer. Ageing and Employment Policies. Amsterdam: OECD Publishing . OECD (2020). Pensions at a glance 2019: OECD and G20 indicators. Paris: OECD Publishing. Oesch, D. (2003). Labour market trends and the Goldthorpe class schema: a conceptual reassessment. Revue suisse de sociologie, 29 , 241-262. Oesch, D. (2006). Coming to grips with a changing class structure: An analysis of employment stratification in Brit ain, Germany, Sweden and Switzerland. International Sociology, 21 , 263 -288. Oostrom, J. K. , Pennings, M. & Bal, P. M. (2016). How do idiosyncratic deals contribute to the employability of older workers? Career Development International, 21. Oshagbemi, T. (1999). Overall job satisfaction: how good are single versus multiple ‐item measures? Journal of Managerial Psychology, 14. Phelps, E. S. (1972). The statistical theory of racism and sexism. The American Economic Review, 62 , 659-661. Phillipson, C. & Smith, A. (2005). Extending working life: A review of the research literature , CDS. Prakash, K., Virtanen, M., Pentti, J., Kivimäki, M., Vahtera, J. & Stenholm, S. (2021). Does working beyond the statutory retirement age have an impact on health and functional capacity? The Finnish Retirement and Aging cohort study. Occupational and Environmental Medicine, 78 , 509-515. Principi, A., Bauknecht, J., Di Rosa, M. & Socci, M. (2020). Employees’ Longer Working Lives in Europe: Drivers and Barriers in Companies. Interna tional Journal of Environmental Research and Public Health, 17 , 1658. PRO (2015). Ålderdom utan fattigdom - om äldrefattigdom och kvinnors låga pensioner [Old age without poverty - about elderly poverty and women's low pensions]. Stockholm: The National Or ganization of Pensioners. Rad, E. H., Rashidian, A., Arab, M. & Souri, A. (2017). Comparison the effects of poor health and low income on early retirement: a systematic review and meta - analysis. Industrial Health, 55 , 306-313. Radl, J. (2013). Labour marke t exit and social stratification in Western Europe: The effects of social class and gender on the timing of retirement. European Sociological Review, 29 , 654 -66 8 . Regeringskansliet. (2019). Trygga framtidens pensioner – uppföljning av den blocköverskridand e pensionsöverenskommelsen [Goverment offices of Sweden - Secure future pensions: a follow -up of the pension agreement] [ Online]. 2019 -02-02: Regeringskansliet. Available: https://www.regeringen.se/pressmeddelanden/2019/02/trygga -framtidens - pensioner-- uppfoljning -av - den- blockoverskridande- pensionsoverenskommelsen/ [Accessed]. 87 Richert- Kaźmierska $ Stankiewicz, K. (2016). Work – life balance: Does age matter? Work, 55 , 679 - 68 8 . Robroek, S. J., Schuring, M., Croezen, S., Stattin, M. & Burdorf, A. (2013). Poor health, unhealthy behaviors, and unfavorable work characteristics influence pathways of exit from paid employment among older workers in Europe: a four year follow- up study. Scand J Work Environ Health, 39 , 125 - 33. Rosen, C. C., Slater, D. J., Chang, C. - + -RKnsRn 5 (   /eW·s make a deaO Development and validation of the ex post i- deals scale. Journal of Management, 39 , 709 - 742. Rowe, J. W. & Kahn, R. L. (1987). Human aging: usual and successful. Science, 237 , 143 - 149. Rundström, M. (2017). Passé : de ofrivilliga pensionärerna [Passé: the involuntary pensioners.], Stockholm, Wahlström Widstrand. SALAR (2018a). Heltidsarbete som norm [Full - time employment as norm]. Stockholm: Swedish Association of Local Au thorities and Regions. SALAR (2018b). Sveriges viktigaste jobb finns i välfärden. Rekryteringsrapport 2018 [Swedish most important work is in the welfare. Recruitment report 2018]. Stockholm: Swedish Association of Local Authorities and Regions. SALAR (201 9). Köp av verksamhet 2019 - Kommuner och regioner 2006 - 2018 [Purchase of services 2019 - Municipalities and regions 2006 - 2018]. Stockholm: Swedish Association of Local Authorities and Regions. SALAR (2020). Möt Kompetensutmaningen - Rekryteringsrapport 2020 [Meet the Competence Challenge - Recruitment Report 2020]. Stockholm: Swedish Association of Local Authorities and Regions. Sangregorio, I. - L . (2018). Blåsningen : så har det nya pensionssystemet lurat oss alla, Stockholm, Karneval förlag. Scharn, M., S ewdas, R., Boot, C. R., Huisman, M., Lindeboom, M. & Van Der Beek, A. J. (2018). Domains and determinants of retirement timing: A systematic review of longitudinal studies. B MC Public Health, 18 , 1083. Scheaffer, R. L., Mendenhall III, W., Ott, R. L. & Ger ow, K. G. (2011). Elementary survey sampling, USA, Cengage Learning. Scherman, K. G. (2014). Pensioner på villovägar: orsaker och lösningar, Stockholm, Jure Förl. Schmidt, L., Sjöström, J. & Antonsson, A. - B. (2017). Understanding the challenges facing occupational health services in the Swedish public sector. Nordic Journal of Working Life Studies, 7 , 85 - 104. Seldén, D., Hasselgren, C., Jonsson, R. & Dellve, L. (2020). Arbete och Pensionering - uppfattningar bland anställda i Göteborgs stad [Work and Retire ment - Attitudes among employees in the City of Gothenburg]. Gothenburg Studies in Work Science. Rapport Nr 21, 2020 ed. Gothenburg. Sen, A. (2005). Human rights and capabilities. Journal of Human Development, 6 , 151 - 166 . Sen, A. K. (2009). The idea of jus tice, Boston, Harvard University Press. Siegrist, J., Wahrendorf, M., Von dem Knesebeck, O., Jürges, H. & Börsch - Supan, A. (2007). Quality of work, well - being, and intended early retirement of older employees— baseline results from the SHARE Study. The European Journal of Public Health, 17 , 62 - 68. 86 Nowell, L. S., Norris, J. M., White, D. E. & Moules, N. J. (2017). Thematic Analysis: Striving to Meet the Trustworthiness Criteria. International Journal of Qualit ative Methods, 16 , 1– 13. Nussbaum, M. C. (1997 ). Capabilities and human rights. Fordham L. Rev., 66 , 273. Odeberg, E. (2020). Trygghet på arbetsmarknaden. En rapport om arbetsrätt och arbetsvillkor [Security in the labor market. A report on labor law and w orking conditions]. Stockholm: The Swedish Municipal Workers’ Union. OECD (2006). Live Longer, Work Longer. Ageing and Employment Policies. Amsterdam: OECD Publishing . OECD (2020). Pensions at a glance 2019: OECD and G20 indicators. Paris: OECD Publishing. Oesch, D. (2003). Labour market trends and the Goldthorpe class schema: a conceptual reassessment. Revue suisse de sociologie, 29 , 241 - 262. Oesch, D. (2006). Coming to grips with a changing class structure: An analysis of employment stratification in Brit ain, Germany, Sweden and Switzerland. International Sociology, 21 , 263 - 288. Oostrom, J. K. , Pennings, M. & Bal, P. M. (2016). How do idiosyncratic deals contribute to the employability of older workers? Career Development International, 21. Oshagbemi, T. (1999). Overall job satisfaction: how good are single versus multiple ‐item measures? Journal of Managerial Psychology, 14. Phelps, E. S. (1972). The statistical theory of racism and sexism. The American Economic Review, 62 , 659 - 661. Phillipson, C. & Smith, A. (2005). Extending working life: A review of the research literature , CDS. Prakash, K., Virtanen, M., Pentti, J., Kivimäki, M., Vahtera, J. & Stenholm, S. (2021). Does working beyond the statutory retirement age have an impact on health and functional capacity? The Finnish Retirement and Aging cohort study. Occupational and Environmental Medicine, 78 , 509 - 515. Principi, A., Bauknecht, J., Di Rosa, M. & Socci, M. (2020). Employees’ Longer Working Lives in Europe: Drivers and Barriers in Companies. Interna tional Journal of Environmental Research and Public Health, 17 , 1658. PRO (2015). Ålderdom utan fattigdom - om äldrefattigdom och kvinnors låga pensioner [Old age without poverty - about elderly poverty and women's low pensions]. Stockholm: The National Or ganization of Pensioners. Rad, E. H., Rashidian, A., Arab, M. & Souri, A. (2017). Comparison the effects of poor health and low income on early retirement: a systematic review and meta - analysis. Industrial Health, 55 , 306 - 313. Radl, J. (2013). Labour marke t exit and social stratification in Western Europe: The effects of social class and gender on the timing of retirement. European Sociological Review, 29 , 654 - 66 8 . Regeringskansliet. (2019). Trygga framtidens pensioner – uppföljning av den blocköverskridand e pensionsöverenskommelsen [Goverment offices of Sweden - Secure future pensions: a follow -up of the pension agreement] [ Online]. 2019 - 02- 02: Regeringskansliet. Available: https://www.regeringen.se/pressmeddelanden/2019/02/trygga - framtidens- pensioner-- uppfoljning - av- den- blockoverskridande- pensionsoverenskommelsen/ [Accessed]. 87 Richert-Kaźmierska $ Stankiewicz, K. (2016). Work – life balance: Does age matter? Work, 55 , 679-68 8 . Robroek, S. J., Schuring, M., Croezen, S., Stattin, M. & Burdorf, A. (2013). Poor health, unhealthy behaviors, and unfavorable work characteristics influence pathways of exit from paid employment among older workers in Europe: a four year follow-up study. Scand J Work Environ Health, 39 , 125- 33. Rosen, C. C., Slater, D. J., Chang, C. -+ -RKnsRn 5 (   /eW·s make a deaO Development and validation of the ex post i-deals scale. Journal of Management, 39 , 709- 742. Rowe, J. W. & Kahn, R. L. (1987). Human aging: usual and successful. Science, 237 , 143 - 149. Rundström, M. (2017). Passé : de ofrivilliga pensionärerna [Passé: the involuntary pensioners.], Stockholm, Wahlström Widstrand. SALAR (2018a). Heltidsarbete som norm [Full -time employment as norm]. Stockholm: Swedish Association of Local Au thorities and Regions. SALAR (2018b). Sveriges viktigaste jobb finns i välfärden. Rekryteringsrapport 2018 [Swedish most important work is in the welfare. Recruitment report 2018]. Stockholm: Swedish Association of Local Authorities and Regions. SALAR (2019). Köp av verksamhet 2019 - Kommuner och regioner 2006 -2018 [Purchase of services 2019 - Municipalities and regions 2006 -2018]. Stockholm: Swedish Association of Local Authorities and Regions. SALAR (2020). Möt Kompetensutmaningen - Rekryteringsrapport 2020 [Meet the Competence Challenge - Recruitment Report 2020]. Stockholm: Swedish Association of Local Authorities and Regions. Sangregorio, I. -L . (2018). Blåsningen : så har det nya pensionssystemet lurat oss alla, Stockholm, Karneval förlag. Scharn, M., S ewdas, R., Boot, C. R., Huisman, M., Lindeboom, M. & Van Der Beek, A. J. (2018). Domains and determinants of retirement timing: A systematic review of longitudinal studies. B MC Public Health, 18 , 1083. Scheaffer, R. L., Mendenhall III, W., Ott, R. L. & Ger ow, K. G. (2011). Elementary survey sampling, USA, Cengage Learning. Scherman, K. G. (2014). Pensioner på villovägar: orsaker och lösningar, Stockholm, Jure Förl. Schmidt, L., Sjöström, J. & Antonsson, A. -B. (2017). Understanding the challenges facing occupational health services in the Swedish public sector. Nordic Journal of Working Life Studies, 7 , 85-104. Seldén, D., Hasselgren, C., Jonsson, R. & Dellve, L. (2020). Arbete och Pensionering - uppfattningar bland anställda i Göteborgs stad [Work and Retire ment - Attitudes among employees in the City of Gothenburg]. Gothenburg Studies in Work Science. Rapport Nr 21, 2020 ed. Gothenburg. Sen, A. (2005). Human rights and capabilities. Journal of Human Development, 6 , 151-166 . Sen, A. K. (2009). The idea of jus tice, Boston, Harvard University Press. Siegrist, J., Wahrendorf, M., Von dem Knesebeck, O., Jürges, H. & Börsch -Supan, A. (2007). Quality of work, well - being, and intended early retirement of older employees—baseline results from the SHARE Study. The European Journal of Public Health, 17 , 62-68. 88 Skirbekk, V. (2004). Age and individual productivity: A literature survey. Vienna Yearbook of Population Research 2 , 133- 153. SOU (2012). Längre liv, längre arbetsliv - Förutsättningar och hinder för äldre att arbeta längre [Longer lives, Longer working lives - preconditions, and barriers for older people to work longer]. Stockholm: Swedish Government Official Reports. SOU (2013). Åtgärder för ett längre arbetsliv [Measures to extend working lives]. Stockholm: S wedish Government Official Reports. SOU (2020). Äldre har aldrig varit yngre – allt fler kan och vill arbeta längre [older people have never been younger - more people can and want to continue working]. Stockholm: Swedish Government Official Reports. Sousa-Ribeiro, M., Bernhard -Oettel, C., Sverke, M. & Westerlund, H. (2021). Health -and Age - Related Workplace Factors as Predictors of Preferred, Expected, and Actual Retirement Timing: Findings from a Swedish Cohort Study. International Journal of Environmental Research and Public Health, 18 , 2746. SSIA (2015). Sjukskrivningar 60 dagar eller längre - En beskrivning av sjukskrivna åren 1999 – 2014 efter kön, ålder, arbetsmarknadsstatus, yrke, sjukskrivningslängd och diagnospanorama [Sick leave 60 days or longer - A description of sick leave in the years 1999 – 2014 by sex, age, labor market status, occupation, length of sick leave and diagnostic panoramas]. Stockholm: The Swedish Social Insurance Agency SSIA (2020). Sjukfrånvaro i psykiatriska diagnoser - En register studie av Sveriges arbetande befolkning i åldern 20 – 69 år [Sick leave in psychiatric diagnoses - A register study of Sweden's working population aged 20 – 69] . Stockholm: The Swedish Social Insurance Agency. Statistics Sweden. (2003). Undersökning av levnads förhållanden 2002 -2003 [Survey of Living Conditions 200 2 -2003 ] [ Online]. Stockholm: Statistics Sweden. Available: https://www.scb.se/contentassets/35d a017ddbc3439a932bd8af95c58601/le0 101_bs_2002_2003.pdf [Accessed 2021 -10-30]. Statistics Sweden. (2012). S S Y K 2012 Swedish Standard Classification of Occupations [ Online]. Stockholm: Statistics Sweden. Available: https://www.scb.se/dokumentation/klassifikationer -och -standarder/standard - for-svensk -yrkesklassificering - ssyk/ [Accessed 2021 -03-05]. Statistics Sweden. (2019). Price Basic Amount (1960 -2019 ) [ Online]. Stockholm: Statistics Sweden. Available: https://www.scb.se/hitta -statistik/statistik -efter - amne/priser -och -konsumtion/konsumentprisindex/konsumentprisindex - kpi/pong/tabell -och -diagram/prisbasbelopp/prisbasbelopp/ [Accessed 2021 - 09-20]. Statistics Sweden. (2020a). Average monthly salary, 1973 – 2 0 1 9 [ Online]. Stockholm: Statistics Sweden. Available: https://www.scb.se/en/finding - statistics/statistics - by-subject -area/labour -market/wages -salaries -and -labour - costs/wage -and -salary -structures -and -employment -in -the - primary- municipalities/pong/tables -and -graphs/average -monthly -salary/ [Accessed 2021-04-05]. Statistics Sweden. (2020b). Undersökningarna av levnadsförhållanden (ULF / S IL C) [Survey of Living conditions (ULF / S IL C) ] [Online]. Stockholm: Statistics Sweden. Available: https://www.scb.se/hitta - statistik/statistik -efter - 8 9 amne/levnadsforhallanden/ levnadsforhallanden/undersokningarna - av- levnadsforhallanden- ulf- silc/#_Dokumentation [Accessed 2020 - 08 - 05]. Stenholm, S., Westerlund, H., Salo, P., Hyde, M., Pentti, J., Head, J., Kivimäki, M. & Vahtera, J. (2014). Age - related trajectories of physical func tioning in work and retirement: the role of sociodemographic factors, lifestyle and disease. J Epidemiol Community Health, 68 , 503 - 509. SWEA (2012a). Arbetsmiljön 2012 [The Work Environment 2011]. Stockholm: Swedish Work Environment Authority. SWEA (2012b) . Kunskapsöversikt: Jobba längre - vad vet vi om äldre i arbetslivet? [Research review: Work longer - What do we know about older in workforce?]. Stockholm: Swedish Work Environment Authority. SWEA (2013). Belastning, genus och hälsa i arbetslivet, kunskaps sammanställning [Stress, gender and health in working life, a research review]. Stockholm: Swedish Work Environment Authority. SWEA (2014). Sambandet mellan arbetsmiljö och beslutet att lämna arbetskraften (Rapport 2014:8) [The associaiton between working environment and the decision to leave the workforce]. . Stockholm: Swedish Work Environment Authority. SWEA (2015). Arbetsmiljölagen: och dess förordning med kommentarer 1 januari 2015 [The Work Environment Act: and its regulation with comments 1 january 2 015]. Stockholm: Swedish Work Environment Authority. SWEA (2016). Den organisatoriska och sociala arbetsmiljön: viktiga pusselbitar i en god arbetsmiljö [The organizational and social work environment: important pieces for a sound work environment]. Stockh olm: The Swedish Work Environment Authority. SWEA (2016:2). The Work Environment 2015. Stockholm: The Swedish Work Environment Authority. SWEA (2020a). Arbetsmiljön 2019 [The Work Environment 2019]. Stockholm: The Swedish Work Environment Authority. SWEA ( 2020b). Register över samtliga författningar med mera som är utfärdade av Arbetsmiljöverket under 2015 - 2020 [Register of all regulations that were issued by the Swedish Work Environment Authority during 2015 - 2020]. Sweden: Swedish Work Environment Authorit y. Swedish Parliament (2019). En riktålder för höjda pensioner och följsamhet till ett längre liv [A target age for increased pensions and adherence to a longer life]. Stockholm: Socialförsäkringsutskottets Swedish Research Council (2002). Forskningsetisk a principer inom humanistisk- samhällsvetenskaplig forskning [Research ethics principles in humanities and social science research]. Stockholm: Swedish Research Council. SVT. (2016). De vill folkbilda om pensioner [They want to educate people about pensions]. 2016 - 06 - 19. Söderbacka, T., Nyholm, L. & Fagerström, L. (2020). Workplace interventions that support older employees’ health and work ability - a scoping review. B MC Health Services Research, 20 , 1- 9. 88 Skirbekk, V. (2004). Age and individual productivity: A literature survey. Vienna Yearbook of Population Research 2 , 133 - 153. SOU (2012). Längre liv, längre arbetsliv - Förutsättningar och hinder för äldre att arbeta längre [Longer lives, Longer working lives - preconditions, and barriers for older people to work longer]. Stockholm: Swedish Government Official Reports. SOU (2013). Åtgärder för ett längre arbetsliv [Measures to extend working lives]. Stockholm: S wedish Government Official Reports. SOU (2020). Äldre har aldrig varit yngre – allt fler kan och vill arbeta längre [older people have never been younger - more people can and want to continue working]. Stockholm: Swedish Government Official Reports. Sousa- Ribeiro, M., Bernhard - Oettel, C., Sverke, M. & Westerlund, H. (2021). Health - and Age - Related Workplace Factors as Predictors of Preferred, Expected, and Actual Retirement Timing: Findings from a Swedish Cohort Study. International Journal of Environmental Research and Public Health, 18 , 2746. SSIA (2015). Sjukskrivningar 60 dagar eller längre - En beskrivning av sjukskrivna åren 1999 – 2014 efter kön, ålder, arbetsmarknadsstatus, yrke, sjukskrivningslängd och diagnospanorama [Sick leave 60 days or longer - A description of sick leave in the years 1999 – 2014 by sex, age, labor market status, occupation, length of sick leave and diagnostic panoramas]. Stockholm: The Swedish Social Insurance Agency SSIA (2020). Sjukfrånvaro i psykiatriska diagnoser - En register studie av Sveriges arbetande befolkning i åldern 20 – 69 år [Sick leave in psychiatric diagnoses - A register study of Sweden's working population aged 20 – 69] . Stockholm: The Swedish Social Insurance Agency. Statistics Sweden. (2003). Undersökning av levnads förhållanden 2002 -2003 [Survey of Living Conditions 200 2 -2003 ] [ Online]. Stockholm: Statistics Sweden. Available: https://www.scb.se/contentassets/35d a017ddbc3439a932bd8af95c58601/le0 101_bs_2002_2003.pdf [Accessed 2021 - 10- 30]. Statistics Sweden. (2012). S S Y K 2012 Swedish Standard Classification of Occupations [ Online]. Stockholm: Statistics Sweden. Available: https://www.scb.se/dokumentation/klassifikationer - och- standarder/standard - for- svensk- yrkesklassificering- ssyk/ [Accessed 2021 - 03 - 05]. Statistics Sweden. (2019). Price Basic Amount (1960 -2019 ) [ Online]. Stockholm: Statistics Sweden. Available: https://www.scb.se/hitta - statistik/statistik - efter- amne/priser - och- konsumtion/konsumentprisindex/konsumentprisindex - kpi/pong/tabell - och- diagram/prisbasbelopp/prisbasbelopp/ [Accessed 2021 - 09 - 20]. Statistics Sweden. (2020a). Average monthly salary, 1973 – 2 0 1 9 [ Online]. Stockholm: Statistics Sweden. Available: https://www.scb.se/en/finding - statistics/statistics - by- subject - area/labour - market/wages - salaries- and- labour- costs/wage - and- salary- structures- and- employment- in- the- primary- municipalities/pong/tables - and- graphs/average - monthly- salary/ [Accessed 2021- 04 - 05]. Statistics Sweden. (2020b). Undersökningarna av levnadsförhållanden (ULF / S IL C) [Survey of Living conditions (ULF / S IL C) ] [Online]. Stockholm: Statistics Sweden. Available: https://www.scb.se/hitta - statistik/statistik - efter- 8 9 amne/levnadsforhallanden/ levnadsforhallanden/undersokningarna -av - levnadsforhallanden-ulf - silc/#_Dokumentation [Accessed 2020 -08-05]. Stenholm, S., Westerlund, H., Salo, P., Hyde, M., Pentti, J., Head, J., Kivimäki, M. & Vahtera, J. (2014). Age -related trajectories of physical func tioning in work and retirement: the role of sociodemographic factors, lifestyle and disease. J Epidemiol Community Health, 68 , 503-509. SWEA (2012a). Arbetsmiljön 2012 [The Work Environment 2011]. Stockholm: Swedish Work Environment Authority. SWEA (2012b) . Kunskapsöversikt: Jobba längre - vad vet vi om äldre i arbetslivet? [Research review: Work longer - What do we know about older in workforce?]. Stockholm: Swedish Work Environment Authority. SWEA (2013). Belastning, genus och hälsa i arbetslivet, kunskaps sammanställning [Stress, gender and health in working life, a research review]. Stockholm: Swedish Work Environment Authority. SWEA (2014). Sambandet mellan arbetsmiljö och beslutet att lämna arbetskraften (Rapport 2014:8) [The associaiton between working environment and the decision to leave the workforce]. . Stockholm: Swedish Work Environment Authority. SWEA (2015). Arbetsmiljölagen: och dess förordning med kommentarer 1 januari 2015 [The Work Environment Act: and its regulation with comments 1 january 2 015]. Stockholm: Swedish Work Environment Authority. SWEA (2016). Den organisatoriska och sociala arbetsmiljön: viktiga pusselbitar i en god arbetsmiljö [The organizational and social work environment: important pieces for a sound work environment]. Stockh olm: The Swedish Work Environment Authority. SWEA (2016:2). The Work Environment 2015. Stockholm: The Swedish Work Environment Authority. SWEA (2020a). Arbetsmiljön 2019 [The Work Environment 2019]. Stockholm: The Swedish Work Environment Authority. SWEA ( 2020b). Register över samtliga författningar med mera som är utfärdade av Arbetsmiljöverket under 2015 - 2020 [Register of all regulations that were issued by the Swedish Work Environment Authority during 2015 -2020]. Sweden: Swedish Work Environment Authorit y. Swedish Parliament (2019). En riktålder för höjda pensioner och följsamhet till ett längre liv [A target age for increased pensions and adherence to a longer life]. Stockholm: Socialförsäkringsutskottets Swedish Research Council (2002). Forskningsetisk a principer inom humanistisk- samhällsvetenskaplig forskning [Research ethics principles in humanities and social science research]. Stockholm: Swedish Research Council. SVT. (2016). De vill folkbilda om pensioner [They want to educate people about pensions]. 2016 -06-19. Söderbacka, T., Nyholm, L. & Fagerström, L. (2020). Workplace interventions that support older employees’ health and work ability -a scoping review. B MC Health Services Research, 20 , 1-9. 90 Taylor, P. & Walker, A. (1998). Employers and older wo rkers: attitudes and employment practices. Ageing & Society, 18 , 641-658. Thomée, S., Eklöf, M., Gustafsson, E., Nilsson, R. & Hagberg, M. (2007). Prevalence of perceived stress, symptoms of depression and sleep disturbances in relation to information and communication technology (IC T) use among young adults – an explorative prospective study. Computers in Human Behavior, 23 , 1300-1321. Thylefors, I. (2016). Chef - och ledarskap inom välfärdssektorn [Management and leadership in the welfare sector], Stockholm, Natur & Kultur. Topa, G., Depolo, M. & Alcover, C. -M. (2018). Early retirement: a meta -analysis of its antecedent and subsequent correlates. Frontiers in Psychology, 8 , 2157. Topa, G., Moriano, J. A., Depolo, M., Alcover, C. -M . & Morales, J. F. (2009). An tecedents and consequences of retirement planning and decision - making: A meta - analysis and model. Journal of Vocational Behavior, 75 , 38-55. Toppinen -Tanner, S. & Vuori, J. (2018). WP 8: Life Course Management – concept as a good practice. In: Naegele, G. & Hess, M. (eds.) Extending working lives of an ageing workforce. Dortmund: Research Association for Gerontology at TU Dortmund University. Truxillo, D. M., Cadiz, D. M. & Hammer, L. B. (2015). Supporting the aging workforce: A review and recommendations for workplace intervention research. Annual Review of Organizational Psychology and Organizational Behavior, 2 , 351-381. United Nations (2017). World Population Ageining 2017 Highlights. New York: United Nations department for economic social affairs. United Nations. (2021). Decent Work and Economic Growth [ Online]. Available: https://www.globalgoals.org/8 -decent -work -and -economic -growth [Accessed]. Walker, A. (2005). The emergence o f age management in Europe. International Journal of Organisational Behaviour, 10 , 685- 69 7 . Van Dalen, H. P., Henkens, K. & Schippers, J. (2010). Productivity of older workers: Perceptions of employers and employees. Population and Development Review, 36 , 309-330. Van der Doef, M. & Maes, S. (1999). The job demand -control ( -support) model and psychological well- being: a review of 20 years of empirical research. Work & Stress, 13 , 87-114. Van der Klink, J. J., Bültmann, U., Burdorf, A., Schaufeli, W. B., Zij lstra, F. R., Abma, F. I., Brouwer, S. & Van der Wilt, G. J. (2016). Sustainable employability— definition, conceptualization, and implications: a perspective based on the capability approach. Scandinavian Journal of Work, Environment & Health , 71 -79. Van Rijn, R. M., Robroek, S. J., Brouwer, S. & Burdorf, A. (2014). Influence of poor health on exit from paid employment: a systematic review. Occup Environ Med, 71 , 295 -301. Van Solinge, H. & Henkens, K. (2014). Work -related factors as predictors in the retirement decision-making process of older workers in the Netherlands. Ageing & Society, 34 , 1551-1574. Wang, M. & Shultz, K. S. (2010). Employee retirement: A review and recommendations for future investigation. Journal of Management, 36 , 172-206. 91 Vickers taff, S., Cox, J. & Keen, L. (2003). Employers and the management of retirement. Social Policy & Administration, 37 , 271 - 287. Wilckens, M. R., Wöhrmann, A. M., Deller, J. & Wang, M. (2020). Organizational Practices for the Aging Workforce: Development and Validation of the Later Life Workplace Index. Work, Aging and Retirement, 7 , 352 – 386,. Virtanen, M., Oksanen, T., Batty, G. D., Ala - Mursula, L., Salo, P., Elovainio, M., Pentti, J., Lybäck, K., Vahtera, J. & Kivimäki, M. (2014). Extending employment beyond the pensionable age: a cohort study of the influence of chronic diseases, health risk factors, and working conditions. PloS one, 9 , e88695. Ybema, J. F., Geuskens, G. A., van den Heuvel, S. G., de Wind, A., Leijten, F. R., Joling, C. I., Blatter, B. M., B urdorf, A., van der Beek, A. J. & Bongers, P. M. (2014). Study on Transitions in Employment, Ability and Motivation (STREAM) : the design of a four- year longitudinal cohort study among 15,118 persons aged 45 to 64 years. British Journal of Medicine and Medi cal Research, 4 , 1383 - 1399. Zacher, H. (2015). Successful aging at work. Work, Aging and Retirement, 1 , 4 - 25. Zacher, H. & Yang, J. (2016). Organizational Climate for Successful Aging. Frontiers in Psychology, 7. Zaniboni, S., Sarchielli, G. & Fraccaroli, F. (2010). How are psychosocial factors related to retirement intentions? International Journal of Manpower, 31 , 271 - 285. Örestig, J., Strandh, M. & Stattin, M. (2013). A Wish Come True? A Longitudinal Analysis of the Relationship between Retirement Prefer ences and the Timing of Retirement. Journal of Population Ageing, 6 , 99 - 118. 90 Taylor, P. & Walker, A. (1998). Employers and older wo rkers: attitudes and employment practices. Ageing & Society, 18 , 641 - 658. Thomée, S., Eklöf, M., Gustafsson, E., Nilsson, R. & Hagberg, M. (2007). Prevalence of perceived stress, symptoms of depression and sleep disturbances in relation to information and communication technology (IC T) use among young adults – an explorative prospective study. Computers in Human Behavior, 23 , 1300 - 1321. Thylefors, I. (2016). Chef - och ledarskap inom välfärdssektorn [Management and leadership in the welfare sector], Stockholm, Natur & Kultur. Topa, G., Depolo, M. & Alcover, C. - M. (2018). Early retirement: a meta - analysis of its antecedent and subsequent correlates. Frontiers in Psychology, 8 , 2157. Topa, G., Moriano, J. A., Depolo, M., Alcover, C. - M . & Morales, J. F. (2009). An tecedents and consequences of retirement planning and decision - making: A meta - analysis and model. Journal of Vocational Behavior, 75 , 38 - 55. Toppinen - Tanner, S. & Vuori, J. (2018). WP 8: Life Course Management – concept as a good practice. In: Naegele, G. & Hess, M. (eds.) Extending working lives of an ageing workforce. Dortmund: Research Association for Gerontology at TU Dortmund University. Truxillo, D. M., Cadiz, D. M. & Hammer, L. B. (2015). Supporting the aging workforce: A review and recommendations for workplace intervention research. Annual Review of Organizational Psychology and Organizational Behavior, 2 , 351 - 381. United Nations (2017). World Population Ageining 2017 Highlights. New York: United Nations department for economic social affairs. United Nations. (2021). Decent Work and Economic Growth [ Online]. Available: https://www.globalgoals.org/8 - decent- work- and- economic- growth [Accessed]. Walker, A. (2005). The emergence o f age management in Europe. International Journal of Organisational Behaviour, 10 , 685 - 69 7 . Van Dalen, H. P., Henkens, K. & Schippers, J. (2010). Productivity of older workers: Perceptions of employers and employees. Population and Development Review, 36 , 309 - 330. Van der Doef, M. & Maes, S. (1999). The job demand - control ( - support) model and psychological well- being: a review of 20 years of empirical research. Work & Stress, 13 , 87 - 114. Van der Klink, J. J., Bültmann, U., Burdorf, A., Schaufeli, W. B., Zij lstra, F. R., Abma, F. I., Brouwer, S. & Van der Wilt, G. J. (2016). Sustainable employability— definition, conceptualization, and implications: a perspective based on the capability approach. Scandinavian Journal of Work, Environment & Health , 71 - 79. Van Rijn, R. M., Robroek, S. J., Brouwer, S. & Burdorf, A. (2014). Influence of poor health on exit from paid employment: a systematic review. Occup Environ Med, 71 , 295 - 301. Van Solinge, H. & Henkens, K. (2014). Work - related factors as predictors in the retirement decision- making process of older workers in the Netherlands. Ageing & Society, 34 , 1551 - 1574. Wang, M. & Shultz, K. S. (2010). Employee retirement: A review and recommendations for future investigation. Journal of Management, 36 , 172 - 206. 91 Vickerstaff, S., Cox, J. & Keen, L. (2003). Employers and the management of retirement. Social Policy & Administration, 37 , 271-287. Wilckens, M. R., Wöhrmann, A. M., Deller, J. & Wang, M. (2020). Organizational Practices for the Aging Workforce: Development and Validation of the Later Life Workplace Index. Work, Aging and Retirement, 7 , 352 – 386,. Virtanen, M., Oksanen, T., Batty, G. D., Ala -Mursula, L., Salo, P., Elovainio, M., Pentti, J., Lybäck, K., Vahtera, J. & Kivimäki, M. (2014). Extending employment beyond the pensionable age: a cohort study of the influence of chronic diseases, health risk factors, and working conditions. PloS one, 9 , e88695. Ybema, J. F., Geuskens, G. A., van den Heuvel, S. G., de Wind, A., Leijten, F. R., Joling, C. I., Blatter, B. M., B urdorf, A., van der Beek, A. J. & Bongers, P. M. (2014). Study on Transitions in Employment, Ability and Motivation (STREAM) : the design of a four-year longitudinal cohort study among 15,118 persons aged 45 to 64 years. British Journal of Medicine and Medi cal Research, 4 , 1383-1399. Zacher, H. (2015). Successful aging at work. Work, Aging and Retirement, 1 , 4 -25. Zacher, H. & Yang, J. (2016). Organizational Climate for Successful Aging. Frontiers in Psychology, 7. Zaniboni, S., Sarchielli, G. & Fraccaroli, F. (2010). How are psychosocial factors related to retirement intentions? International Journal of Manpower, 31 , 271-285. Örestig, J., Strandh, M. & Stattin, M. (2013). A Wish Come True? A Longitudinal Analysis of the Relationship between Retirement Prefer ences and the Timing of Retirement. Journal of Population Ageing, 6 , 99-118. I Gothenburg Studies in Work Science* 1 Jan Holmer (2003): Självvärdering som stöd för personal- och verksamhetsutveckling 2 Göran Fredriksson, Inger Humlesjö, Birgitta Jordansson och Kerstin Nordlander (2003): Könsmaktens förvandlingar: En vänbok till Anita Göransson 3 Gunnar Gillberg (2003): Nätbaserad handledning: Om problembaserat lärande och handledning på distans 4 Monika Andersson Bäck (2008): Conceptions, Conflicts and Contradictions at the Introduction of a Swedish Health Care Centre (Doktorsavhandling) 5 Birger Simonson red. (2008): Perspektiv på arbetsvetenskapens klassiker 6 Jennie Haraldsson (2010): ”Det ska vara lite äventyr”: Styrning av svensk forskarutbildning utifrån reformen 1998 (Doktorsavhandling) 7 Gunnar Gillberg (2010): Individualiseringens villkor: Unga vuxnas föreställningar om arbete och självförverkligande (Doktorsavhandling) 8 Richard Berglund (2010): Engagemang efterfrågas: Hur tre tillverkande företag söker medverkan från sina medarbetare när de inför Lean (Doktorsavhandling) 9 Gunnar Gillberg (2010): Jämställdhet inifrån: berättelsen om en metod 10 Frida Wikstrand (2011): Det tekniska spelet. Förhandlingar om arbete, teknik och kön i relation till införande av nya informationssystem (Doktorsavhandling) 11 A. Sander & T. Jordan (2011): Fanzingo – Entreprenör i samhällsförändringsarbete. Från krigare till barnmorska. 12 Hannes Kantelius (2012): Inhyrningens logik - konsekvenser för individ och organisation (Doktorsavhandling) 13 Lisa Björk (2013): Contextualizing managerial work in local government organizations (Doktorsavhandling) 14 Anna Peixoto (2014): De mest lämpade – en studie av doktoranders habituer ing på det vetenskapliga fältet (Doktorsavhandling) 15 Linda Corin (2016): Job demands, job resources and consequences for managerial sustainability in the public sector – A contextual approach. 16 Pia Andersson (2018): Making room for complexity in group collaborations: the roles of scaffolding and facilitation. 17 Linnéa Åberg (2020): Standardiseringsarbetets kollektiva praktik. En studie om att kvalitetssäkra integrationssatsningen Samhällsorientering för nyanlända. 18 Hanna Uddbäck (2021): Att stanna kvar: arbete, plats och mobilitet i småstaden. 19 Pille Strauss Raats (2021): Temporary Safety. Contextual factors behind job quality in using temporary agency work. 20 Robin Jonsson (2021): Retaining the Aging Workforce: Studies o f the interplay between individual and organizational capability in the context of prolonged working lives. * Vid sammanslagningen med Sociologiska institutionen år 2011, ändrade skriftserien namn från Skrifter från institutionen för arbetsvetenskap vid Göteborgs universitet till Gothenburg Studies in Work Science DOKTORSAVHANDLINGAR VID INSTITUTIONEN FÖR SOCIOLOGI OCH ARBETSVETENSKAP FR O M 1995 59. Börjesson, Mats: Sanningen om brottslingen. Rättspsykiatrin som kartläggning av livsöden i samhällets tjänst under 1900-talet. 1995. 60. Ljung, Margareta: Lyft jorden mot himlen - växande i kvinnogrupper och kvinnliga nätverk. 1995. 61. Björkemarken, Mariann: Implementeringsanalys som komplement vid utvärdering - en fråga om perspektiv och förklaring. 1995. 62. Håkansson, Kristina: Förändringsstrategier i arbetslivet. 1995. 63. Blomsterberg, Marianne: Garanterade karriärer? Om social styrning och sysselsättningspolitik för ungdomar. 1996. 64. Kohlström, Gulli: Identitetsförändring vid anpassning till funktionshinder/handikapp. 1996. 65. Larsson, Patrik: Hemtjänsten ur tre perspektiv - En studie bland äldre, anställda och ledning. 1996. 66. Oskarsson, Hermann: En klasstrukturs uppkomst och utveckling. Akureyri 1860-1940. 1996. 67. Thörn, Håkan: Modernitet, sociologi och sociala rörelser/Rörelser i det moderna: Politik, modernitet och kollektiv identitet i Europa 1789-1989. 1997. 68. Einarsdottir, Torgerdur: Läkaryrket i förändring. En studie av den medicinska professionens heterogenisering och könsdifferentiering. 1997. 69. Åberg, Jan-Olof: Det rationella och det legitima. En studie av utvärderingars teori och praktik. 1997. 70. Pham Van Bich: The Changes of the Vietnamese Family in the Red River Delta. 1997. 71. Lalander, Philip: Anden i flaskan. Alkoholens betydelser i olika ungdomsgrupper. 1998. 72. Eriksson, Birgitta: Arbetet i människors liv. 1998. 73. Bartley, Kristina: Barnpolitik och barnets rättigheter. 1998. 74. Nordström, Monica: Yttre villkor och inre möten. Hemtjänsten som organisation. 1998. 75. Stier, Jonas: Dimensions and Experiences of Human Identity. An Analytical Toolkit and Empirical Illustration. 1998. 76. Jerkeby, Stefan: Slutna cirklar. Om civila moståndsrörelser i Norge och Danmark 1940-45. 1999. 77. Oudhuis, Margareta: Vägen till jämlikhet. En analys av den svenska arbetarrörelsens syn på effektivitet och emancipation i arbetslivet. 1999. 78. Johansson, Anna: La Mujer Sufrida – The Suffering Woman. Narratives on Femininity among Women in a Nicaraguan Barrio. 1999. 79. Theandersson, Christer: Jobbet – för lön, lust eller andra värden . 2000. 80. Carle, Jan: Opinion och aktion. En sociologisk studie av ungdomar och miljö. 2000. 81. Öhrn, Ingbritt: Livet, identiteten och kronisk sjukdom. En socialpsykologisk studie av unga vuxna diabetiker. 2000. 82. Mossberg Sand, Ann-Britt: Ansvar, kärlek och försörjning. Om anställda anhörigvårdare i Sverige. 2000. 83. Berglund, Tomas: Attityder till arbete i Västeuropa och USA. Teoretiska perspektiv och analyser av data från sex länder. 2001. 84. Larsson, Bengt: Bankkrisen, medierna och politiken. Offentliga tolkningar och reaktioner på 90-talets bankkris. 2001. 85. Blomquist, Bo: Förskolebarnets relation till sin familj. Förändrade förutsättningar och föreställningar 1950-1990. 2001. 86. Hansen, Lars: The Division of Labour in Post-Industrial Societies. 2001. 87. Björk, Micael: Upplösningens dialektik. Bildningsmål och politisk modernitet i Sverige kring sekelskiftet 1900. 2002. 88. Gustafson, Per: Place, Place Attachment and Mobility: Three Sociological Studies. 2002. 89. Rigné, Eva Marie : Profession, Science, and State – Psychology in Sweden 1968 -1990. 2002. 90. Persson, Anders: I kräftans tecken. En historiesociologisk studie av cancerforsk- ningens samhälleliga villkor i Sverige och USA under 1900-talet. 2002. 91. Brnic, Anita: Speaking of Nationality. On Narratives of National Belonging and the ’Immigrant’. 2002. 92. Korp, Peter: Hälsopromotion - en sociologisk studie av hälsofrämjandets institu- tionalisering. 2002. 93. Sobis, Iwona: Employment Service in Transition: Adaptation of a Socialist Employment Agency to a Market Economy. A Case Study of Lodz, Poland 1989 -1998. 2002. 94. Hellum, Merete: Förförd av Eros. Kön oc h moral bland utländska kvinnor på en grekisk ö. 2002. 95. Carlson, Marie: Svenska för invandrare – brygga eller gräns? Syn på kunskap och lärande inom sfi-undervisningen. 2002. 96. Hansson, Agneta: Praktiskt taget. Aktionsforskning som teori och praktik – i spåren efter LOM. 2003. 97. Engdahl, Oskar: I finansvärldens bakre regioner. En studie om finansiella offshore- marknader och ekonomisk brottslighet. 2003. 98. Rolandsson, Bertil : Facket, informationsteknologin och politiken. Strategier och perspektiv inom LO 1976-1996. 2003. 99. Schedin, Stefan: Ekonomisk ojämlikhet. Inkomstfördelning och inkomstskillnader i Sverige under 1980- och 1990-talen. 2003. 100. Morner, Claudia G: Självständigt beroende. Ensamstående mammors försörjningsstrategier. 2003. 101. Wennerström, Ulla- Britt: Den kvinnliga klassresan. 2003. 102. Wingfors, Stina S: Socionomyrkets professionalisering. 2004. 103. Tursunovic, Mirzet: Fostran till demokrati: Tre sociologiska delstudier av bosniska ungdomars politiska socialisering. 2004. 104. Thörn, Catharina: Kvinnans plats(er) – bilder av hemlöshet. 2004. 105. Alinia, Minoo: Spaces of Diasporas: Kurdish Identities, Experiences of Otherness and Politics of Belonging. 2004. 106. Chronholm, Anders: Föräldraledig pappa – Mäns erfarenheter av delad föräldraledighet. 2004. 107. Seldén, Daniel: Om det som är. Ontologins metodologiska relevans inom positivism, relativism och kritisk realism. 2005. 108. Winell -Garvén, Irene: Vägen till Parnassen. En sociologisk studie av kvinnligt konstnärskap i Sverige 1864-1939. 2005. 109. Engström, Pär : Samtal och ledarskap. En studie av medarbetarsamtal i grundskolan. 2005 110. Löfstrand, Cecilia: Hemlöshetens politik – lokal policy och praktik. 2005. 111. Eydal, Gu∂ny Björk: Family Policy in Iceland 1944-1984. 2005. 112. Ekbrand, Hans: Separationer och mäns våld mot kvinnor. 2006. 113. Eriksson, Ylva Ulfsdotter: Yrke, Status & Genus. En sociologisk studie om yrken på en segregerad arbetsmarknad. 2006 114. Flisbäck, Marita: Att lära sig konstens regler. En sociologisk studie av osäkra framtidsinvesteringar. 2006. 115. Berntsson, Paula: Lärarförbundet, förskollärare och statushöjande strategier. 2006. 116. Latta, Mia: Public Transfer and Private Help. Support Networks of Marginalised and Poor individuals in Sweden in the 1990s. 2007. 117. Persson, Sofia: Läraryrkets uppkomst och förändring. En sociologisk studie av lärares villkor, organisering och yrkesprojekt inom den grundläggande utbildningen ca 1800- 2000. 2008. 118. Bengtsson, Mattias: Individen stämplar in. Arbetet, facket och lönen i sociologisk belysning. 2008. 119. Sjöstrand, Glenn: Gåvan i Gnosjö. Företagares relationer i ett industriellt distrikt. 2008. 120. Pellbring, Mats: Laissez -faire, systemkritik eller reformism? En studie av den svenska opinionsbildande globaliseringsdiskursen i dagspress, 1992 -2001. 2008. 121. Bjarnason, Tómas: Social Recognition and Employees’ Organizational Support. 2009. 122. Jagudina, Zaira: Social Movements and Gender in Post-Soviet Russia: The Case of the Soldiers’ Mothers NGOs. 2009. 123. Ranagården, Lisbeth: Lärares lärande om elever – en sociologisk studie om yrkespraktik. 2009 124. Shmulyar Gréen, Oksana: Entrepreneurship in Russia: Western ideas in Russian translation. 2009 125. Eriksson, Bo G: Studying ageing: Experiences, description, variation, prediction and explanation. 2010 126. Karlsson, Anette: I moderniseringens skugga? Om förändring och identitet i två administrativa serviceyrken. 2010 127. Wasshede, Cathrin: Passionerad politik. Om motstånd mot heteronormativ könsmakt. 2010 128. Bergström Casinowsky, Gunilla: Tjänsteresor i människors vardag – om rörlighet, närvaro och frånvaro. 2010 129. Jonsson, Christer: I fäders och mödrars spår. Landsortsungdomars identitetsutveckling och vuxenblivande i ett livsformsperspektiv. 2010 130. Söderberg, Johan: Free Software to Open Hardware: Critical Theory on the Frontiers of Hacking. 2011 131. Hedenus, Anna: At the end of the Rainbow – Post -winning life among Swedish lottery winners. 2011 132. Uhnoo, Sara: Våldets regler. Ungdomars tal om våld och bråk. 2011 133. Daoud, Adel: Scarcity, Abundance and Sufficiency: Contributions to Social And Economic Theory. 2011 134. Wahlström, Mattias: The Making of Protest and Protest Policing: Negotiation, Knowledge, Space, and Narrative. 2011 135. Holgersson, Helena: Icke-medborgarskapets urbana geografi. 2011 136. Ericson, Mathias: Nära inpå: Maskulinitet, intimitet och gemenskap i brandmäns arbetslag. 2011 137. Malmqvist, Karl: Offentlighetens gränser: Fem kultursociologiska studier av kontroverser kring litterära självframställningar i Sverige, 1976– 2008. 2012 138. Larsson, Jörgen: Studier i tidsmässig välfärd – med fokus på tidsstrategier och tidspolitik för småbarnsfamiljer. 2012 138. Backman, Christel: Criminal Records in Sweden. Regulation of Access to Criminal Records and the Use of Criminal Background checks by Employers. 2012. 139. Gunnarsson, Andreas: Unleashing Science Popularisation: Studies on Science as Popular Culture. 2012. 140. Kantelius, Hannes: Inhyrningens logik – konsekvenser för individ och organisation. 2012. 141. Westberg, Niklas: Meddelanden från enskildheten – En sociologisk studie av ensamhet och avskildhet. 2012. 142. Calvo , Dolores: What Is the Problem of Gender? Mainstreaming Gender in Migration and Development Policies in the European Union. 2013. 143. Widigson, Mats : Från miljonprogram till högskoleprogram – plats, agentskap och villkorad valfrihet. 2013. 144. Puaca, Gor an: Educational choices of the future. A sociological inquiry into micro- politics in education. 2013. 145. Björk, Lisa: Contextualizing managerial work in local government organizations. 2013. 146. Naldemirci, Önc el: Caring (in) Diaspora: Aging and Caring Experiences of Older Turkish Migrants in a Swedish Context. 2013. 147. Lov én Seldén, Kris ti na: Europafacklig samverkan. Problem och möjligheter. 2014. 148. Petersson, J e sper : Geographies of eHealth: Studies of Healthcare at a Distanc e. 2014. 149. Peix oto, Anna: De mest lämpade – en stu die av doktoranders habituering på det vetenskapliga fältet. 2014. 150. Stretm o, Live: Governing the unaccompanied child – media, policy a nd practice. 2014. 151. Mis cevic, Dank a: Bortom scenen - en sociologisk studie av frilansande skådespelares villkor. 2014. 152. Lil jenbe r g, Mett e. Distributed Leadership in Local School Organisations. Working for School Improvement? 20 15. 153. Nordlander, Erica. On the Mechanisms of Social Inequality. Studies of young people’s educational outcomes, social participation, and well-being. 2015 . 154. Nordholm , Daniel. Organising for School Improvement at the Middle Tier. Studies on Temporary Organisation. 2015. 155. Vulkan, Patrik. The Microfoundations of Flexicurity. Employees’ well -being and attitudes to labour market policy in a Swedish and Nordic welfare state setting. 2016. 156. Corin, Linda. Job demands, job resources and consequences for managerial sustainability in the public sector – A contextual approach. 2016. 157. Boreli us, Ulf. Om befrielseteologins uppkomst i Latinamerika. En sociologisk analys av religiös förändring . 2016. 158. Törnber g, Anton. The Wicked Nature of Social Systems – A Complexity Approach to Sociology. 2017. 159. Reichenber g, Olof. A Mechanism Approach to the Sociology of Teachers’ and students’ Actions: Teaching Practice, Student Disengagement and Instructional Materials. 2017 160. Lyd ahl, Doris. Same and different? Perspectives on the Introduction of Person-Centred Care as Standard Healthcare. 2017. 161. Björk, Sofia. Gender and emotions in family care – Understanding masculinity and gender equality in Sweden. 2017. 162. Wall inder, Ylva. Imagined Independence. Institutional Conditions and Individual Opportunities in European Labour Markets . 2018 163. Andersson, Pia . Making Room for Complexity in Group Collaborations: The Roles of Scaffolding and Facilitation . 2018. 164. Bajqi nca, Nuhi. Mother Tongue Education – The Interest of a N ation. A policy study in Sweden 1957-2017. 2019 . 165. Hassel gren, Ca roline. Inequity in Mind. On the Social and Genetic Risk Factors of Dementia and their Interactions . 2019. 166. Rosqui st, J ohan. Moral i rätten. Utredningar av hedersrelaterat våld i Sverige 1997- 2017. 2020. 167. Åber g, Linné a. Standardiseringsarbetets kollektiva praktik. En studie om att kvalitetssäkra integrationssatsningen Samhällsorientering för nyanlända. 2020. 168. Alfonsson, J ohan. Alienation och Arbete, Unga behovsanställdas villkor i den flexibla kapitalismen. 2020. 169. Uddbäck, Hanna. Att stanna kvar: arbete, plats och mobilitet i småstaden . 2021. 170. Strauss Raats, Pil le. Temporar y Sa fet y. Contextual factors behind job quality in using temporary agency work. 2021 171. J onsson, Robin. Retaining the Aging Workforce: Studies of the interplay between individual and organizational capability in the context of prolonged working lives. 2021. DEPARTMENT OF SOCIOLOGY AND WORK SCIENCE Robin Jonsson is affiliated with the Department of Sociology and Work Science at the University of Gothenburg, Sweden. DEPARTMENT OF SOCIOLOGY AND WORK SCIENCE Robin Jonsson is affiliated wit the Department of S ciology and Work Science at the University of Gothenburg, Sw den. ISBN 978-91-87876-39-4 (PRINT) ISBN 978-91-87876-40-0 (PDF)