DEPARTMENT OF POLITICAL SCIENCE A PENNY FOR LIVES LOST? Analysing the Relationship between Development Financing for Sustainable Development Goal 16 and Battle-Related Deaths Lean Carlo Luis Ong Ipac Master’s Thesis: 30 credits Programme: Master’s Programme in International Administration and Global Governance Date: 13 February 2025 (revised 25 March 2025) Supervisor: Magnus Lundgren Words: 18021 Table of Contents Contents Table of Contents .................................................................................................................................... 2 Acknowledgements ................................................................................................................................. 4 List of Acronyms ..................................................................................................................................... 5 Abstract ................................................................................................................................................... 6 Introduction ............................................................................................................................................. 7 Literature Review .................................................................................................................................. 10 Development financing: inducing civil war ...................................................................................... 11 Reducing violence with development financing ............................................................................... 14 The case of ineffective development financing against civil war ..................................................... 18 Institutional Reform and Quality: ...................................................................................................... 18 Summing up the literature ................................................................................................................. 20 Theoretical Framework ......................................................................................................................... 22 Linking institutions and development ............................................................................................ 22 Institutions and civil war ............................................................................................................... 23 SDG 16 development financing and civil war ............................................................................... 24 Limitations of the theory ................................................................................................................ 26 Research Data and Design ..................................................................................................................... 28 Variables ....................................................................................................................................... 28 Preparation of Variables .................................................................................................................... 37 Statistical Technique ......................................................................................................................... 40 Results and Discussion .......................................................................................................................... 42 Base model testing: ........................................................................................................................... 42 Hypothesis: SDG 16 development financing has a negative association with battle related deaths with state capacity as a moderator variable ....................................................................................... 45 Hypothesis: SDG 16 development financing has a negative relationship with battle related deaths with military expenditure as a percentage of GDP as a moderator variable ..................................... 48 Hypothesis: SDG 16 development financing has a negative relationship with battle related deaths with state capacity and military expenditure as a percentage of GDP as moderator variables ......... 50 Robustness test .................................................................................................................................. 52 Conclusion ............................................................................................................................................. 54 Bibliography .......................................................................................................................................... 56 Appendices ............................................................................................................................................ 67 Appendix A: Sustainable Development Goal 16 – Common Reporting System Purpose Codes ..... 67 Appendix B: List of donor countries ................................................................................................. 69 Appendix C: List of multilateral organisations donors ..................................................................... 71 Appendix D: List of Private Donors in SDG Financing Database .................................................... 73 Appendix E: List of recipient countries in the study ......................................................................... 75 Appendix F: Stata Do-File for Zero-Inflated Negative Binomial Regression and result of Likelihood Ratio test ............................................................................................................................................ 79 Acknowledgements I would like to express my sincerest gratitude to the following: 1) To the Swedish Institute, for providing the opportunity to learn alongside the smartest peers and to be thought by the wisest teachers in Gothenburg University. My deepest thanks for affording me to make more sense in this complex world. 2) Salamat sa aking pamilya: sa aking Daddy, Mommy, Kuya, at Ate. You have provided me the strength support to go through this journey. Your sacrifices and efforts are recognized and honored in this work. 3) To Magnus Lundgren, my thesis supervisor. I apologise for the inconvenience that I have caused. I am thankful for your unending compassion and grace that you gave despite the hurdles of the past year. 4) Salamat sa Gothenburgers!!!! Amay, Dianne, Boss Kenneth, Jessy, Merl, Juno, JL, Roselle, Tom, Anjelou, Deb, Bonn, Lloyd, Lars, and Krishna. May we continue to enjoy our blessings and give relief to each other struggles in every gathering. 5) To Tabitha and Manel. Our three-person group chat provided laughter, relief, and understanding of the world. May we continue to enrich our lives with love, fortune, and light. 6) Tack så mycket to my classmates in the IAGG. We have struggled together in this endeavour. May we find success and happiness in our future. Special mention to Vlad, Jorick, Alice, Marius, Johanna, Nadyn, Lisa, Mona, Cecilia, and Roosa. You have made this journey meaningful and fun. List of Acronyms AFP Armed Forces of the Philippines CBI Central Bank Independence CCT Conditional Cash Transfer CDD Community-Driven Development CERP Commander’s Emergency Response Program CSP Community Stabilization Program FARC Fuerzas Armadas Revolucionarias de Colombia FASD Financing the 2030 Agenda for Sustainable Development GDP Gross Domestic Product MDG Millenium Development Goals NPA New People’s Army NREGA National Rural Employment Guarantee Act OECD CRS Organisation for Economic Co-operation and Development – Creditor Reporting System OECD DAC Organisation for Economic Co-operation and Development - Development Assistance Committee QOGD Quality of Government Database PRIO Peace Research Institute Oslo RGC Revenue Generation Capacity SATP South Asian Terrorism Portal SC State Capacity SDG Sustainable Development Goals SDGI Sustainable Development Goals Index SDG 16 Sustainable Development Goals 16 UCDP GED Uppsala Conflict Data Program Georeferenced Event Dataset USAID United States Agency for International Development Abstract This thesis is an attempt to test whether Sustainable Development Goal 16 development financing has a relationship with civil war in a country per year. It builds on the scholarship of development financing literature and the focus on the institutional effects on development as studied by Fjelde, Knutsen, and Nygård (2021). The literature on development financing has been focused on the tangible outcome of development. Studies on its intangible outcomes, such as financing institutional reform, and its effects to civil war have been lacking. To address this gap, this research used the Financing the 2030 Agenda for Sustainable Development database by AidData and tested it with data from the Uppsala Conflict Database Battle Related Deaths Database. State capacity and military expenditure were used as moderator variables in the study, as the literature identified its moderating role on the effect of development financing and civil war. Using a Zero-inflated Negative Binomial Regression with a lagged independent variable for one year (development financing), the analysis reveals that SDG-16 can likely reduce battle related deaths when there is high levels of state capacity. Liberal democratic governance and gender equality demonstrated statistically significant results to likely reduce and prevent civil war. Thus, policies increasing levels of state capacity in relation with SDG 16 objectives, liberal democratic governance, and gender equality elicits the likelihood of reduction and prevention of civil war. The effect of SDG 16 development financing can be studied further when 1) a more disaggregated data for SDG financing is released (forthcoming per AidData), and that 2) theoretical precedents must be established to account the best time for lagging development financing for institutional reform and its effects to development and/or civil war, and 3) expand control variables in statistical models to potentially address issues of endogeneity with SDG 16 financing. Introduction What is the relationship between development financing and civil war? The end of the Millennium Development Goals (MDGs) in 2015 brought lessons that necessitated broader, grander, and more participatory ways of working to address poverty, climate crisis, and other problems in society (Woodridge, 2015). Thus, the 193 United Nations member states adopted the 17 Sustainable Development Goals (SDGs) in September 2015 (United Nations, 2015). One of the 17 goals is Sustainable Development Goal 16 (SDG 16) - Peace, Justice, and Strong Institutions. Fundamentally, the goal seeks to bring about a peaceful society by establishing accountable, effective, and transparent institutions that can provide justice, uphold human rights, foster inclusivity, and reduce violence (United Nations, 2024). Significant resources have been disbursed to achieve this goal. According to the database of Burgess, Bengtson, and Lautenslager (2023), 2.289 trillion USD have been disbursed by donor countries, multilateral international organisations, and selected private sector actors from 2010 to 2021 for the Sustainable Development Goals. 273 billion USD or 11.92% have been disbursed for SDG 16, the second highest among the 17 goals just behind SDG 3 - Good Health and Well-Being. This study is an effort to understand the effects of SDG-16 financing to the magnitude of civil war. SDG-16 financing involves investments to improve governance institutions and organisations, conflict prevention and reconciliation, peacebuilding, and disaster risk mitigation, reduction and management (Burgess, Bengston, and Lautenslager, 2023). Studies on the effects of development financing on security outcomes have yielded mixed results. Generally, causal studies have identified a positive, negative or non-significant relationship between development financing and the frequency and/or magnitude of organized civil war. A positive relationship was observed in studies that analysed humanitarian aid (Nunn and Qian, 2014; Wood and Sullivan, 2015; Wood and Molfino, 2016; Narang, 2014; and Narang, 2015), conditional cash transfers or CCTs (Zizumbo-Colunga, 2023; Weintraub, 2016), community- driven development programmes or CDDs (Crost, Felter, and Johnston, 2014; Khana and Zimmerman, 2017, Sexton, 2016). Negative relationships were observed in studies that looked into employment programmes (Dasgupta, Gawande, and Kapur, 2016 and Fetzer, 2014), some CDDs (Berman et. al., 2013 and Beath, Fotino, and Enkipolov, 2012), and some CCTs (Crost, Felter, and Johnston, 2016). There were also studies that yielded insignificant changes between the two variables (Child, 2012 and 2014). However, the general trend of the literature is focused on the tangible outcomes of aid and its effects to civil war. Limited studies on the intangible outcomes of aid have been made. This study postulates institutional reform as an intangible outcome of development financing. Since SDG 16 is focused on reforming institutions, it forwards a theory exploring the relationship of financing institutional reform and deaths caused by civil war. Notably, studies made on aid efficacy generally agree that the degree of state capacity and/or military presence in an area where development financing towards peacebuilding is implemented plays a significant conditioning role to determine the relationship. Higher degrees of state capacity and/or military presence reduces the magnitude, persistence, and incidents of civil war when aid is given. Inversely, lower degrees of state capacity and/or military presence can increase levels of civil war whether as retribution, capture of resources, weakening of state legitimacy, among others. These variables are then deployed as moderators in the study of development financing for institutional reform (SDG 16 financing) and the magnitude of civil war (battle- related deaths) This study contributes to the literature of development financing and civil war by analysing country-level effects of development financing by measuring aid aimed toward institutional reform. It is an attempt to understand whether the amount of development financing for SDG 16 can influence the number of deaths caused by civil war at the country level. It expands the current discussion on development financing scholarship by focusing on the intangible outcome of development financing as seen in SDG 16 aid and its effects to civil war. It builds on the literature of the effect of institutions to civil war as studied by Fjelde and Knutsen. It posits the theory where SDG 16 financing, a development agenda focused on policy reform, affects civil war. It will analyse the general trend with country-level data using variables that can influence the flow of aid, and the number of battle related deaths using a zero-inflated negative binomial regression. The study used a new database made by AidData where machine learning was used to encode development financing data from the Organisation for Economic Co- operation and Development - Development Assistance Committee (OECD-DAC) member countries, non-OECD DAC member countries (except China), and selected charities. Data from the Uppsala Conflict Data Program Battle Related Deaths Database is used to account for battle related deaths. It also deployed state capacity and military capacity as moderator variables as identified by the literature as moderating variables. The control variables to be used are ethnic fractionalization, liberal democratic governance, deaths caused by natural calamities, education expenditure as a percentage of GDP, health expenditure as a percentage of GDP, building human resources rating, and gender equality rating. The analysis revealed that SDG 16 development financing affects battle related deaths when there is a high level of state capacity. However, when interacting with military capacity, it was observed to have a statistically insignificant association to battle related conflict. Liberal democratic governance and gender equality tend to reduce and prevent battle related deaths. This suggests that enhancements for strengthening state capacity in line with the objectives of SDG 16, promote gender equality, and enrich liberal democratic governance are key in reducing and preventing civil war. Further disaggregation of the SDG 16 development financing is recommended to yield nuanced analysis on the effects of specific institutional reforms and its corresponding funding to civil war. The model should also expand to include other variables to account for other factors that affects the relationship of the two main variables. Still, state capacity and military capacity are observed to be critical moderators and determinants for civil war. Literature Review In this section, I survey the existing literature investigating the relationship between development financing and civil war. The studies discussed are selected because they directly examine a relationship between development financing and civil war. 17 studies were included in this review, published from 1991 to 2023. This study defines development financing or aid based on the definition of the Organisation for Economic Co-operation and Development (OECD) definition of official development assistance (ODA). OECD defines ODA as “financial support - either grants or "concessional" loans from OECD-DAC (Organisation for Economic Co-operation and Development - Development Assistance Committee) member countries to developing countries” (OECD, 2024). This study expands this definition to include non-OECD DAC member countries, other multilateral organisations, and private charities that finance SDG projects and programmes. Studies on the efficacy of development financing on development outcomes have provided rich analysis for scholars and policymakers. Burnside and Dollar’s (2000) seminal work on the effects of foreign aid on economic development underscores the importance of effective institutions for growth. It accounts for the fact that foreign aid can lead to macro- level changes in countries. The argument is echoed by Kotsadam et. al. (2018) who observed that ODA reduces infant mortality rates in Nigeria. Dreher et. al. (2021) also resonates with Burnside and Dollar’s argument that observed Chinese development financing boosts short- term economic growth. Banerjee et. al. (2015) also observed the effectiveness of multifaceted foreign-funded development programmes in improving conditions in Ghana, Ethiopia, Honduras, India, Pakistan, and Peru. Within this general debate, there has been a focus on the effect of aid on civil war. Scholarship on the relationship between development financing and organised civil war has yielded varying causal analyses. Grossman (1991) argued that foreign aid could increase the risks of civil war among rebel groups due to increased financial resources of governments that induce rent-seeking behaviour. De Ree and Nillesen (2009) contends that the amount of foreign aid has a negative relationship with the probability of prolonged violent conflict. However, they also claimed to observe that there is no statistically significant result to indicate that foreign aid can affect the start of violent conflict. Diving deeper into the diverging arguments, the succeeding sub-sections will discuss attempts to establish causality between development financing and civil war. Three strands are observed by the literature in this field so far: 1) the violence-inducing effect of aid, 2) it’s ability to reduce civil war, 3) and the possibility where no significant effects were observed. Across the discussion, the feature of state capacity of the recipient area provides a salient feature the conditions the relationship of the two variables. Development financing: inducing civil war Several statistical studies to understand the relationship between development financing and civil war observed a positive relationship between the two. Chiefly among the literature are studies on humanitarian aid which are mostly attributed to having an increasing effect on civil war. Nunn and Qian (2014) studied the correlation of US Food Aid in 125 non-OECD countries using panel data of violent incidents and publicly funded US food aid disbursement. They found that US Food aid (wheat) is correlated with the increase in the duration of civil war among the countries studied but does not affect its onset. Wood and Sullivan (2015) have observed increasing violent incidents in Sub-Saharan Africa due to humanitarian aid. Wood and Sullivan used UCDP GED to approximate the effect humanitarian aid, from publicly funded bilateral and multilateral disbursement data from AidData and analysed it through cross-sectional and country-level differences-in-differences statistical technique from 1989 to 2009. Wood and Molfino (2016) echo the same findings of Wood and Sullivan (2015) with a more precise model that measures the amount of aid and violent incidents by disaggregated politico-administrative districts using the same database. The authors conducted panel data regression analyses and zero-inflated Poisson quasi-experimental model of humanitarian aid received/disbursed to politico-administrative units in Sub-Saharan Africa from 1990 to 2008 in relation with humanitarian aid. The results suggest that there is a positive relationship between the frequency of battles engagements with the amount of aid received in Sub-Saharan Africa. Narang (2014) observed that humanitarian aid may shorten peace duration in a war that ended with a decisive victory. They argued that under such conditions of victory, the losing side tends to receive more humanitarian aid that will help them regain the political and military capital to restart a war; an observation that is not found in wars that ended through negotiated settlements. Narang (2015) also yielded substantially the same conclusions in their panel data analysis of humanitarian aid and the duration of civil war from 1969-2008. Humanitarian aid, especially in areas located on the fringes of state control, tends to prolong civil wars as they have higher risks of misappropriation that can contribute to galvanising resources for rebel forces. Studies on conditional cash transfers (CCT) such as the Bolsa Familia in Brazil and the Pantawid Pamilyang Pilipino Program in the Philippines have also yielded positive relationships with civil war. Conditional cash transfers are programs that provide cash grants to generally poor households on the condition that their children are sent and retained in school, inoculated with mandatory vaccines, and participate in social development programmes, among others (Fsizbein, Schady, and Ferreira, 2009). These programs have been observed to improve the socioeconomic conditions of poor households (Fsizbein, Schady, and Ferreira, 2009). However, they have been observed to increase incidents and magnitude of organised violence in Latin America (Zizumbo-Colunga, 2023). Zizumbo-Colunga (2023) studied the effects of PROGRES/Opportunidades in Mexico where its sudden expansion is linked to increased violent incidents due to low rule of law. They also surveyed the rest of South America in their study and observed that CCT programmes increase the probability of violent incidents in the continent. Weintraub (2016) echoes this argument in their study of the Familias en Acción, a CCT programme, in an experimental set-up involving 122 municipalities in Colombia. They used data from the Human Rights Observatory under the President of Colombia for the study. Weintraub argued that the increase in killings done by the Fuerzas Armadas Revolucionarias de Colombia (FARC) are retributive actions aimed at decreasing the level of cooperation between the beneficiaries of the CCT programme and the government. further. Another form of a development programme that has been found to have a violence- increasing effect on civil war is community-driven development (CDD). CDD is a form of development programming where external factors, such as the state, a donor, or financial institution, provide financing and/or capacity development for local communities to sustainably implement projects and programmes (Boräng, Grimes, and Ahlborg, 2022). They are aimed at achieving development goals such as the provision of basic utilities and services, reducing poverty, and encouraging local ownership. Crost, Felter, and Johnston (2014), studied the Asian Development Bank’s National Community-Driven Development Project in the Philippines (KALAHI-CIDS) in an experimental set-up involving 4000 barangays (villages) in 184 municipalities. Panel data analysis was conducted from 2003 to 2008. The study used information from project documents to 1) determine the amount disbursed to villages that functioned as the independent variable, and 2) develop a standard for “treated villages” based on their eligibility for the program. Civil war data were sourced from the Significant Activities database of the Armed Forces of the Philippines (AFP) and used as the dependent variable. The study found that KALAHI-CIDS is attributed to the increase in violent incidents done by the New People’s Army (NPA) among the treatment villages. However, the authors provided a caveat that the KALAHI-CIDS may not be a cause of the spike in incidents of civil war. Rather, it provided incentives for communist rebels to sabotage the cooperation that KALAHI-CIDS might foster between the government and poor communities, a political base for the NPA. This resonates with the findings of Khana and Zimmerman (2017) where a cash-for-work programme in India induces violence against civilians by Maoist rebels as a form of retaliation, therefore increasing civil war incidents in the short run. Retributory actions instigated by rebel groups against development programmes contribute to their political consolidation and military gain. The Commander’s Emergency Response Program (CERP) is another form of development financing with an explicit aim to reduce violent conflict in countries under civil war. CERP is a flexible financing program of the United States military operating in Iraq and Afghanistan that provides immediate humanitarian and reconstruction needs (Special Inspector General for Afghanistan Reconstruction, 2018). CERP has been documented to have violence- inducing effect in Afghanistan. Sexton (2016) argued that insurgents seek to sabotage CERP projects in communities with low levels of government control. They arrived at this conclusion by using week-by-week CERP spending per district and the violent incidents that occurred, while controlling for troop presence. Although CERP indicated a violence-reducing effect in districts where there are high levels of government control, they observed that CERP can trigger violent incidents from rebels in areas where there is low government control or military presence. Adams (2015) supports the argument of CERP’s violence-increasing effect in Afghanistan. Empirical analyses reveal that CERP can significantly reduce violence in Afghanistan under conditions of small-scale (under 50,000 USD), timely, and flexible projects. However, CERP can increase violence in areas where projects that are implemented are large- scale (over 50,000 USD) with low levels of administrative capacity significantly increases violence. Adams also conducted interviews of nine Civil Affairs Officers who had operational knowledge of CERP in Afghanistan. The interviews reveal that CERP’s inflexibility may add to the risks of increasing incidents of civil war. It can be seen in the literature discussed above that the violence-increasing effect of development financing is mediated by or made more profound by one or more factors. These include 1) the relative weakness of military or state control where development programmes or projects are implemented as compared to situations where there are highly functioning military and state organisations in control (Zizumbo-Colunga, 2023; Narang 2015; and Sexton, 2016), 2) the incentives that development financing can potentially provide rebel forces to exact retribution or sabotage against the state and military (Weintraub, 2016; Crost, Felter, and Johnston, 2014; and Khana and Zimmerman, 2017), 3) whether the conditions for victory in post-conflict societies were gained through decisive military action or a negotiated settlement (Narang, 2014), and 4) the scale and degree of flexibility that development financing has in terms of implementing its programmes and projects (Wood and Molfino, 2016 and Adams, 2015). These findings represent the complexity of causality in the study of aid, especially in the analysis of its effectiveness to prevent, cease, or shorten civil war. Reducing violence with development financing Some studies have observed a negative relationship between development financing and civil war. Most notably, studies on the relationship between aid directed towards employment programmes and civil war have observed a trend of the former’s violence-reducing effect. One example is the National Rural Employment Guarantee Act (NREGA) of India. Scholars have analysed how NREGA influences the number of incidents of civil war in different parts of the country. The NREGA is a 100-day cash-for-work programme of the Mahatma Gandhi Ministry of Rural Development where the unemployed work in unskilled manual labour (Mahatma Gandhi Ministry of Rural Development, 2012). In an experimental study, Dasgupta, Gawande, and Kapur (2016) analysed the effects of NREGA on civil war from 1999 to 2009 sourced from media reports in 144 districts in India. They found that NREGA has a significant and long-term violence-reducing effect on districts where the programme is implemented, especially in cities with a relatively high state capacity (Dasgupta, Gawande, and Kapur, 2016). Furthermore, the programme has a higher violence-reducing effect in areas affected by extreme weather conditions. Dasgupta, et. al. theorized that income provided by NREGA to the unemployed decreased the incentives to participate in the Maoist insurgency. Fetzer (2014) supports this argument and argues that NREGA reduces civil war incidents and intensity in different parts of India. Using the South Asian Terrorism Portal (SATP) for conflict data as the dependent variable, Fetzer conducted regression analyses with 1) NREGA participation from the Management Information System of the government, 2) weather data from the Tropical Rainfall Measuring Mission, and 3) wage data from the Agricultural Wages India. They found that NREGA provided a source of livelihood for poor households in India, especially in climate- dependent agricultural regions. The reduction in civil war incidents and intensity can be linked to the mitigation of income shocks among agricultural households brought about by extreme weather conditions. This reduces the incentive to participate in Maoist rebel political galvanisation, thus lowering civil war incidents and intensity (Fetzer, 2014). A study on CCT has also observed its violence-reducing effect. In an experimental study done by Crost, Felter and Johnston (2016) of 130 villages in the Philippines, they argued that conditional cash transfers have substantially decreased conflict-related incidents in villages that have CCT programmes compared to those that have not. The Pantawaid Pamilyang Pilipino (PPP) programme was a World Bank funded human development programme patterned after CCT programmes in South America such as the Bolsa Familia in Brazil (Official Gazette of the Philippines, 2024). In the study, the villages were randomly divided into two; 65 villages that received cash grants through the PPP were designated as the treatment group and the other 65 villages that were not beneficiaries of the PPP were assigned to the control group. The findings show that there was a reduction in the civil war incidents in the control villages. Crost, et. al. theorised that this may be explained by 1) the higher costs belligerents shoulder to operate in treatment villages and/or 2) the PPP was instrumental in developing trust in the government, thus reducing the political persuasion of the rebel forces. Studies on CDD have also observed its violence-reducing effect. One example is the United States Agency for International Development (USAID) Community Stabilization Program (CSP) in Iraq. The CSP is a 41-month-long programme implemented in Iraq from May 2006 to October 2009. It consisted of financing the building of essential infrastructure, vocational training for long-term employment, training on business development, and social programmes for the youth (United States Agency for International Development, 2010). Berman et. al. (2013) argued that the programme provides violence-reducing effects when 1) projects are small-scale and targeted, 2) the area is militarily secured, and 3) there is professional development expertise among implementers. Beath, Fotino, and Enikopolov (2012), observed the same patterns in Afghanistan in the case of the National Solidarity Program (NSP). The NSP was a World Bank-funded, nationwide, community development- focused programme that empowered rural communities to initiate, lead, and manage small- scale development projects within their communities through capacity development and provision of grant funding for said projects (Centre for Public Impact, 2016). Beath et. al. (2012) was commissioned for the study by the funder to do an experimental study involving 500 villages within 10 districts. The villages were divided into 2 groups: 250 villages were the treatment group that received an average of 30,000 USD from the NSP programme while the other 250 were the control group. Beath et. al. observed a decrease in violent incidents among the villages in the treatment group in the eight districts while no significant effect was observed in the treated villages of the other two districts. The villages in the eight districts were found to be relatively more secure as compared with the villages in the two districts. Contrary to Maslow’s theory, the authors argued that the NSP was key in the provision of public goods in the eight secured districts, thus satisfying their economic needs which may lead to heightened trust towards the government, military, and NGOs (Beath, Fotino, and Enikopolov, 2012). In the treatment group of the other two districts, the priority for their general sense of security was lacking in the community, thus public goods provision did not elicit the intended effect of reducing violent incidents. Studies have also shown that CERP has a violent-reducing effect in Iraq. Berman, Shapiro and Felter (2011) observed how CERP can have a violence-reducing effect through their study that employed a first-difference design of the changes in aid disbursement given through CERP and the changes in the number of attacks against United States and Iraqi government. They regressed the values while controlling for troop strength and previous incidents of violence. They found that CERP projects under 50,000 USD has a significant violence-reducing effect in the areas where they are implemented (Berman, Shapiro, and Felter, 2011). Further improving the study, Berman, Shapiro, Felter, and Tolman (2013) introduced the control variable of troop presence. They arrived in the same conclusion where CERP projects reduce civil war incidents if there is a large troop presence in the area. The literature on the negative relationship of aid and civil war further provides validity on the claim the mediating factors influence the relationship between the two. State control and military presence play a critical role on the effect of development financing to reducing incidents, persistence, and magnitude of civil war. This was exhibited in the studies of Berman, Shapiro, Felter and Tolman (2013), Beath, Fotino, and Enikopolov (2012) and Dasgupta, Gawande, and Kapur, (2016), and Crost, Felter, and Johnston (2016). As discussed in the preceding section, lower levels of state capacity results to higher incidents and magnitude of civil war due to aid. Inversely, the discussion in this section proves that higher state capacity reduces the likelihood of aid causing more civil war. This illustrates the point of the importance of including a measure of state capacity of the country in the model that this study will use in its analysis. Military presence also reduces the likelihood of aid resulting to more incidents and magnitude of civil war. Berman, Shapiro, Felter and Tolman (2013) demonstrate this point in their study. As mentioned in the preceding section, lower levels of military presence and control in an area increases the likelihood of aid encouraging acts of civil war. Otherwise, higher levels of it reduces the likelihood of aid as an aggravating factor for civil war. This point further proves the inclusion of military presence as a control variable in this study. The disincentives for rebel forces to launch civil war operations influenced by aid figures saliently in this section. Crost, Felter, and Johnston, 2016 theorised how CCT can make military operations for rebels more costly. Taken alone, this point is tangent to the arguments made in the preceding section where aid can provide incentives for retribution and retaliation from rebel forces. However, Crost et. al. was clear that this must be coupled with high levels of state capacity for CCT to become a disincentive for rebels. CCT here is therefore a tool to further augment state capacity in communities by providing social services and trust in government. This makes CCT, or development financing, as a mechanism to enhance state capacity. Hence, it can be theorised that state capacity can be proxied as an incentive/disincentive that can influence the motivations behind the act of civil war. A peculiar confounding variable discussed in this section is weather condition. Dasgupta, Gawande, and Kapur, 2016 theorised how extreme weather conditions can cause shocks that make communities vulnerable to civil war. This point suggests that weather condition, or the severity of it, must be accounted in the model of this study. The case of ineffective development financing against civil war Some studies have also observed the ineffectiveness of development financing to affect civil war. Chou’s (2012) study on the CERP and CDD programmes in Afghanistan observed its ineffectiveness to affect organised civil war. Using civil war data from the United States military’s Combined Information Data Network Exchange, Chou performed regressions with data on public spending for these development programmes and found that they have no significant effect to civil war in Afghanistan. However, when tested for small-scale projects (under 50,000 USD), it was observed that these programs exhibited a stabilizing effect. Child (2014) arrived at a similar conclusion with CERP in Afghanistan and found it to be ineffective in affecting civil war. Using the Worldwide Incidents Tracking System for their civil war data, Child performed panel regressions from 2005 to 2009 across 227 districts in Afghanistan with CERP spending data from the North Atlantic Treaty Organisation’s Afghanistan Country Stability Picture database. The study found that CERP has statistically nil influence on civil war and that its effect on violence is contingent to other factors (Child, 2014). It can be derived from this section that the causal or correlational link CERP and organised civil war in Afghanistan are mediated by intervening factors. Whether it is the scale of the project as observed by Chou (2012) as underscored by Child (2014), it is important to account for these mediating mechanisms in developing a theory to understand the relationship between development financing and organised civil war. This is done to come up with a more precise model to empirically explain the relationship between the dependent and independent variable that this study seeks to understand. Hence, an expansive statistical model that accounts for the variables discussed in this section and in the preceding two must be considered in undertaking this task. Institutional Reform and Quality: Among the various forms of development financing discussed in the previous sections of this review, it can be noted that apart from being publicly funded, they are focused in delivering tangible public goods to communities. Whether it was infrastructure, employment, cash, or humanitarian aid, development financing has been biased for the logistically convenient measurement of change. This section digresses from that discussion focuses on institutional reform as a catalyst for peacebuilding. Acemoglu, Johnson and Robinson (2005) have underscored that institutions are key to economic growth provided that political institutions are effective in controlling elite capture, rent-seeking behaviour, and enforce property rights. Reforming these institutions have also been proven to elicit changes to society. Zhao, Madni, Anwar, and Zahra (2021) demonstrated this point in their study of 122 developing countries’ impacts of institutional reform to economic growth with a panel data from 1996 to 2019. They deployed a difference-in-difference analysis of the changes in their measures of economic freedom as measured by the Heritage Foundation and degree of democracy provided by the POLITY IV dataset. They concluded two salient observations: 1) countries who focused first on political reform had slower rates of economic growth, and 2) countries that prioritized economic reform enjoyed higher growth and investments through a strong and efficient institutional structure. Indeed, the literature in the preceding sections have underscored the fundamental role of state capacity. While the literature discussed have varying definitions of state capacity; from institutional quality, bureaucratic capacity, rule of law, or military power, among others, this paper shall use Hendrix’s (2010) definition of state capacity. Hendrix argues that the measuring state capacity must have multivariate dimensions that can apprehend the effectiveness of its processes to extract resources with its rentier autocratic tendency and possess the rationality to utilize it. Hence, they propose that state capacity is most aptly measured by its percentage of tax revenue relative to its gross domestic product and its bureaucratic quality. This definition has been loosely applied in studies that analyse the relationship between state capacity and civil war. Wade (2017) found the importance of establishing institutional state capacity, particularly in accountability and control of donor funds, to be important in their qualitative case study of post-conflict Liberia. Wade’s theory of state capacity identifies itself more on the bureaucratic quality spectrum of Hendrix’s definition. Meanwhile, Sobek’s (2010) analysis identifies itself more on the extractive capacities and Weberian quality of state capacity. They observed in their review of state capacity literature that states that can address the demands of its citizens, address grievances, and repel challenges to authority can minimize civil war. However, they recognise the endogeneity of the relationship between state capacity and civil wars. Civil wars weaken state capacity and state capacity can prevent or cease civil wars (Sobek, 2010). As mentioned in the previous sections, state capacity plays a significant role in the risk of countries to plunge into civil war. The literature on institutional analyses of state capacity and its relationship to civil war generally argues for its positive relationship between the two. Furthermore, reforms in institutions yield desirable changes within states, despite the variances of the duration of its onset. Summing up the literature The literature review presented the divergent arguments and varied methodologies that illustrate the desired and undesired impacts of development financing to organised civil war. In the same breadth, it also demonstrates the significant effects of mediating variables to the relationship between the two. Generally, three strands of relationship were observed by past studies. Humanitarian aid (Nunn and Qian, 2014; Wood and Sullivan, 2015; Wood and Molfino, 2016; Narang, 2014; and Narang, 2015), CCTs (Zizumbo-Colunga, 2023; Weintraub, 2016), CDD programmes (Crost, Felter, and Johnston, 2014; Khana and Zimmerman, 2017, Sexton, 2016) were observed to have a positive relationship with civil war, provided that military presence and/or state capacity were weak in the areas where they are implemented. Inversely, employment programmes (Dasgupta, Gawande, and Kapur, 2016 and Fetzer, 2014), CDD programmes (Berman et. al., 2013 and Beath, Fotino, and Enkipolov, 2012), and CCTs (Crost, Felter, and Johnston, 2016) provided evidence on their negative relationship with civil war if state capacity and military presence is strong in areas where they are implemented. Interesting, development financing may not yield significant results if it lacks scale to effect security and development dynamics of an area (Child, 2012 and 2014). The studies show how state capacity (e.g. institutional quality, military presence and power, degree of government reach, among others) significantly effects the relationship of aid and civil war (Nunn and Qian, 2014, Wood and Sullivan, 2015; Wood and Molfino, 2016; Narang, 2014; Narang 2015; Zizumbo-Colunga, 2023; Weintraub, 2016; Crost, Felter, and Johnston, 2014; Khana and Zimmerman, 2017; Dasgupta, Gawande, and Kapur, 2016; Berman et. al., 2013; and Beath, Fotino, and Enikopolov, 2012). These studies have shown that higher levels of state capacity result to more the likelihood of more effective aid to reduce the onset, frequency, duration, and magnitude of civil war. Wade (2017) and Sobek (2010) emphasized the role of the state, particularly its institutions, in effectively reducing incidents and magnitude of organised civil war. Furthermore, reforms towards economic openness and democratisation elicit levels of organised civil war. It can be theorised that the intrinsic value of state capacity to effectively provide and protect its citizens mitigate the risks of onset, exacerbation, and prolongment of civil war. Given this, the efficacy of development financing to achieve its desired result is likely to be significantly influenced by state capacity of the receiving area to implement development programmes, including the management of civil war. However, a salient gap in the literature on aid and conflict is that studies that try to establish causality between the two have been focused on the tangible outcomes of development financing. Indeed, intangible outcomes of development financing logistically and theoretically difficult to parsimoniously and precisely quantify. As demonstrated in the literature review, the scholarship has been focused on development financing disbursed for conditional cash transfers, community development programmes, funding salaries for employment, and infrastructure. The focus on tangible outcomes of aid has turned a blind eye on the effects of financing institutional reform to civil war. Studies on the effects of institutions on civil war have been centred on its political determinants to control executive power, legitimacy and trust, elections and regime type (Deglow and Fjelde, 2021). The lacuna on the effect of development financing for institutional reform to civil war is due to the lack of data that can provide information on the disbursements made directed for enhancing institutions. Theoretical Framework The arguments made by the studies in the literature review illustrate that there is a relationship between development financing and civil war, with state capacity as a moderating mechanism operating to facilitate it. Across the studies discussed, it emphasized the phenomenon that there is potent likelihood that the variables will have a positive relationship when state capacity is weak. When aid is given under conditions of strong state capacity and military presence, civil war tends to decrease. However, the current literature on the relationship of aid and civil war demonstrates a strong focus on the tangible outcomes of the former on the latter. The focus on humanitarian aid, employment, community development programmes, among others diverts the scholarly gaze from the effects of development financing to intangible variables, such as institutions. The study on development financing that is intended for intangible outcomes leaves much room for analysis. Primary of this is development financing for institutional reform. Yet, there exists a theoretical gap in linking aid with institutional reform. Linking institutions and development To close the gap in the theory, it is imperative to synthesize the intervening nature of development financing and the effects of institutions to development outcomes, peace and violent conflict among them. North (1990) defines institutions as: “the rules of the game in a society or, more formally, are the humanly devised constraints that shape human interaction. In consequence, they structure incentives in human exchange, whether political, social, or economic.” (pg. 3) This definition of institutions will be used in this study. As a set of rules, institutions are intangible, or does not have physical presence, that can affect development outcomes. It provides the motivations and incentives for actors, governments or rebel organisations in this case, to behave in a specific way aligned to their objectives. This may come in the form of advocating for or defending of democratic institutions (rule or expectation of electing leaders, rule of law, participation in decision-making, etc.) which actors in a civil war may use to galvanize support from the people or the international community. The effect of institutions on development have been theorised and studied. Rothstein and Stolle (2003) proposed an institutional theory of states as measured in the trust given by citizens to governments through impartial implementation. Shirley (2005) echoed this and argued that states should establish institutions that are consistent and durable. In effect, they can encourage trust, lower transaction costs, foster market economies, and protect property rights towards development. In the same vein, McFaul’s (1995) studied the weakness of political institutions in Russia to facilitate the enforcement private property rights due to elite capture. The studies discussed provide evidence on the efficacy of strong institutions to influence development by providing the motivations and incentives of individuals and organisations to behave. This, in turn, affects development outcomes, specifically civil war in the case of this study. However, weak institutions are at risk to produce unintended and/or undesired outcomes as studied by McFaul (1995). Yet, institutions can be strengthened through reform. Still, institutional reform cost resources: economic, political, and social. Since development financing intended for institutional reform and strengthening is the objective of Sustainable Development Goal 161 (SDG 16), it can be theorised that SDG 16 financing affects development trajectories of states through funding institutional reform. As stated by the United Nations (2024), SDG 16 aims to “promote peaceful and inclusive societies” by “building effective, accountable and inclusive institutions at all levels”. Since SDG 16 financing prioritizes the democratisation, accountability, and inclusivity of institutions, it is therefore assumed that SDG 16 is financing for institutional reform, an intangible outcome of development financing. The recent release of AidData’s Financing the 2030 Agenda for Sustainable Development (FASD) data makes it possible to analyse the effects of SDG aid, both publicly and to some extent privately funded, to available social data. Disbursements of aid are parsed and categorised according to the projects that helped progress SDG attainment, SDG 16 included. Hence, it can be theoretically and methodologically possible to study the effects of development financing for institutional reform to civil war. Institutions and civil war 1 See Appendix B: Sustainable Development Goal 16 – Common Reporting System Purpose Codes The discussion on the criticality of institutions in influencing civil war has been empirically studied. Fjelde, Knutsen, and Nygård (2021) observed that the complimentary relationship of vertical (electoral) and horizontal (judicial, legislative) institutional constraints on executive power may affect the onset of civil wars. Petrova (2022) found that high levels of trust to local state institutions reduced the risk of conflict exacerbated by flooding in Sub- Saharan Africa. Da Silva (2023) argued that institutions that foster political accountability, especially elections, share a negative relationship with percentage of GDP for military spending; more potent political accountability mechanisms tend to reduce military spending of states. Soysa, Finseraas, and Vadlamannati (2024) claimed that policies promoting private ownership, economic freedom, and the free market reduce the likelihood of the onset of civil war. Reverte (2022) also argued that public policies improving institutional quality, governance systems, education spending, and support for innovation are critical in achieving the SDGs, peace and stability among them. The arguments presented above underpin the theory of this study in which institutions influence civil war. The characteristics of transparent, accountable impartial, and consistent institutions are the same with the goals of institutional reforms outlined by SDG 16. Given the empirically studied influence of institutions with SDG 16 characteristics, development funding disbursed for it can be assumed to have an effect to civil war. SDG 16 development financing and civil war In the previous subsections, it was discussed how institutions are linked to development and how it can influence civil war. Then, a link was established between development financing for institutional reform through SDG 16 funding. Intuitively, it can be theorised that development financing for institutional reform can influence civil war. Evidence on the effect of institutions on civil war point towards the dampening effect of strong and democratic institutions on violent conflict (Fjelde, Knutsen, and Nygård, 2021; Petrova, 2022; Da Silva, 2023). Since SDG 16 fosters institutional reform that are aimed to make institutions more effective and democratic, development financing for this goal would generate similar reducing effects on violent conflict. Notably, studies on development financing have provided significant evidence on moderating effect of state capacity and military capacity on the relationship of aid and violent conflict. At this juncture, it must be clarified that since SDG 16 funds institutional reform, and that institutions are intangible, logic will arrive to the point where SDG 16 funds intangible outcomes of development. Recall that in the literature review, there is a heavy focus on the study of tangible outcomes of development financing. These come in the form of infrastructure, conditional cash transfers, salary for employment programmes, and other material benefits of development financing. This study, in turn, focuses on the intangible aspect of development financing, specifically institutional reform aligned with SDG 16. Hence, this study hypothesizes that funding institutional reforms aligned with SDG 16 reduces deaths caused by violent conflict when state capacity and/or military capacity is potent. It posits that funding institutions to make them more accountable, inclusive, and transparent can affect conditions in which violent conflict can more likely be prevented at the onset, stymied, or ceased. Burnside and Dollar (2000) provide intuitive support for the hypothesis. Their findings reveal that aid is more effective when there are good economic policies (institutions) are in place. To prove the hypothesis, this study proposes a theory that synthesizes the work of Fjelde, Knutsen, and Nygård (2021) on their work with institutions together with the scholarship on the causal relationships of aid and its effects with violent conflict. State capacity and military capacity play a moderating role in the relationship of development financing with violent conflict. The studies discussed in the literature review presents how higher levels of state capacity and military capacity in an area result in reductions in violent conflict when aid is given. Inversely, lower levels of state capacity and military capacity in territory elicits an increase in the instances and/or damages of violent conflict when aid is provided. Thus, establishing a causal relationship between SDG 16 financing and deaths caused by violent conflict must account for the moderating role state capacity plays in the relationship. Hence, this study proposes the following framework: Figure 1: Theoretical framework of the dependent and independent variables with moderators Further interrogations can be made with the given theory when accounting for the permutations of the dependent, independent, and moderator variables. Besides testing the relationship with two moderator variables, hypotheses can be tested when considering just one moderator variable. Hence, this study further posits two hypotheses: 1. There is a negative relationship between SDG 16 development financing and deaths caused by civil war with state capacity as a moderator variable. 2. There is a negative relationship between SDG 16 development financing and deaths caused by civil war with military capacity as a moderator variable. Limitations of the theory Although the causal mechanisms of development financing and institutions with civil war have been empirically studied, there are limitations in the assumptions forwarded by the theory proposed. Chief among these is the role of informal institutions. The measurement provided by SDG 16 development financing is based on the data provided by the Organisation for Economic Co-operation and Development Common Reposting Standard (OECD CRS). The OECD CRS is an instrument used by over 120 countries to agree to share financial flows facilitated by tax agencies, monetary institutions, banks, and other agents of commerce with the common goal to monitor and prevent money laundering, tax evasion and avoidance, among others (Organisation for Economic Co-operation and Development, 2025). The data used for SDG 16 aid disbursements are then only covered by organisations who enact the formal institutions provided by the OECD CRS. Hence, it does not cover organisations who are outside the ambit of these formal institutions. However, under of conditions of low state capacity and conflict, informal institutions are more significant in governing the behaviour of individuals and organisations (Deglow and Fjelde, 2021). The theory hence cannot analyse the influence of informal institutions operating in conflict-affected areas. Rather, the analysis of the theory is only limited to organisations operating under formal institutions that are compliant with the OECD CRS mechanism. These includes governments, international bilateral and multilateral organisations, and civil society organisations who operate under the auspices of the state. Research Data and Design In this section, the author will discuss the variables and statistical techniques that will be used to test the theory proposed in the previous section. It will also define the variables deployed; thus, formulating a model of analysis. The relationship between SDG 16 development financing and deaths caused by civil war with state capacity and military expenditure as a moderators can be studied using a panel regression analysis with moderator variables. It will use 1) deaths caused by natural calamities, 2) democratic governance, 3) ethnolinguistic fractionalization and 4) mountainous terrain as control variables. Variables Independent variable The independent variable, SDG 16 development financing, will be sourced from AidData’s Financing the 2030 Agenda for Sustainable Development (FASD). The study will focus on the data of the disbursed development financing for Sustainable Development Goal 16 – Peace, Justice, and Strong Institutions. Development financing for SDG 16. Burgess, Bengston, and Lautenslager, (2023) defines SDG 16 financing as financial disbursements aimed to improve institutions and organisations, in government or civil society, to 1) prevent or cease civil war, 2) foster reconciliation, 3) build peace, 4) mitigate, reduce, and manage the risks of disasters, and 5) enable connectivity2. This study posits the assumption that aid for SDG 16 to countries are financial resources for policy reform. Hence, aid that is focused on intangible changes in communities. The institutional reforms are aimed to allow governments and/or organisations induce institutional changes that align with SDG 16. The United Nations (2024) outlines specific targets under SDG 16. These include reduction of all forms of violence and deaths, promotion of rule of law at all levels, reduce corruption and bribes, establish effective and accountable institutions, promote participation and inclusion in decision-making processes at all levels, and build capacity to prevent crime and terrorism. These specific targets align with the discussed institutional reforms that have been studied to reduce civil war. However, this study focuses its analysis on the financing for the institutional reforms themselves and not to the tangible outcomes of aid. Thus, it presents a 2 See Annex B for a complete list of keywords. novel measurement of aid where the assumption is that disbursement is allocated to institutions that enable to reduction of deaths caused by civil war. The study will use county-level data from the years 2010 to 2021 lagged at one year in the regression. This will account for the effects of development financing to countries given a year of disbursement of funds. It will be expressed in United States dollars with 2020 fixed prices. This provides the study with an ample amount of time on the implementation of the SDGs. The values are expressed are logged to scale its distribution among the countries. While the SDGs were formalized in 2015, financing that aligns with SDG 16 have started before the ratification of the Goals. Thus, the analysis goes back five years before the ratification of the SDGs. However, a major concern on Aid Data’s FASD is the aggregated nature of its data. Since the database is relatively new, a fully disaggregated dataset is yet to be released. Indeed, FASD has partially disaggregated to parse disbursement of development financing according to year, country, and specific SDG allotment at the project level. However, there is concern on the mutual exclusivity of the variables with respect to their SDG allocation. FASD only provides the amount of development financing for a specific SDG but does not inform whether these amounts overlap with other goals, especially in projects where multiple SDGs are sought to be achieved. The current dataset cannot be verified on this level of granularity and thus risks of mutual exclusivity issues on the variables. Hence, an experimental approach cannot be conducted with the given dataset. It only allows to put one SDG at a time for analysis and scrutiny. Another issue is the risk of conceptual tangency between the indicators used in FASD and the outcomes outlined by the United Nations on SDG 16. FASD categorised SDG 16 development financing using machine learning according to the functional and purposive nature of a project’s budget. On the other hand, the United Nations outlined specific outcomes for SDG 16. This presents a conceptual gap between the indicators. Although it can be assumed that all the functions and purpose of the projects lead towards the attainment of SDG 16, the partially disaggregated nature of FASD presents difficulties in verifying it. Furthermore, since FASD used majority of its keywords for categorising relating to policy reform, it also cannot be verified whether such projects where purely focused on the intangibility of the processes of policy reform or a tangible infrastructure that enables for policy reform to take place. Still, it can be safely assumed that these projects strengthen institutions aligned with SDG 16 characteristics. This suffices the requirement of the theory where development financing for institutional reform towards SDG 16 can affect civil war. Dependent variable The dependent variable for this study will be the number of battle-related deaths from the UCDP Battle-Related Deaths Dataset (UCDP-BRDD). The UCDP BRDD defines battle related deaths as: battlefield fighting, guerrilla activities (e.g. hit-and-run attacks/ambushes) and all kinds of bombardments of military bases, cities and villages etc. The target for the attacks is either the military forces or representatives for the parties, though there is often substantial collateral damage in the form of civilians being killed in the crossfire, indiscriminate bombings, etc. (Petterson, 2024, pg. 5) This study will use country-level data from the years 2010 to 2021. This study will use the sum of the best estimate from UCDP BRRD to capture the number of battle-related deaths in a specific country per year. This variable will express the magnitude of civil war in the analysis. The Battle Related Deaths Database has been extensively used to quantify civil war activity in past studies. Bienek (2024) used the UCDP/PRIO Armed Conflict Dataset to quantify Conflict as a control variable in their study of immigration and development financing. They identified conflict as one of the push factors for people to immigrate to the European Union. In Bienek’s study, they integrated Gleditsch’s et. al. (2002) data in the PRIO dataset to quantify conflict incidents and deaths. Similarly, Ahossey (2024) also used the UCDP BRRD as a control variable in their study of European Union aid effectiveness in Sub-Saharan Africa to the countries’ democracy scores. They used the UCDP BRRD to quantify armed conflict in their study where they found that increased amounts of EU aid, provided with democratic reform conditionalities, increased a country’s democracy score, with engagement with Chinese foreign direct investment as a significant variable in predicting the relationship. Soules and Avdan (2025) deployed logged values of UCDP BRRD as control variable in their novel dataset of magical practices of rebel organizations. The new dataset was tested in prevalence of child recruitment of rebel groups and found strong associations when tested with the control variable. Moderator variables 1. State Capacity The studies discussed in the literature review section highlighted the importance of state capacity that facilitate the relationship of the two main variables to be used in this study. Hendrix (2010) argues that to measure state capacity, focus should be given in a country’s bureaucratic quality and revenue generation. The Quality of Government Database (QOGD, Teorell, et. al., 2024) provides suitable data to measure these variables. In this study, it will use tax revenue as a percentage of a country’s gross domestic product (GDP) that can be sourced from the QOGD. This data can provide the rate of governments relative to their country’s production capacity to collect taxes. It provides balanced data of a state’s revenue generation capacity (RGC) relative to its share of economic output. To measure bureaucratic quality, this paper will base its measurement to the theory of Nistoskaya and Cingolani (2014) where the central bank independence (CBI) is used as proxy. They argue that for central banks to be efficacious with monetary policy, a high degree of political insulation and long term political-economic commitment is required. This minimizes political oversight of the bureaucracy by politicians; thus, less swayed by capricious pollical agenda. The meritorious designation and longevity of central bank governors is an indicator of a state’s capacity to have impartial and consistent governance qualities. The criticality of a central bank in a country’s economic system can represent the impartiality of a state’s civil service, hence a proxy of its bureaucratic quality. This study will express state capacity as a product of the rates of tax revenue as a share of GDP and rate of central bank independence divided by two. Z scores are first standardized and are multiplied with each other to create an interaction effect. The number generated from this formula is the state capacity score. Hence, state capacity is expressed as: 𝑆𝐶 = 𝑅𝐺𝐶 𝑥 𝐶𝐵𝐼 Where: SC is State Capacity RGC is Revenue generation capacity (tax revenue as percentage of GDP) CBI is Central bank independence The RGC and CBI of each country for each year is assumed to represent state capacity. This variable creates an interaction effect between RGC and CBI. It is a two-fold assumption. First, it approximates its ability to generate resources for its own use. A state’s ability to extract financial resources from its economic output. Second, it signifies the consistency of its operations across all government branches. its impartiality of operations represents its status in a year. Hence, it generates a score of a specific country’s government to effectively follow and implement its institutions. The interaction of these two variables denote a government’s efficacy to implement its political objectives. The multiplication of two variables is a statistical practice to create interaction effects. 2. Military expenditure Military expenditure will also be used as a moderator variable in this study. Data on military expenditure as a percentage of GDP from SIPRI as expressed in QOGD will be used for this variable. Citing SIPRI’s adoption of the North Atlantic Treaty Organization definition, this study defines this variable as: all current and capital expenditures on the armed forces, including peacekeeping forces; defense ministries and other government agencies engaged in defense projects; paramilitary forces, if these are judged to be trained and equipped for military operations; and military space activities (pg. 1479) While the literature tells that military capacity plays a significant role in influencing the relationship between aid and civil war, this study reflects on the arguments made by Hendrix (2010) on the reason why military capacity (or expenditure) is not included in the calculation of state capacity. They argued that military expenditure has three pitfalls. First, Hendrix cited Gupta, de Mello and Sharan (2001) where they found that there is a strong and positive association between military spending and corruption. This renders the impartial feature of state capacity endogenous as is negatively affects the political insulation of policy formulation and implementation. The second pitfall is that states may invest more in military capacity if they foresee more challenges against their legitimacy and rule. Hendrix cites the argument made by Mason and Krance (1989) where military capacity can become a tool of repressive political priority instead that can lead to more civil conflict. Lastly, Hendrix cites Sepp (2005) where they emphasized the role of law enforcement as a deterrent against insurgency rather than military strength. Sepp argued that quality of law enforcement and an impartial judiciary are factors that deter insurgency. This resonates with the argument of Hendrix in their measurement of state capacity. Nevertheless, military capacity has been empirically proven to shift the relationship between aid and conflict. Hence, it shall be considered as a moderator variable separate from state capacity. The author expresses military capacity of a state as its military expenditure as a percentage of its GDP. It provides a proxy of a polity’s capacity to prevent and defend against threats to its sovereignty and legitimacy while accounting for its economic output. It will cover the years 2010 to 2021. Control variables 1. Extreme weather conditions Extreme weather conditions are also noted as a significant control variable in the relationship between development financing and deaths caused by civil war. Dasgupta, Gawande, and Kapur (2016) noted this in their study of NREGA in India. Unmet basic needs during and after natural calamities can become triggers of civil war since they can act as motivation to challenge state rule. This study shall measure the human impact directly caused by natural calamities as reflected by the data from the University of Louvain’s Centre for Research on the Epidemiology of Disasters EM-DAT database (2023). Humanitarian aid accounts as development financing hence can affect disbursements to a country if a natural calamity occurs. Natural calamities can also instigate civil war. The EM-DAT uses an inclusion criterion for a natural disaster event when on of the following events has transpired: 1) there are at least 10 deaths, 2) it affected at least 100 people through injury, damages, and homelessness, or 3) the government calls for international assistance or has declared a state of emergency. By accounting the number of deaths, injured people, and affected population, this study can provide a measurement of the magnitude of the natural calamity. This study expresses this data in integers which counts for the number of people died from natural calamities and will use the sum of deaths of a natural calamity that ended on the years 2010 to 2021. 2. Democratic governance Another control variable is the democratic governance. Fjelde, Knutsen, and Nygård (2021) tackled this variable in their study where institutional mechanisms of control over the executive branch of government significantly influence the onset of armed conflict. It mainly argues that when there are potent and democratic institutional controls over the exercise of power, onset of civil conflict becomes unlikely. Their measurement of the degree of democracy a country was categorised into two: vertical constraints and horizontal constraints. Vertical constraints refer to the accountability of government officials through elections. They measured this using data on freedom of expression, suffrage, and clean elections, among others. Horizontal constraints are the mechanisms wielded by co-equal branches of government through the legislative and judiciary branch. It measures the capacity of these branches of government to make the executive branch in check and accountable. While their study separates the variables, for purposes of brevity this study adopts this institutional mechanism as a control variable through the Liberal Democracy Index from QOGD. The Liberal Democracy Index measures the constraints imposed upon the executive power of government to prevent abuse and undue discretion. The data focuses on the degree of protection of individual and minority rights through “constitutionally protected civil liberties, strong rule of law, an independent judiciary, and effective checks and balances” (Teorell, et. al, 2024, pg. 48). This measurement synthesizes the measurement of vertical and horizontal constraints used by Fjelde et. al. to measure democratic governance of states. It accounts for electoral democracy, civil liberties, and the powers of the branches of government over each other. This variable is expressed in positive integers between 1 and 0 where 1 is the most democratic while 0 is the least. The years 2010 to 2021 data will be used from the dataset. 3. Fractionalization Country-level studies on development financing have used Fractionalization as a control variable. De Ree and Nillesen (2009) used Ethnolinguistic and Religious Fractionalization as one of their control variables in their country-level analysis on the effects of aid to civil conflict in Sub-Saharan Africa. Burnside and Dollar (2000) also used ethnolinguistic fractionalization as a variable in their study on the effect aid to economic growth. Fjelde, Knutsen, and Nygård (2021) have also used ethnic fractionalization variable in their cross-country study on the relationship of constraints on executive power and civil conflict. Fractionalization is a measure of a country’s heterogeneity (Alesina, Devleeschauwer, Easterly, Kurlat, and Wacziarg, 2003). They found that fractionalization can influence institutional effectivity, civil war, and economic growth. They argue that higher degrees of heterogeneity within a state is negatively correlated with the quality of government and its economic growth. Fractionalization was distinguished into three types: ethnic, linguistic, and religious fractionalization. People within a country who have varying languages, ethnicity, and religion tend to come with different practices, motivations, and rules that govern them. Hence, development interventions may have varying effects if not considered. Their analysis reveals how the heterogeneity of societies likely results to fragmentation, which can create tensions, competition for resources, and discrimination. These may lead to civil conflict and, as studied, less effective aid towards economic growth. This study will use only ethnic fractionalization as recorded in QOGD to avoid collinearity issues. Fractionalization is expressed in positive integers between 0 and 1, with 1 being the perfectly heterogenous and 0 being perfectly homogenous. The data for years 2010 to 2021 will be used in this study. 4. Country Policy and Institutional Assessment building human resources rating The building human resources rating is the World Bank’s effort to measure the quality of policies that are aimed to reduce poverty through health and education (World Bank, 2025). It tells whether national policies for health and education are in place as implemented by the public and private sectors. Witter et. al. (2015) found plausible empirical evidence on the linkage of health systems in Afghanistan, Timor Leste, and Burundi. Their study reveals that strengthened public administration and development of labour for health systems correlate with state-building processes under fragile conditions and in territories transitioning from violent conflict to peace. This finding merits the inclusion of this variable as a control variable in this study. The values are expressed in the interval of 1 as the lowest and 6 as the highest. Higher scores denote that policies are in place and coherent to provide access to education and health services, especially in the prevention and treatment of diseases. The study will use data from the period 2010 to 2021. 5. Country Policy and Institutional Assessment gender equality rating The gender equality rating is a measure of the World Bank on the extent of policies and programmes a country must enforce and promote gender equality between men and women in education, health, the economy, justice (World Bank, 2025). Ekvall (2013) found strong associations between the presence of policies leaning towards gender equality and the presence of civil war. Using data from Uppsala Conflict Data Project and World Values Survey as the sources of their main variables, Ekvall found that higher political and economic gender equality policies is linked to less instances of armed conflict in a country. Given this, this study adopts the country policy and institutional assessment gender equality rating as a control variable. Like the building human resources rating variable, the values of this variable are expressed in the interval 1 to 6, with 1 as the lowest and 6 as the highest. A low score under this variable signifies that a country lacks the policies and programmes to ensure and promote gender equality each year. For this study, the period 2010 – 2021 will be used. 6. Government spending on health and education This section consists of two variables: education expenditure as a percentage of GDP and health expenditure as a percentage of GDP. Since GDP has been used a control variable in the studies discussed in the literature review, the author decided to include this metric without compromising collinearity with the moderator variable military expenditure as a percentage of GDP. The use of these variables accounts for economic conditions of a country per year through its fiscal rate of expenditure. This is consistent with the moderator variable yet is mutually exclusive among each other. This study will use data from QOGD from the year 2010 to 2021. Preparation of Variables In the UCDP BRDD, data on battle related deaths were trimmed to retain the years 2010 to 2021 in line with the sample duration from FASD. All recipient countries identified by the database received development financing for SDG 163. However, not all had instances of civil war the period. Thus, the data was further trimmed down whether the countries have recorded battle related deaths caused by civil conflict. Thus, the database was left with 50 countries: Afghanistan, Algeria, Angola, Azerbaijan, Bangladesh, Benin, Burkina Faso, Burundi, Cameroon, Central African Republic, Chad, Colombia, Congo, Democratic Republic of the Congo, Egypt, Ethiopia, India, Indonesia, Iran, Iraq, Ivory Coast, Jordan, Kenya, Lebanon, Libya, Malaysia, Mali, Mauritania, Mozambique, Myanmar, Niger, Nigeria, Pakistan, Peru, Philippines, Rwanda, Senegal, Somali, South Sudan, Sudan, Syria, Tajikistan, Tanzania, Thailand, Togo, Tunisia, Turkey, Uganda, Ukraine, and Yemen. The study uses four datasets: UCDP BRRD, FASD, QOGD, and EM-DAT. To facilitate merging of datasets and their variables, the sum of the count values for UCDP BRRD, FASD, and EM-DAT were acquired to generate one observation of a country’s disbursed SDG 16 development financing and battle related deaths per each year. Since the values of the variables from QOGD are interval data per country per year, they were retained to logically align with the analysis of one observation per country per year. The following table shows the summary statistics of the variables to be used in the study: Table 1: Descriptive statistics of variables: Descriptive Statistics Variable Obs Mean Std. Dev. Min Max Battle Related Deaths 529 1829.197 8439.698 0 124546 SDG 16 Financing 529 3.641e+08 5.736e+08 0 4.137e+09 State Capacity 289 .068 .781 -2.356 2.777 Military Expenditure 529 2.399 2.228 0 15.48 Ethnic Fractionalization 529 .686 .527 0 2.786 Liberal Democracy 529 .319 .311 0 2.2 Deaths caused by Natural 529 397.612 1758.345 0 22104 Calamities Building Human Resources 270 3.506 .618 2 4.5 rating Health Expenditure 440 5.117 2.299 1.752 21.828 Education Expenditure 376 3.623 1.397 .396 7.516 3 See Annex E for the list of recipient countries according to FASD. Gender Equality Ra~g 270 3.035 .675 1.5 4.5 It can be seen that three variables exhibit a high level of variability: Battle Related Deaths, SDG 16 development financing, and Deaths Caused by Natural Calamities. SDG 16 Development Financing has high variability due to the scale of the values of the data expressed in millions of US dollars. The values are then logged to fix scaling issues. Hence, the values become: Table 2: Descriptive statistics of SDG 16 Development Financing - unlogged and logged SDG 16 Financing Descriptive Statistics Variable Obs Mean Std. Dev. Min Max SDG16 Financing unlogged 529 3.641e+08 5.736e+08 0 4.137e+09 SDG16 Financing logged 462 18.857 1.455 14.214 22.143 The values for the variable Deaths Caused by Natural Calamities are also logged. As shown in the table below, the variability of the values decreased thus resolving issues of scale. Table 3: Descriptive statistics of unlogged (Deaths Natural Calamities) and logged (Death NaturalCalamities) values of the variable Deaths Caused by Natural Calamities: Descriptive Statistics Variable Obs Mean Std. Dev. Min Max Deaths caused by Natural 529 397.612 1758.345 0 22104 Calamities unlogged Deaths caused by Natural 337 4.346 1.973 0 10.004 Calamities logged As for the dependent variable, the research decided to retain the raw values to avoid biases and methodological problems. The high variability of the values of the dependent variable represents the varying conditions of the territories under study. Some countries exhibit relatively peaceful conditions while others are in a state of war. Ommission bias is also avoided by retaining the outliers. These can yield insights on whether SDG 16 can influence the magnitude of civil war as represented by the death count. To test whether the variables have a normal distribution, a skewness and Kurtosis Test was conducted. The results of the test are as follows: Table 4: Kurtosis test of variables: Skewness and kurtosis tests for normality ----- Joint test ----- Variable Obs Pr(skewness Pr(kurtosis) Adj chi2(2) Prob>chi2 ) Battle Related 529 0.000 0.000 610.870 0.000 Deaths SDG16 462 0.000 0.226 13.110 0.001 Financing State Capacity 289 0.001 0.002 16.770 0.000 Military 529 0.000 0.000 141.180 0.000 Expenditure Ethnic 529 0.000 0.000 51.950 0.000 Fractionalization Liberal 529 0.000 0.000 186.910 0.000 Democracy Score Deaths caused by 337 0.004 0.707 7.750 0.021 Natural Calamities Building Human 270 0.001 0.432 10.680 0.005 Resources Rating Health 440 0.000 0.000 163.240 0.000 Expenditure GDP Education 376 0.077 0.253 4.440 0.109 Expenditure GDP Gender Equality 270 0.002 0.510 9.480 0.009 Rating The results of the skewness and Kurtosis test reveal that the dependent variable battle- related deaths, the independent variable lagged SDG 16 development financing, the moderator variables state capacity and military expenditure as a percentage of GDP, and the control variables liberal democracy, ethnic fractionalization, deaths caused by natural calamities, and health expenditure as a percentage of GDP have p-values of 0.0000. This means that the variables have significant number of outliers and does not follow a normal distribution. Meanwhile, Expenditure on Education as a percentage of GDP has a p-value of 0.1052, indicating a closeness to normal distribution. Since the focus of this study is Battle Related Deaths (BRD), the author chose to focus on the outliers of the distribution of values in this variable. It reveals high values in the countries Syria, Afghanistan, and Iraq throughout the period of 2010-2021. These countries with high values of BRD are in state of war thus registering high values. On the other hand, the countries of Benin, Congo, Ivory Coast, Jordan, Malaysia, Mauritania, Peru, Senegal, and Tanzania showed none or just 1 observation of Battle-Related Deaths in a year. This signifies that these countries are generally peaceful, thus having minimal or zero BRD. Furthermore, the variance of the dependent variable has greater value than the mean as shown in the summary statistics below: Table 5: Summary statistics of Battle Related Deaths Descriptive Statistics Variables Obs Mean Std. Min Max p1 p99 Skew. Kurt. Dev. Battle Related 609 1939.793 10247.60 0 163217 0 43650 10.8 143.407 Deaths 9 Statistical Technique Given the distribution of the variables and the focus on the dependent variable, this study will use a zero-inflated negative binomial regression (ZINBR) with clustered countries with moderator and control variables to analyse the relationship between development financing for SDG-16 and battle related deaths caused by civil war at the country level from 2010 to 2021. The countries are clustered since the zero-inflated negative binomial regression can only be done with cross section data. By clustering the countries, Stata assumes that within country observations are correlated. This method is selected because of the 1) skewness and Kurtosis issues of the data, 2) significant number of outliers, and 3) the variance of the dependent variable is greater than the mean, and 4) there are a lot of zero values in the variables. The observations where there were zero values of BRD were retained to avoid biases in eliminating them. Clustering with zero-inflated negative binomial regression to analyse panel data has been used in studies. Eric, Melanie, Yuan, Jiajia, and Bankole (2020), pooled their observations according to years in their study of the association between employment and self-reported mental health in the USA for 2011-2017. Nødtvet, Dohoo, Sanchez, Conboy, DesCôteaux, Keefe, Leslie, and Campbell (2002) also pooled their age group variable in their study of gastrointestinal parasites among Canadian dairy cows. These studies provide precedence on the method selected for this study where clustering according to country is valid when analysing panel data using zero-inflated negative binomial regression. An analysis of the dispersion parameter will be done to determine whether the statistical technique is appropriate. The dispersion parameter, represented by the alpha score, indicates whether there is overdispersion in the model. Values closer to zero indicate lower levels of dispersion in the data. This will function as a robustness test for the statistical models deployed for this study since it provides evidence whether the usage of a negative binomial regression is justified. Results and Discussion Base model testing: To test the baseline relationship of this study’s dependent and independent variable, a basic model of the was run. The dependent, independent, and control variables were deployed in the ZINBR. Table 6: Results of zero-inflated negative binomial regression on basic model Zero-inflated negative binomial regression Battle Related Coef. St.Err. t-value p-value [95% Conf Interval] Sig Deaths SDG16 Financing .332 .16 2.08 .037 .02 .645 ** Ethnic .625 .291 2.15 .032 .055 1.195 ** Fractionalization Liberal Democracy -1.294 .523 -2.47 .013 -2.32 -.269 ** Deaths Caused by .198 .071 2.78 .005 .058 .337 *** Natural Calamities Building Human .531 .48 1.11 .269 -.409 1.471 Resources Health .156 .065 2.39 .017 .028 .284 ** Expenditure Education -.053 .159 -0.33 .74 -.365 .259 Expenditure Gender Equality -.222 .278 -0.80 .424 -.767 .322 Rating Constant -2.477 2.829 -0.88 .381 -8.023 3.068 Battle Related -1.483 .036 -41.59 0 -1.552 -1.413 *** Deaths SDG16_Financing -2.225 .49 -4.54 0 -3.186 -1.264 *** Ethnic -6.994 1.325 -5.28 0 -9.59 -4.397 *** Fractionalization Liberal Democracy 13.216 3.46 3.82 0 6.434 19.999 *** Score Constant 64.371 9.184 7.01 0 46.37 82.372 *** lnalpha .015 .102 0.14 .885 -.185 .214 Mean dependent var 1226.170 SD dependent var 3072.646 Number of obs 135 Chi-square 150.543 Prob > chi2 0.000 Akaike crit. (AIC) 1372.706 *** p<.01, ** p<.05, * p<.1 Salient values of the regression reveal that SDG 16 development financing has a positive and significant association with battle related deaths having a coefficient of .332 and with a p score of 0.037. This means that higher SDG 16 development financing is associated with higher numbers of battle related deaths in the basic model. This may seem counterintuitive given the discussion on the Theory Chapter of this study. This may mean that endogeneity issues with SDG 16 development financing elicits an increase in battle related deaths. Increasing SDG 16 disbursement can be a reaction to past conflict which may be on an upward trend as civil war intensifies without accounting for state capacity and military expenditure. Hence, SDG 16 may be influenced by battle related deaths instead. However, it should be noted that the moderators, which have been empirically proven, are not operating in this model. This accentuates the role of state capacity and military expenditure as moderator variables when analyzing the effect of SDG 16 development financing to battle related deaths. Ethnic Fractionalization has a significant positive association with battle-related deaths having a coefficient of .62 and a p score of 0.032. This supports previous research made that argued that ethnic heterogeneity in a country increases the risk of civil war. Competing and contrary interests divided along ethnic factional interests over limited resources may greatly contribute to the risk of civil war. The variable, deaths caused by natural calamities, has significantly positive association with battle related deaths with a p = 0.005 and a coefficient of .19. This shows that natural calamities are strongly associated with the increase of battle related deaths in the basic model. It echoes the theory of Dasgupta, Gawande, and Kapur (2016) in their study of NREGA’s effect to civil war in India after natural calamities. Health expenditure also has a positive significant association with battle related deaths. The basic model yielded p = 0.017 with a coefficient of 0.15. This may be due to higher health spending under situations of armed conflict. Inversely, a higher liberal democracy score yielded a statistically negative relationship with battle related deaths. With a p score of 0.013 and a coefficient of -1.29, the basic model provides support to the claim that countries with higher levels of liberal democracy are less likely to have battle related deaths. This resonates with the findings of Fjelde et. al. (2021) where higher levels of control over the executive branch of government is linked with lower incidents of violent conflict. In terms of predicting zero instances of battle related deaths, the regression reveals that SDG 16 development financing has a significantly negative association with battle related deaths. The basic model yielded a coefficient of -1.48 with a p value of p < 0.001. This demonstrates that under situations of zero battle related deaths, SDG 16 drives the retention of this condition. Hence, SDG 16 is significantly associated with avoiding battle related deaths. Interestingly, Ethnic Fractionalization and Liberal Democracy also has a statistically significant association in avoiding battle related deaths. The basic model yielded p < 0.001 with a coefficient of -6.99 for Ethnic Fractionalization and p < 0.001 with a coefficient of 13.21 for Liberal Democracy. The basic model demonstrates that peaceful coexistence among ethnic groups with strong liberal democratic governance prevents battle related deaths. This runs counter with previously cited studies where ethnic heterogeneity contributes to the risk of violent conflict. However, under conditions of high levels of liberal democratic governance, battle related deaths may be prevented. Given the result of the regression of the baseline model, the hypothesis forwarded is rejected as SDG 16 development financing displays significant effect to increase battle related deaths. This may be due to endogeneity issues since the moderator variables are not used in the model. Yet, under conditions where there are zero battle related deaths, SDG 16 is seen to have a potent preventive influence against civil war. Furthermore, the baseline model ZINB regression illustrates strong evidence that deaths caused by natural calamities share a positive association with battle related deaths. Increased deaths caused by natural calamities is strongly associated with leading to more battle related deaths. Hypothesis: SDG 16 development financing has a negative association with battle related deaths with state capacity as a moderator variable In this section, ZINBR will be conducted with State Capacity as the moderating variable. An interaction effect variable is also included in the model to measure the likelihood of SDG 16 development financing and State Capacity to affect civil war. The results are as follows: Table 7: Hypothesis testing for SDG development financing and battle related deaths with state capacity as moderator in ZINBR Zero-inflated negative binomial regression Battle Related Coef. St.Err. t-value p-value [95% Conf Interval] Sig Deaths SDG16 Financing .118 .142 0.83 .409 -.162 .397 State Capacity 7.568 1.927 3.93 0 3.792 11.345 *** Interaction of -.425 .097 -4.37 0 -.616 -.234 *** State Capacity and SDG16 Financing Ethnic .646 .289 2.24 .025 .08 1.213 ** Fractionalization Liberal Democracy -1.261 .638 -1.98 .048 -2.511 -.01 ** Deaths caused by .013 .073 0.17 .861 -.131 .156 Natural Calamities Building Human .525 .531 0.99 .323 -.516 1.565 Resources Rating Health .081 .066 1.24 .215 -.047 .21 Expenditure Education .766 .127 6.04 0 .518 1.014 *** Expenditure Gender Equality .208 .271 0.77 .442 -.323 .738 Rating Constant -1.371 2.233 -0.61 .539 -5.748 3.005 Battle Related -1.913 .041 -46.91 0 -1.993 -1.833 *** Deaths SDG16_Financing -6.244 1.369 -4.56 0 -8.927 -3.561 *** Ethnic -21.794 4.298 -5.07 0 -30.218 -13.37 *** Fractionalization Liberal Democracy 50.777 12.575 4.04 0 26.13 75.425 *** score Constant 147.289 25.983 5.67 0 96.363 198.215 *** lnalpha -.669 .19 -3.51 0 -1.042 -.295 *** Mean dependent var 1732.700 SD dependent var 3861.200 Number of obs 80 Chi-square 544.041 Prob > chi2 0.000 Akaike crit. (AIC) 842.454 *** p<.01, ** p<.05, * p<.1 Salient results of the ZINBR with State Capacity as moderator show that the independent variable, SDG 16 development financing, did not yield statistically significant results. It yielded a p value of 0.409, hence does not explain the variation in battle related deaths- However, State Capacity exhibits significant positive association with battle related deaths of a country the following year. The p value of < 0.001 with a coefficient of 7.568 illustrate that a 1 unit increase in State Capacity would likely result to a 7.768 unit increase in battle related deaths. This can be due to higher state capacity may indicate autocratic rule that can motivate civil war. This is supported when considering Liberal Democracy Score where it yielded a p-value of 0.048 with a coefficient of -1.26. This suggests that higher levels of democratic governance are associated with fewer battle related deaths. Interestingly, when testing for interaction effect of SDG 16 development financing and state capacity it resulted to a negative association with p score of <0.001 and a coefficient of - 0.425. This forwards an interesting insight where a higher level of state capacity, when receiving funding for SDG 16 is associated with lower battle related deaths. Funding reforms aligned with SDG 16 goals is associated with the reduction of battle related deaths when there is strong state capacity. Consistently, the Ethnic Fractionalization variable has aa statistically significant positive association with battle related deaths in the model. The resulting p value of 0.025 with a coefficient of 0.646 provides evidence of this association. Unexpectedly, Education Expenditure yielded a statistically significant positive association with battle related deaths. Further analysis is required to identify whether there are issues of endogeneity or collinearity with the variable. In the zero-inflated model, SDG 16 financing yielded a statistically significant positive relationship. This echoes the analysis made in the baseline model where conflict-prone areas receive more SDG 16 financing. Higher level of Ethnic Fractionalization is seen as reducing the chance to prevent battle related deaths. This supports the claims made by Fjelde et. al. (2021) where higher levels of ethnic heterogeneity make it more likely to increase violent conflict. As with the previous model, higher levels of liberal democratic governance is associated with preventing battle related deaths. Given that the p value for the independent variable yielded statistically insignificant result, the hypothesis that battle related deaths has a negative association to SDG 16 development financing with state capacity as moderator is rejected. Furthermore, SDG 16 financing is linked with reducing the likelihood of zero battle related deaths. State capacity is seen as having a positive association with battle related deaths in the model. However, when interaction effects are accounted between the independent variable and the moderator variable together with battle related deaths, the result become statistically significant with a negative association. It can be assumed that in the ZINBR model with state capacity as moderator, SDG 16 development financing alone is ineffective in predicting the values of battle related deaths. However, when interacting with State Capacity, it provides evidence that financing SDG 16 reform for states with high levels of state capacity reduces the likelihood of battle related deaths. Consistently, liberal democracy is observed to have a negative association with battle related deaths. Hypothesis: SDG 16 development financing has a negative relationship with battle related deaths with military expenditure as a percentage of GDP as a moderator variable Results of the ZINBR for the model are as follows: Table 8: Hypothesis testing for SDG development financing and battle related deaths with military expenditure as moderator Zero-inflated negative binomial regression Battle Related Coef. St.Err. t-value p-value [95% Conf Interval] Sig Deaths SDG16 Financing .231 .181 1.28 .202 -.124 .586 Military .625 1.086 0.58 .565 -1.503 2.753 Expenditure Interaction of -.049 .057 -0.86 .39 -.162 .063 Military Expenditure and SDG 16 financing Ethnic .321 .26 1.24 .216 -.188 .831 Fractionalization Liberal Democracy 1.489 .905 1.64 .1 -.285 3.263 Score Deaths caused by .222 .074 3.01 .003 .078 .367 *** Natural Calamities Building Human .498 .482 1.03 .301 -.447 1.444 Resources Rating Health .077 .06 1.27 .203 -.041 .195 Expenditure Education .136 .185 0.73 .463 -.227 .499 Expenditure Gender Equality -.807 .338 -2.39 .017 -1.469 -.145 ** Rating Constant .983 3.48 0.28 .778 -5.839 7.804 Battle Related -1.573 .036 -43.65 0 -1.643 -1.502 *** Deaths SDG16 Financing -2.29 .505 -4.53 0 -3.279 -1.3 *** Ethnic -7.355 1.355 -5.43 0 -10.01 -4.699 *** Fractionalization Liberal Democracy 13.37 3.519 3.80 0 6.473 20.267 *** Score Constant 67.157 9.476 7.09 0 48.585 85.729 *** lnalpha -.107 .123 -0.87 .386 -.348 .134 Mean dependent var 1226.170 SD dependent var 3072.646 Number of obs 135 Chi-square 152.427 Prob > chi2 0.000 Akaike crit. (AIC) 1363.718 *** p<.01, ** p<.05, * p<.1 Variables that elicited statistically significant p values are Deaths caused by Natural Calamities and Gender Equality Rating. Higher values of battle related deaths may be due to the state instability caused by natural calamities. The shock of human, economic, and political losses can induce fragility to countries making it more likely to fall into civil war. Inversely, higher levels of gender equality are associated with reducing battle related deaths. This is consistent with the finding of Ekvall (2013) where gender equal policies and laws were observed to reduce the likelihood of triggering violent conflict. However, in this model, the introduction of the moderator variable military expenditure and its interaction with the independent variable exhibited statistically insignificant results, having a p value of 0.565 and 0.390 respectively. This tells that higher military expenditure cannot explain the variances in the values of battle related deaths. This runs counter to the arguments made by the literature where military capacity has shown statistically significant negative relationship with violent conflict. The zero-inflated model saw SDG 16 financing as having a strong association with reducing the likelihood of zero battle related deaths. This is consistent with the previous analysis where SDG 16 financing tends to be disbursed in conflict prone areas. Similar with the previous model, higher levels of Liberal Democracy are shown to have a potent preventing effect against battle related deaths. Given the results of the regression, the hypothesis whether SDG 16 development financing has a negative association with battle related deaths is rejected. Hypothesis: SDG 16 development financing has a negative relationship with battle related deaths with state capacity and military expenditure as a percentage of GDP as moderator variables The result for the regression of the model is as follows: Table 9: Hypothesis testing for SDG development financing and battle related deaths with state capacity military expenditure as moderators Zero-inflated negative binomial regression Battle Related Coef. St.Err. t-value p-value [95% Conf Interval] Sig Deaths SDG16 Financing -.03 .147 -0.21 .837 -.318 .257 State Capacity 6.372 1.728 3.69 0 2.985 9.76 *** Military -.507 .858 -0.59 .555 -2.188 1.174 Expenditure Interaction of -.36 .09 -4.02 0 -.535 -.184 *** State Capacity and SDG16 Financing Interaction of .01 .043 0.24 .812 -.075 .096 Military Expenditure and SDG16 Financing Ethnic .261 .307 0.85 .397 -.342 .863 Fractionalization Liberal Democracy 1.21 1.127 1.07 .283 -.999 3.42 Score Deaths caused by .047 .079 0.60 .552 -.108 .202 Natural Calamities Building Human .412 .49 0.84 .401 -.549 1.372 Resources Rating Health .031 .071 0.43 .665 -.108 .17 Expenditure Education .836 .149 5.62 0 .544 1.127 *** Expenditure Gender Equality -.337 .297 -1.14 .256 -.918 .245 Rating Constant 3.528 2.543 1.39 .165 -1.455 8.512 Battle Related -1.589 .04 -39.26 0 -1.668 -1.509 *** Deaths SDG16 Financing -6.447 1.593 -4.05 0 -9.569 -3.326 *** Ethnic -23.576 5.064 -4.66 0 -33.501 -13.651 *** Fractionalization Liberal Democracy 46.5 13.434 3.46 .001 20.17 72.831 *** Score Constant 148.745 30.487 4.88 0 88.992 208.497 *** lnalpha -.806 .189 -4.27 0 -1.175 -.436 *** Mean dependent var 1732.700 SD dependent var 3861.200 Number of obs 80 Chi-square 533.817 Prob > chi2 0.000 Akaike crit. (AIC) 838.442 *** p<.01, ** p<.05, * p<.1 The result of the ZINBR show that the independent variable does not have a significant association to predict the values of battle related deaths the following year. The p value of 0.837 indicates non-statistically significant association in predicting the values of battle related deaths. This may be due to the insufficiency of lag time of the independent variable. In the literature review, a one-year lag for development financing was used to study the effect of tangible outcomes of development financing. With SDG 16 development financing, however, intangible outcomes such as institutional changes may take more time to have effect to battle related deaths. Currently, theoretical precedents to create this lagging effect are insufficient. However, the moderator variable of state capacity statistically significant result. The p value for state capacity is < 0.000 with a positive coefficient of 6.37. This shows that a one unit increase in state capacity is associated with an increase of battle related deaths of 6.37 units. This is similar with the model where state capacity is the only moderator variable. Yet, the influence of state capacity over the dependent variable is inverted when SDG 16 financing interacts with it. The regression resulted to a p value of < 0.000 with a coefficient of -0.36. This illustrates the strong association SDG 16 financing and state capacity over battle related deaths. This can be due to SDG 16 financing when directed to countries with higher levels of state capacity, create the necessary conditions for the instances and magnitude of civil war to decrease. However, the analysis of further lagging the SDG 16 variable must be further studied. Counterintuitively, the moderator variable military expenditure as a percentage of GDP has yielded statistically insignificant results both as a stand alone variable and its interaction with the independent variable. The regression yielded a p value of 0.555 and 0.812 respectively. Again, this runs contrary to previous finding where military capacity and presence have been observed to diminish the magnitude and incidents of civil war. This may be due to the issue of insufficient lagging of SDG 16 to account for its effects. Since development financing literature on tangible outcomes of aid are lagged for just one-year, institutional reform may require increased duration for lagging to account for the relatively protracted effects of institutional reform to civil war. This reflects the statistically insignificant results yielded by the SDG 16 financing variable. Robustness test The dispersion parameter of the four models are as follows: Table 10: Zero values dispersion of models Model Number of Non-zero Zero values Logged Alpha Observations values observations of alpha score observations of the dependent score the dependent variable variable Baseline model 135 84 (62.2% of 51 (37.8% of -0.107 0.899 observations) observations) Model with 80 51 (63.75% of 29 (36.25% of -0.806 0.447 State Capacity observations) observations) as moderator Model with 135 84 (62.2% of 51 (37.8% of -1.155 0.315 Military observations) observations) Expenditure as moderator Model with 80 51 (63.75% of 29 (36.25% of -0.559 0.572 State Capacity observations) observations) and Military Expenditure as moderators Across the four models, it can be observed that there are high proportion of zero observations in the dependent variable. More than 30% of the observations have zero values. This indicates the high number of observations of absence of battle related deaths in a country in a year. Alpha scores in all models are greater than zero. This indicates overdispersion in the dataset thus justifying the use of zero-inflated negative binomial regression in all four models. Conclusion This study is an effort to test whether SDG 16 development has a relationship with battle related deaths of countries per year. Studies on the relationship of development financing and civil war show that there has been a focus on the tangible outcome of aid. The lack of scholarship on the relationship of civil war with the intangible outcomes of development financing is a lacuna in the literature that this study seeks to fill. It theorised that SDG 16 development financing, given SDG 16’s focus on institutional reform, can account for the intangible outcomes of aid. It is analysed with battle related deaths as a proxy for the magnitude of civil war. Based on the literature, state capacity and military capacity act as salient moderators in influencing the relationship between aid and civil war. A zero-inflated negative binomial regression was deployed to analyse the relationship. The technique was chosen because the dependent variable, battle related deaths, has skewness and kurtosis issues due to the ample outliers in the dataset. Furthermore, the observations for battle related deaths contains zero values in the panel data. Liberal democracy scores, ethnic fractionalization, deaths caused by natural calamities, building human resource rating, gender equality rating, education expenditure as a percentage of GDP, and health expenditure as a percentage of GDP were chosen as control variables. SDG 16 development financing was lagged for one year to give time for it to take effect on the dependent variable. The results show that SDG 16 development financing, on its own, is not associated with the number of battle related deaths. This may be due to the long duration for institutional reform to take effect on civil war. Unlike tangible outcomes, intangible outcomes such as institutional reform takes time to be implemented and take on civil war. Thus, further study should be made to determine the optimal lag for use in analysing SDG 16 development financing and its effect to civil war. Furthermore, confounding effects should be studied to fully distinguish and parse the FASD database. The partially disaggregated database cannot provide for nuanced analysis of SDG 16 financing. Disaggregation at the specific types and functionalities of institutional reform should be made to effectively analyse its effects with other social phenomena. The moderator variables exhibited mixed results. Higher levels of state capacity show a positive relationship with battle related deaths. This may be due to higher levels of repression that motivate civil violence. When interacted with SDG 16 financing, the relationship is inversed. This suggests that countries with high levels of state capacity are more likely to reduce civil war when SDG 16 financing is disbursed. Thus, policy reforms should be directed towards creating more inclusive institutions that respect the rule of law and human rights. When states have higher levels of state capacity and are receive funding to have more inclusive institutions that respect human rights and rule of law, deaths caused by civil war tend to decrease. Notably, the more complex models yielded analysis that may provide endogeneity issues of SDG 16 when lagged for just one year. In the zero-inflated logit model, SDG 16 financing tends to be disbursed to countries with higher levels of civil conflict. To understand whether SDG 16 financing affects battle related deaths, further study should be made to address endogeneity issues of the variable. One such proposal is further disaggregation of the FASD dataset to parse specific institutional reform line items. On the other hand, Military capacity, as proxied by military expenditure as a percentage of GDP, consistently yielded statistically insignificant results with the dependent variable. The study also reveals that higher levels of liberal democratic governance yielded statistically significant negative relationship with battle related deaths. This suggests that further democratisation of states, aligned in liberal democratic values reduce conflict. Guarantees on free speech and religion, effective peaceful means for redress and conflict resolution, and participatory governance are some of the policy directions this insight entails in terms of reform. In conclusion, the study established an association with SDG 16 development financing and battle related deaths under conditions of strong state capacity. Liberal democratic governance and gender equality consistently exhibited likelihood to reduce and prevent civil war. Nevertheless, all hypotheses proposed in this study is rejected given the current model used. Further disaggregation of FASD, address endogeneity of SDG 16 financing with military expenditure, and expanding the model with other confounding variables is key to further provide a more robust analysis of the relationship of the two main variables. Bibliography Acemoglu, Daron, Simon Johnson, and James A. Robinson (2005). Chapter 6 Institutions as a Fundamental Cause of Long-Run Growth in Philippe Aghion and Steven N. Darauf (eds.) Handbook of Economic Growth Volume 1, Part A. Elsevier. 385-472 Adams, Greg (2015). Honing the proper edge: CERP and the two-sided potential of military- led development in Afghanistan. Economics of Peace and Security Journal 10(2). 53-60. Ahossey, Kossi (2024). Do Colonial Legacy and Engagement with China Impact the Effectiveness of the European Aid in Promoting Democracy in African States? International Area Studies Review 27(4). 341-352. Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlate, and Romain Wacziarg (2003). Fractionalization. Journal of Economic Growth 8. 155-194. Banerjee, Abhijit, Ether Dulfo, Nathaniel Goldberg, Dean Karlan, Robert Ose, William Parienté, Jeremy Shapiro, Bram Thuysbaert and Christopher Udry (2015). A multifaceted program causes lasting progress for the very poor: Evidence from six countries. Science, 348(6236). Beath, Andrew, Christia Fotini, and Ruben Enikopolov (2012). Winning Hearts and Minds Through Development: Evidence from a Field Experiment in Afghanistan. World Bank Policy Research Working Papers. Accessed at: https://elibrary.worldbank.org/doi/epdf/10.1596/1813-9450-6129 Berman, Eli, Joseph H. Felter, Jacob N. Shapiro, Erin Troland (2013). Modest, Secure, and Informed: Successful Development in Conflict Zones. American Economic Review 103(3). 512-517. Berman, Eli, Jacob N. Shapiro, and Joseph H. Felter (2011). Can Hearts and Minds Be Bought? The Economics of Counterinsurgency in Iraq. Journal of Political Economy 119(4). 617-820. Bienek, Jan (2024). A mixed-methods study on the effect of development aid from the European Union and its member states on asylum seeker movements (unpublished master thesis). Ghent University. Burgess, Bryan, Andrew Bengtson,, and Brook Lautenslager (2023). Financing the 2030 Agenda for Sustainable Development, Version 1.0. Williamsburg, VA. AidData. Accessed at http://aiddata.org/sdg Burgess, Bryan, John Custer, and Samantha Custer (2023). Financing Agenda 2030: Are international aid donors missing the mark on the Sustainable Development Goals? Williamsburg, VA: AidData at William & Mary. Burnside, Craig and David Dollar (2000). Aid, Policies, and Growth. The American Economic Review 90(4). 847-868. Boräng, Frida, Marcia Grime, and Helene Ahlborg (2022). Introducing Uncertainty: Community Driven Development and Local Collective Action Capacity. Quality of Government Working Paper Series 2022:8. Accessed at https://gupea.ub.gu.se/bitstream/handle/2077/74586/2022_8_Borang_Grimes_Ahlborg. pdf?sequence=3&isAllowed=y Centre for Public Impact (2016). Building trust in government: Afghanistan's National Solidarity Programme (NSP). Accessed at: https://www.centreforpublicimpact.org/case- study/building-trust-in-government-afghanistans-national-solidarity-program Centre for Research on the Epidemiology of Disasters (2023). EM-DAT. https://public.emdat.be/data Child, Travers B. (2014). Hearts and Minds Cannot be Bought: Ineffective Reconstruction in Afghanistan. The Economics of Peace and Security Journal 9(2). 43-49. Chou, Tiffany (2012). Does development assistance reduce violence? Evidence from Afghanistan. Economics of Peace and Security Journal, 7(2), 5–13. Coppedge, Michael, John Gerring, Carl Henrik Knutsen, Staffan I. Lindberg, Jan Teorell, David Altman, Fabio Angiolillo, Michael Bernhard, Cecilia Borella, Agnes Cornell, M. Steven Fish, Linnea Fox, Lisa Gastaldi, Haakon Gjerløw, Adam Glynn, Ana Good God, Sandra Grahn, Allen Hicken, Katrin Kinzelbach, Kyle L. Marquardt, Kelly McMann, Valeriya Mechkova, Anja Neundorf, Pamela Paxton, Daniel Pemstein, Oskar Rydén, Johannes von Römer, Brigitte Seim, Rachel Sigman, Svend-Erik Skaaning, Jeffrey Staton, Aksel Sundström, Eitan Tzelgov, Luca Uberti, Yi-ting Wang, Tore Wig, and Daniel Ziblatt. 2024. "V-Dem Codebook v14" Varieties of Democracy (V-Dem) Project. Crost, Benjamin, Joseph Felter, and Patrick Johnston (2014). "Aid under Fire: Development Projects and Civil Conflict." American Economic Review, 104 (6): 1833-1856 Crost, Benjamin, Joseph H. Felter, and Patrick B. Johnston (2016). Conditional cash transfers, civil conflict and insurgent influence: Experimental evidence from the Philippines. Journal of Development Economics 118. 171-182. Custer, Samantha, Axel Dreher, Thai-Binh Elston, Andreas Fuchs, Siddharta Ghose, Joyce Jiahui Lin, Ammar A. Malik, Bradley C. Parks, Brooke Russell, Kyra Solomon, Austin Strange, Michael J. Tierney, Katherine Walsh, Lincoln Zaleski, and Sheng Zhang (2021). Tracking Chinese Development Finance: An Application of AidData’s TUFF 2.0 Methodology. Williamsburg, VA: AidData at William & Mary. Da Silva, Diego Lopes (2023). Political Accountability and Military Spending. Defence and Peace Economic 34(5). 563-580. DOI: 10.1080/10242694.2022.2129128 Dasgupta, Aditya, Kishore Gawande, and Devesh Kapur (2016). (When) Do Antipoverty Programs Reduce Violence? India's Rural Employment Guarantee and Maoist Conflict. International Organization 71. 605-632 Davies, Shawn, Garoun Engström, Therese Pettersson & Magnus Öberg (2024). Organized violence 1989-2023, and the prevalence of organized crime groups. Journal of Peace Research 61(4). De Ree, Joppe and Eleonora Nillesen (2009). Aiding violence or peace? The impact of foreign aid on the risk of civil conflict in sub-Saharan Africa. Journal of Development Economics 88(2). 301-313. De Soysa, Indra, Henning Finseraas and Krishna Vadlamannati. Group Grievances, Opportunity, and the Onset of Civil War: Some Theory and Tests of Competing Mechanisms, 1990-2017. Peace Economic Peace Science Publication 30(2). 171-205.. Deglow, Annekatrin and Hanne Fjelde (2021). ‘The Quality of Government and Civil Conflict’ in Andreas Bågenholm, Monika Bauhr, Marcia Grimes, and Bo Rothstein (eds.) The Oxford Handbook of the Quality of Government. Oxford University Press. Oxford. 781-799 Dreher, Axel, Andreas Fuchs, Bradley Parks, Austin Strange, and Michael J. Tierney (2021). Aid, China, and Growth: Evidence from a New Global Development Finance Dataset. American Economic Journal: Economic Policy, 13(2), 135-174. Ekvall, Åsa (2013). Gender Equality, Attitudes to Gender Equality, and Conflict. Gendered Perspectives on Conflict and Violence: Part A Advances in Gender Research 18A. 272- 295. Eric, Chinaeke, Gwynn Melanie, Hong Yuan, Zhang Jiajia, and Olatosi Bankole (2020). The positive association between employment and self-reported mental health in the USA: a robust application of marginalized zero-inflated negative binomial regression (MZINB). Journal of Public Health 42(2). 340-352. Fetzer, Thiemo (2014) Can workfare programs moderate violence? Evidence from India. Economic Organisation and Public Policy Discussion Papers (EOPP 53). Suntory and Toyota International Centres for Economics and Related Disciplines, London, UK. Fjelde, Hanne and Carl Henrik Knutsen (2021). Which Institutions Matter? Re-Considering the Democratic Civil Peace. International Studies Quarterly 65. 223-237 Fiszbein, Ariel, Norbert Schady, and Fran Ferreira (2009). Conditional cash transfers: reducing present and future poverty. World Bank. Accessed at: https://documents1.worldbank.org/curated/en/914561468314712643/pdf/476030PUB0 Cond101Official0Use0Only1.pdf Gleditsch, Nils Petter, Peter Wallensteen, Mikael Eriksson, Margareta Sollenberg, and Håvard Strand (2002) Armed Conflict 1946-2001: A New Dataset. Journal of Peace Research 39(5) Gupta, Sanjeev, Luiz de Mello and Raju Sharan (2001). Corruption and Military Spending. European Journal of Political Economy 17(4). 749-777 Grossman, Herscell I. (1991). A General Equilibrium Model of Insurrections. The American Economic Review 81(4). 912-921. Hendrix, Cullen S. (2010). Measuring state capacity: Theoretical and empirical implications for the study of civil conflict. Journal of Peace Research, 47(3). 273-285. https://doi.org/10.1177/0022343310361838 Högbladh Stina, 2023, “UCDP GED Codebook version 23.1”, Department of Peace and Conflict Research, Uppsala University Jabbari, Mehdi, Majid Shafiepour Motlagh, Khosro Ashrafi, and Ghareman Abdoli (2020). Differentiating countries based on the sustainable development proximities using the SDG indicators. Environment, Development, and Sustainability 22. 6405-6423. Khana, Gaurav and Laura Zimmerman (2017). Guns and butter? Fighting violence with the promise of development. Journal of Development Economics 124. 120-141. Kotsadam, Andreas, Gudrun Østby, Siri Aas Rustad, Andreas Forø Tollefsen, and Henrik Urdal (2018). Development aid and infant mortality. Micro-level evidence from Nigeria. World Development, 105, 59-69. Mahatma Gandhi Ministry of Rural Development (2012). Mahatma Gandhi National Rural Employment Guarantee Act (Mahatma Gandhi NREGA). Accessed at: https://rural.gov.in/sites/default/files/02%20_CIC_PartII_MG_NREGA%28F%29_0.pd f Mason, T David and Dale A. Krane (1989). The political economy of death squads: Toward a theory of the impact of state-sanctioned terror. International Studies Quarterly 33(2). 175-198 McFaul, Michael (1995). State Power, Institutional Change, and the Politics of Privatization in Russia. World Politics 47(2). 210-243. Myxter-iino, Erik and Juliet Lu (Host) (2018-2024). Ammar Malik, China AidData, and the Data Debate over Chinese Lending [Audio Podcast]. Accessed at: https://open.spotify.com/episode/6ref7zgcxHPynmeIRi6osO?si=47af55fadd214e2a Narang, Neil (2015). Assisting Uncertainty: How Humanitarian Aid can Inadvertedly Prolong Civil War. International Studies Quarterly 59(1). 184-195. Narang, Neil (2014). Humanitarian Assistance and Duration of Peace after Civil War. The Journal of Politics 76(2). 446-460 North, Douglas (1990). Institutions, Institutional Change and Economic Performance. Cambridge University Press. Published online on June 2012. https://doi.org/10.1017/CBO9780511808678 Nunn, Nathan and Nancy Qian (2014). US Food Aid and Civil Conflict. American Economic Review 104(6). 1630-1366 Nødtvet Ane, Ian Dohoo, Javier Sanchez, Gary Conboy, Luc DesCôteaux, Greg Keefe, Ken Leslie, and John Campbell (2002). The use of negative binomial modelling in a longitudinal study of gastrointestinal parasite burdens in Canadian dairy cows. Can J Vet. Res. 66(4). 249-257. Official Gazette of the Philippines (2024). Pantawid Pamilyang Pilipino Program. Accessed at: https://www.officialgazette.gov.ph/programs/conditional-cash-transfer/ Organisation for Economic Co-operation and Development (2024). AEOI standard's implementation status by jurisdiction. Accessed at: https://www.oecd.org/tax/automatic- exchange/crs-implementation-and-assistance/crs-by-jurisdiction/#d.en.345489 Organisation for Economic Co-operation and Development (2024). Development finance standards - understand how we measure and collect data. Accessed at: https://www.oecd.org/development/understanding-development-finance.htm Organisation for Economic Co-operation and Development (2025). CRS Information. Accessed at: https://web-archive.oecd.org/tax/automatic-exchange/international- framework-for-the-crs/ Petrova, Kristina (2022). Floods, communal conflict and the role of local state institutions in Sub-Saharan Africa. Political Geography 92. 1-11. Petterson, Therese (2024). UCDP Battle-related Deaths Dataset Codebook v 24,1 (https://ucdp.uu.se/downloads/). Reverte, Carmelo (2022). The importance of institutional differences among countries in SDGs achievement: A cross-country empirical study. Sustainable Development. 1882- 1899. doi: 10.1002/sd.2354 Rothstein, Bo and Dietlind Stolle (2003). Social Capital, Impartiality and the Welfare State: An Institutional Approach in Marc Hooghe and Dietlind Stolle (eds.) Generating Social Capital. Palgrave Macmillan. New York. 191-209. Sachs, J.D., Guillaume Lafortune, Grayson Fuller, and Eamon Drumm (2023). Implementing the SDG Stimulus. Sustainable Development Report 2023. Paris: SDSN, Dublin: Dublin University Press, 2023. 10.25546/102924 Sepp, Kalev I (2005). Best practices in counterinsurgency. Military Review 85. 8-12. Sexton, Renard (2016). Aid as a Tool against Insurgency: Evidence from Contested and Controlled Territory in Afghanistan. American Political Science Review 110(4). 731- 749. doi:10.1017/S0003055416000356 Shirley, Mary M. (2005). Institutions and Development in Claude Menard and Mary M. Shirley (eds) Handbook of New Institutional Economics. Springer. New York. 611-638. Sobek, David (2010). Managing of their domains: The role of state ccapacity in civil wars. Journal of Peace Reserach 47(3). 267-356. Special Inspector General for Afghanistan Reconstruction (2018). Commander’s Emergency Response Program: DOD Has Not Determined the Full Extent to Which Its Program and Projects, Totaling $1.5 Billion in Obligations, Achieved Their Objectives and Goals in Afghanistan from Fiscal Years 2009 through 2013. Accessed at https://www.sigar.mil/pdf/audits/SIGAR-18-45-AR.pdf Soules, Michael and Nazli Avdan. The magical practices of rebel organizations: Introducing the Magical Acts by Groups in Civil Conflicts dataset. Journal of Peace Research, 0(0). https://doi.org/10.1177/00223433241295466 Sundberg, Ralph, and Erik Melander, 2013, “Introducing the UCDP Georeferenced Event Dataset”, Journal of Peace Research, vol.50, no.4, 523-532 Teorell, Jan, Aksel Sundström, Sören Holmberg, Bo Rothstein, Natalia Alvarado Pachon, Cem Mert Dalli, Rafael Lopez Valverde & Paula Nilsson. 2024. The Quality of Government Standard Dataset, version Jan24. University of Gothenburg: The Quality of Government Institute, https://www.gu.se/en/quality-government doi:10.18157/qogstdjan24 United Nations (2015). Historic New Sustainable Development Agenda Unanimously Adopted by 193 UN Members. Access at: https://www.un.org/sustainabledevelopment/blog/2015/09/historic-new-sustainable- development-agenda-unanimously-adopted-by-193-un-members/ United States Agency for International Development (2010). Community Stabilization Program Final Report. Accessed at: https://pdf.usaid.gov/pdf_docs/PDACR642.pdf United Nations (2024). Goal 16: Promote just, peaceful and inclusive societies. Accessed at: https://www.un.org/sustainabledevelopment/peace-justice/ Wade, Lila (2017). Can You Pay for Peace? The Role of Financing Frameworks in Effective Peace Operations. Journal of International Peacekeeping 21. 152-175 Witter, Sophie, Jean-Benoit Falisse, Maria Paola Bertone, Alvaro Alonso-Garbayo, João S. Martins, Ahmad Shah Salehi, Enrico Pavigani, and Tim Marrtineau (2015). State- building and human resources for health in fragile and conflict-affected states: exploring the linkages. Human Resources for Health 13(33). https://doi.org/10.1186/s12960-015- 0023-5 Wood, Reed M. and Christopher Sullivan (2015). Doing Harm by Doing Good? The Negative Externalities of Humanitarian Aid Provision during Civil Conflict. The Journal of Politics 77(3). https://doi.org/0022-3816/2015/7703-0011$10.00 Wood, Reed M. and Emily Molfino (2016). Aiding Victims, Abetting Violence: The Influence of Humanitarian Aid on Violence Patterns During Civil Conflict. Journal of Global Security Studies 1(3). 186-203 Woodridge, Michael (2015). From MDGs to SDGs: What are the Sustainable Development Goals? International Council for Local Environmental Initiatives Briefing Sheet - Urban Issues, No. 1. https://www.local2030.org/library/251/From-MDGs-to-SDGs-What-are- the-Sustainable-Development-Goals.pdf Weintrub, Michael (2016). Do All Good Things Go Together? Development Assistance and Insurgent Violence in Civil War. The Journal of Politics 78(4). 957-1295 World Bank (2025). Metadata Glossary. Databank website of the World Bank Group. https://databank.worldbank.org/metadataglossary/country-policy-and-institutional- assessment/series/IQ.CPA.GNDR.XQ World Bank (2025). Metadata Glossary. Databank website of the World Bank Group. https://databank.worldbank.org/metadataglossary/country-policy-and-institutional- assessment/series/IQ.CPA.HRES.XQ Zizumbo-Colunga, Daniel (2023). The Unintended Consequences of Conditional Cash Transfer Programs for Violence: Experimental and Survey Evidence from Mexico and the Americas. Latin American Politics and Society 65(3): 47-71. doi:10.1017/lap.2022.67 Appendices Appendix A: Sustainable Development Goal 16 – Common Reporting System Purpose Codes 1) education policy and administrative management; 2) primary education; 3) secondary education; 4) higher education; 5) health policy and administrative management; 6) population policy and administrative management; 7) water resources policy and administrative management; 8) government and civil society, 9) purpose unspecified or does not fit under any other applicable codes; 10) economic and development policy/planning; 11) public sector financial management; 12) legal and judicial development; 13) government administration; 14) strengthening civil society; 15) conflict prevention and resolution, 16) peace and security, 17) purpose unspecified or does not fit under any other applicable codes; 18) security system management and reform; 19) civilian peace-building, conflict prevention and resolution; 20) post-conflict peace-building (un); 21) reintegration and small arms and light weapons (SALW) control; 22) land mine clearance; 23) child soldiers (prevention and demobilisation); 24) social/ welfare services; 25) employment policy and administrative management; 26) housing policy and administrative management; 27) transport policy and administrative management 28) communications policy and administrative management; 29) information and communication technology (ict); 30) energy policy and administrative management; 31) electrical transmission/ distribution; 32) financial policy and administrative management; 33) monetary institutions; business support services and institutions; 34) agricultural policy and administrative management; 35) forestry policy and administrative management; 36) fishing policy and administrative management; 37) industrial policy and administrative management; 38) mineral/mining policy and administrative management; 39) construction policy and administrative management; 40) trade policy and administrative management; 41) trade facilitation; tourism policy and administrative management; 42) environmental policy and administrative management; 43) women in development; 44) urban development and management; 45) rural development; 46) disaster prevention and preparedness Appendix B: List of donor countries 1. Austria 2. Belgium 3. Denmark 4. France 5. Germany 6. Italy 7. Netherlands 8. Norway 9. Portugal 10. Sweden 11. Switzerland 12. United Kingdom 13. Finland 14. Iceland 15. Ireland 16. Luxembourg 17. Greece 18. Spain 19. Slovenia 20. Czech Republic 21. Slovak Republic 22. Hungary 23. Poland 24. Lithuania 25. Canada 26. United States 27. Japan 28. Korea 29. Australia 30. New Zealand 31. Monaco 32. Cyprus 33. Malta 34. Turkey 35. Croatia 36. Liechtenstein 37. Bulgaria 38. Romania 39. Estonia 40. Latvia 41. Russia 42. Algeria 43. Libya 44. Mexico 45. Iraq 46. Israel 47. Kuwait 48. Qatar 49. Saudi Arabia 50. United Arab Emirates 51. Azerbaijan 52. Kazakhstan 53. Chinese Taipei 54. Thailand 55. Timor-Leste 56. Canada Appendix C: List of multilateral organisations donors 1. EU Institutions 2. Nordic Development Fund 3. United Nations Environmental Program 4. Global Environmental Facility 5. Montreal Protocol 6. International Bank for Reconstruction and Development 7. Multilateral Investment Guarantee Agency 8. International Finance Corporation 9. International Development Association 10. Caribbean Development Bank 11. International Monetary Fund 12. Inter-American Development Bank 13. Central American Bank for Economic Integration 14. African Development Bank 15. African Development Fund 16. Asian Development Bank 17. Arab Fund (AFESD) 18. United Nations Peacebuilding Fund 19. Council of Europe 20. World Health Organisation 21. Food and Agriculture Organisation 22. International Atomic Energy Agency 23. United Nations Economic Commission for Europe 24. OPEC Fund for International Development 25. Organization of Arab Petroleum Exporting Countries 26. Arab Bank for Economic Development in Africa 27. Special Arab Aid Fund for Africa 28. IMF Trust Fund 29. International Monetary Fund (Concessional Trust Funds) 30. United Nations Development Program 31. United Nations Transitional Authority 32. United Nations Conference on Trade and Development 33. United Nations Children’s Fund 34. United Nations Relief and Works Agency 35. United Nations High Commissioner for Refugees 36. United Nations Programme on HIV/AIDS 37. United Nations Population Fund 38. Organization for Security and Co-operation in Europe 39. Islamic Monetary Fund 40. Arab Fund for Technical Assistance to African and Arab Countries 41. Black Sea Trade & Development Bank 42. GODE 43. Other Arab Agencies 44. International Fund for Agricultural Development 45. European Bank for Reconstruction and Development 46. Global Partnership for Education 47. Climate Investment Funds 48. Adaptation Fund 49. Council of Europe Development Bank 50. Private Infrastructure Development Group 51. Development Bank of Latin America 52. Green Climate Fund 53. Credit Guarantee and Investment Facility 54. Global Energy Efficiency and Renewable Energy Fund 55. IDB Invest 56. Central Emergency Response Fund 57. World Tourism Organisation 58. Asian Infrastructure and Investment Bank 59. Center of Excellence in Finance 60. International Investment Bank 61. United Nations Institute for Disarmament Research 62. United National Capital Development Fund 63. Eurasian Fund for Stabilization and Development 64. New Development Bank 65. North American Development Bank 66. United Nations Women 67. COVID-19 Response and Recovery Multi-Partner Trust Fund 68. Joint Sustainable Development Goals Fund 69. International Commission on Missing Persons 70. World Health Organization – Strategic Preparedness and Response Plan 71. International Centre for Genetic Engineering and Biotechnology 72. Global Alliance for Vaccines and Immunization 73. Global Fund 74. Global Green Growth Institute 75. World Trade Organization – International Trade Centre 76. United Nations Industrial Development Organization Appendix D: List of Private Donors in SDG Financing Database List of private donors 1. Bill & Melinda Gates Foundation 2. Dutch Postcode Lottery 3. Swedish Postcode Lottery 4. People's Postcode Lottery 5. MetLife Foundation 6. Mastercard Foundation 7. Grameen Crédit Agricole Foundation 8. IKEA Foundation 9. Bernard van Leer Foundation 10. MAVA Foundation 11. Oak Foundation Fondation Oak 12. H&M Foundation 13. Laudes Foundation 14. Charity Projects Ltd (Comic Relief) 15. Children's Investment Fund Foundation 16. Gatsby Charitable Foundation 17. Conrad N. Hilton Foundation 18. David & Lucile Packard Foundation 19. John D. & Catherine T. MacArthur Foundation 20. Carnegie Corporation of New York 21. Michael & Susan Dell Foundation 22. Omidyar Network Fund, Inc. 23. Rockefeller Foundation 24. William & Flora Hewlett Foundation 25. Arcus Foundation 26. Gordon and Betty Moore Foundation 27. Ford Foundation 28. Wellcome Trust 29. UBS Optimus Foundation 30. World Diabetes Foundation 31. McKnight Foundation 32. Citi Foundation 33. LEGO Foundation 34. Norwegian Postcode Lottery 35. BBVA Microfinance Foundation 36. Jacobs Foundation 37. Arcadia Fund 38. Margaret A. Cargill Foundation 39. La Caixa Banking Foundation 40. Bloomberg Family Foundation 41. Susan T. Buffett Foundation 42. Howard G. Buffett Foundation 43. Open Society Foundations 44. Fondation Botnar 45. CHANEL Foundation 46. Bezos Earth Fund 47. German Postcode Lottery Appendix E: List of recipient countries in the study 1. Turkey 2. Kosovo 3. Serbia 4. Bosnia and Herzegovina 5. Montenegro 6. North Macedonia¨ 7. Albania 8. Ukraine 9. Belarus 10. States Ex-Yugoslavia unspecified 11. Europe, regional 12. Moldova 13. Algeria 14. Libya 15. Morocco 16. Tunisia 17. Egypt 18. North of Sahara, regional 19. South Africa 20. Angola 21. Botswana 22. Burundi 23. Cameroon 24. Cabo Verde 25. Central African Republic 26. Chad 27. Comoros 28. Congo 29. Democratic Republic of the Congo 30. Benin 31. Ethiopia 32. Gabon 33. Gambia 34. Ghana 35. Guinea 36. Guinea-Bissau 37. Equatorial Guinea 38. Côte d'Ivoire 39. Kenya 40. 2Lesotho 41. Liberia 42. Madagascar 43. Malawi 44. Mali 45. Mauritania 46. Mauritius 47. Mozambique 48. Niger 49. Nigeria 50. Zimbabwe 51. Rwanda 52. Sao Tome and Principe 53. Senegal 54. Eritrea 55. Sierra Leone 56. Somalia 57. Djibouti 58. Namibia 59. Saint Helena 60. Sudan 61. South Sudan 62. Eswatini 63. Tanzania 64. Togo 65. Uganda 66. Burkina Faso 67. Zambia 68. South of Sahara, regional 69. Africa, 70. 3Costa Rica 71. Cuba 72. Dominican Republic 73. El Salvador 74. Guatemala 75. Haiti 76. Honduras 77. Belize 78. Jamaica 79. Mexico 80. Nicaragua 81. Panama 82. Dominica 83. Grenada 84. Saint Lucia 85. Saint Vincent and the Grenadines 86. Montserrat 87. Caribbean & Central America,regional 88. Argentina 89. Bolivia 90. Brazil 91. Colombia 92. Ecuador 93. Guyana 94. Paraguay 95. Peru 96. Suriname 97. Venezuela 98. South America, regional 99. America, regional 100. Iran 101. Iraq 102. Jordan 103. West Bank and Gaza Strip 104. Lebanon 105. Syrian Arab Republic 106. Yemen 107. Middle East, regional 108. Armenia 109. Azerbaijan 110. Georgia 111. Kazakhstan 112. Kyrgyzstan 113. Tajikistan 114. Turkmenistan 115. Uzbekistan 116. Central Asia, regional 117. Afghanistan 118. Bhutan 119. Myanmar 120. Sri Lanka 121. India 122. Maldives 123. Nepal 124. Pakistan 125. Bangladesh 126. South Asia, regional 127. South & Central Asia, regional 128. Cambodia 129. China (People's Republic of) 130. Indonesia 131. Democratic People's Republic of Korea 132. Lao People's Democratic Republic 133. Malaysia 134. Mongolia 135. Philippines 136. Thailand 137. Timor-Leste 138. Viet Nam 139. Far East Asia, regional 140. Asia, regional 141. Fiji 142. Kiribati 143. Nauru 144. Vanuatu 145. Niue 146. Marshall Islands 147. Micronesia 148. Palau 149. Papua New Guinea 150. Solomon Islands 151. Tokelau 152. Tonga Tonga 153. Tuvalu Tuvalu 154. Wallis and Futuna 155. Samoa 156. Oceania, regional 157. Developing countries, unspecified 158. Eastern Africa, regional 159. Middle Africa, regional 160. Southern Africa, regional 161. Western Africa, regional 162. Caribbean, regional 163. Central America, regional 164. Melanesia, regional 165. Micronesia, regional Appendix F: Stata Do-File for Zero-Inflated Negative Binomial Regression Hypothesis testing A. Regression for ZINBR baseline model zinb BattleRelatedDeaths SDG16_Financing Ethnic_Fractionalization Liberal_Democracy_Score Deaths_Natural_Calamities Building_Human_Resources_Rating Health_Expediture_GDP Education_Expenditure_GDP Gender_Equality_Rating, inflate( BattleRelatedDeaths SDG16_Financing Ethnic_Fractionalization Liberal_Democracy_Score) vce(robust) B. Regression for ZINBR with State Capacity as moderator zinb BattleRelatedDeaths SDG16_Financing State_Capacity State_Capacity_SDG16 Ethnic_Fractionalization Liberal_Democracy_Score Deaths_Natural_Calamities Building_Human_Resources_Rating Health_Expediture_GDP Education_Expenditure_GDP Gender_Equality_Rating, inflate( BattleRelatedDeaths SDG16_Financing Ethnic_Fractionalization Liberal_Democracy_Score) vce(robust) C. Regression for ZINBR with Military Expenditure as moderator zinb BattleRelatedDeaths SDG16_Financing Military_Expenditure_GDP Military_Expenditure_SDG16 Ethnic_Fractionalization Liberal_Democracy_Score Deaths_Natural_Calamities Building_Human_Resources_Rating Health_Expediture_GDP Education_Expenditure_GDP Gender_Equality_Rating, inflate( BattleRelatedDeaths SDG16_Financing Ethnic_Fractionalization Liberal_Democracy_Score) vce(robust) D. Regression for ZINBR with State Capacity and Military Expenditure as moderators zinb BattleRelatedDeaths SDG16_Financing State_Capacity Military_Expenditure_GDP State_Capacity_SDG16 Military_Expenditure_SDG16 Ethnic_Fractionalization Liberal_Democracy_Score Deaths_Natural_Calamities Building_Human_Resources_Rating Health_Expediture_GDP Education_Expenditure_GDP Gender_Equality_Rating, inflate( BattleRelatedDeaths SDG16_Financing Ethnic_Fractionalization Liberal_Democracy_Score) vce(robust)