ECONOMIC STUDIES 
DEPARTMENT OF ECONOMICS 
SCHOOL OF BUSINESS, ECONOMICS AND LAW 
UNIVERSITY OF GOTHENBURG 
232 
________________________ 
   Incentives and Forest Reform: 
      Evidence from China 
Yuanyuan Yi
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ISBN 978-91-88199-19-5 (printed) 
 ISBN 978-91-88199-20-1 (pdf) 
 ISSN 1651-4289 (printed) 
 ISSN 1651-4297 (online) 
 
Printed in Sweden, 
Gothenburg University 2017 
In loving memory of my father
To my mother and my daughter, Yingzhuo
Acknowledgements
First and foremost I wish to thank my supervisors Gunnar Ko¨hlin and Fredrik Carlsson at
the University of Gothenburg, and my advisor Jintao Xu at Peking University. They have
supported me not only by their expertise, inspiration, and enthusiasm in environmental and
natural resource economics, but also their guidance, advice and encouragement in many aspects
throughout the writing of this thesis.
I gratefully acknowledge the generous financial and capacity building support from the
Swedish International Development Cooperation Agency (Sida) at the Environmental Eco-
nomics Unit at the University of Gothenburg. I would also like to express my enormous
gratitude to the Environment for Development Initiative (EfD) and the Richard C. Malmstens
Foundation for their financial support.
Cyndi Berck is highly appreciated for her expertise in editing and publication manage-
ment. Her advice and comments on the entire thesis and her help in clarifying my language
have been indispensable. I want to especially thank Yonas Alem and Jessica Coria for their
thorough reviews and constructive comments. I would like to express my deep appreciation to
H˚akan Eggert, Erik Hjalmarsson, Randi Hjalmarsson, Martin Holmen, Mitesh Kataria, Elina
Lampi, A˚sa Lo¨fgren, Ola Olsson, Johan Stennek, Thomas Sterner, and Ma˚ns So¨derbom at the
University of Gothenburg. I would also like to extend my sincere gratitude to Peter Berck at
UC Berkeley, Allen Blackman at the Inter-American Development Bank, Martin Dufwenberg
at the University of Arizona, Pamela Jagger at the University of North Carolina at Chapel
Hill, Billy Pizer at Duke University, Elizabeth Robinson at University of Reading, Andy White
at the Rights and Resources Group, Runsheng Yin at Michigen State University, and William
F. Hyde.
Heartfelt thanks to Wolfgang Habla. He has been not only a brilliant co-author, but has
shared with me his generosity and admirable research attitude.
I warmly acknowledge all my classmates and friends for their deep understanding of the
papers in my thesis and their very useful comments in my seminars. I equally appreciate my
colleagues at the department for their encouragement and moral support.
A special thanks to my family. They are the most precious fortune to me and have blessed
me with a life of joy in research.
Yuanyuan Yi
Gothenburg, May 2017
Contents
Acknowledgments i
1 Introduction 1
Collective forests and household management . . . . . . . . . . . . . . . . . . . . . . . 2
State-owned forests and state forest enterprieses . . . . . . . . . . . . . . . . . . . . . 4
Policy implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Forest Devolution Reform in China: A Trigger for Investment or Deforestation?
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Evidence of devolution and forest conditions . . . . . . . . . . . . . . . . . . . . . . . . 3
The Chinese Collective Forest Tenure Reform . . . . . . . . . . . . . . . . . . . . . . . 5
Data and descriptive statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Empirical strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Figures and tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3 Allocative Efficiency or Agglomeration? The Emergence of Forestland Rental Mar-
kets and the Forest Devolution Reform in China
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Studies on agricultural land rental markets and efficiency . . . . . . . . . . . . . . . . 3
China’s forest devolution reform and emerging forestland rental markets . . . . . . . . 5
A conceptual model of forestland rental market participation . . . . . . . . . . . . . . 7
Testable hypotheses and empirical strategy . . . . . . . . . . . . . . . . . . . . . . . . 8
Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Figures and tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4 Managerial Incentives for Environmental Protection in Chinese-Style Federalism
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Theoretical framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Empirical strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Econometric results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Figures and tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1 Introduction
In the past two decades, the world’s forests had a net loss of 129 million hectares,
though the rate of forest loss is declining (FAO, 2015). An increasing number of devel-
oping countries have recognized local ownership and control over forests. Meanwhile,
governments adopted forest tenure reforms by devolving control of forests to (indige-
nous) community management, or by a further step, to household management. And
some governments maintain their forests as public. By 2015, 76 percent of the world’s
total forests of 3,999 million hectares were under public ownership and less than 20
percent were under private ownership. The existence of such reforms, legislation or
regulations, however, is not always coupled with effective enforcement and forest con-
servation (FAO, 2015; RRI, 2014). Meanwhile, governments in general, in spite of
concerns with environmental protection and climate change mitigation, are more fo-
cused on economic development, food security and political stability. The overriding
concerns of sustainability, efficiency and distribution in forest devolution reforms and
forest management systems have never been higher.
China, as one of the world’s recent “economic development miracles”, has a total
forestland area of over 208 million hectares as of 2014, and is ranked fifth in terms
of percentage of the world’s total (FAO, 2015). China’s forests are categorized into
two ownership types: village collective-owned and state-owned. The collective forests
consist of 60 percent of the national forest area and 32 percent of the forest stock
volume; the state forests account for 40 percent of the national forest area and 68
percent of the forest stock (State Forestry Administration, 2011).
In this context, the thesis explores the extent and depth of recognition of rights to
forest land and resources in the two types of forest in China, and investigates the rela-
tionship between managerial incentives and associated environmental outcomes. The
first two chapters focus on collective forests and evaluate the impacts of the devolution
of forest tenure rights from villages to households. The third chapter shows how the
institutional setting for state-owned forests, in which the actual forest manager is an
agent to two principals – the central government and jurisdictional (local) government –
leads to deforestation. The findings show that devolution of forest rights to households
improves households’ investment and the efficiency of forestland reallocation among
households, as well as household welfare and forest biomass. With respect to state-
owned forests, however, the research shows that forests are being depleted because local
governments strongly incentivize the forest manager to maximize revenue, while the
1
central government provides only limited incentives to protect the environment. These
findings contribute substantially to the knowledge base for the forest policy sector.
2 Collective Forests and Household Management
Whether devolution (decentralization) brings about deforestation is a question that is
ceaselessly asked. An increasing number of countries have devolved forests to the com-
munity level in forest-rich areas of Asia, Africa, and Latin America, where more than
27 countries had the total forest area under local community management increased
from 383 million hectares to 511 million hectares in 2002-2013 (RRI, 2014). A large
body of literature has been devoted to studying the effect of this step of devolution on
forest conditions, but the findings are inconclusive (e.g., Kaimowitz et al., 1998; Foster
et al., 2002; Baland et al., 2010; Coleman and Fleischman, 2012). A key question, then,
is whether to take a further step, by devolving forest management to the household
level. China made an early move.
Given the history of unsuccessful village collective management, a wave of forest de-
volution reform was announced soon after the millennium, starting in Fujian province,
and featuring the devolution of forest management rights to households. The main
measures include reallocation of village collectively-owned forestland, formal acknowl-
edgement of household property rights to these forest plots, in particular by forest-
land certificates with legalized tenure terms, and encouragement of forestland transfer
(rental) markets. This was very successful and encouraged the central government
to implement a few more pilots starting in 2003 (State Council, 2003) and finally to
promulgate the policy document “Collective Forest Tenure Reform in the Southern
Collective Forest Areas in China” in 2008. This promulgation reflects both the recog-
nition by the benefits of devolution by the highest legislation and the extent and the
depth of the government’s ambition to devolve forest tenure rights. During that pe-
riod, any decision and implementation of the reform required village-level democratic
consensus.
The Collective Forest Tenure Reform in China is the most extensive devolution of
communal forests to households ever seen. By 2008, the reform had devolved 62 million
hectares of the total 100 million hectares of the forests from collective ownership to
individual households (Xu et al., 2010). The reform has involved 600 million people
in rural areas in more than twenty provinces of China (State Forestry Administration,
2
2011).
My Chapter I, Forest Devolution Reform in China: A Trigger for Investment
or Deforestation? (single authored), evaluates the impact of the Collective Forest
Tenure Reform on households’ investment in forestland. I also investigate the effect
of household management on forest resource conditions. Specifically, the investment
analysis is based on a panel dataset of a two-round survey of 3,000 households in
eight provinces before and after the implementation of the forest devolution reform.
The identification strategy exploits the variations in villages’ decisions to select the
reform and in households’ forest investment across time. Using a difference-in-difference
propensity score matching model, I find that the devolution reform resulted in more
investment per area unit of a forest plot, in terms of annual labor input days and value
of silvicultural treatments. The analysis on resource conditions is based on satellite
imagery on forest cover and vegetation during 2001-2012. At the county level, where
more forestland is under household management, improved forest conditions are found
soon after the reform. As the channels for the investment effect, I investigate the
following two: (i) the effect of tenure security, i.e., holding a forestland certificate
and (ii) the reallocation effect from obtaining more forestland during reform. The
effects of devolution and improved tenure on increased private investment and resource
conditions provide evidence that well-defined and protected property (tenure) rights
for households offer an effective alternative to common-pool resources management in
small-scale forestry in China.
In Chapter II, Allocative Efficiency or Agglomeration? The Emergence of Forest-
land Rental Markets and the Forest Devolution Reform in China (single authored),
I focus on the emerging forestland rental markets. I investigate whether the devo-
lution reform of forestland to household management had an effect on allocative ef-
ficiency and household welfare through households’ participation in forestland rental
markets. Using a household panel dataset from three Chinese provinces, I find that the
emerging forestland rental markets improved allocative efficiency, using an indicator of
factor equalization. Based on multinomial logit model estimates for households that
chose among renting out their forestland, renting in forestland, and not participat-
ing in the rental market, I find that, with the reform, forestland is transferred (rented)
from forestland-rich, labor-constrained households to forestland-constrained, labor-rich
households. I also find that forestland is transferred to households with higher levels of
productivity in forestry. I do not find any evidence for agglomeration of forestland to
3
households that were already land-rich, nor to wealthier or politically powerful house-
holds. Furthermore, I compare the differences in welfare between the no-renters and
renting households of similar characteristics, based on a propensity score matching ap-
proach. I find participation in forestland rental markets increases household per-capita
income and decreases the likelihood of having an income below the poverty line.
3 State-owned Forests and State Forest Enterprises
The state-owned forest sector presents a different system from that in the villages’ col-
lective forests. In the state forest areas, forests are owned by the state (i.e., the central
government on behalf of the state). These forests are managed by state-owned forest
enterprises (SFEs). A manager of an SFE has obligations to the central government
on sustainable use of forests. Meanwhile, he or she also signs a contract with the ju-
risdictional sub-national government on revenue sharing and other societal goals such
as job creation, payment of pension benefits, provision of schooling and health care.
In the past thirty years, China’s annual growth rate of gross domestic product
(GDP) was remarkable, at almost 10 percent. However, the fast speed of economic
growth has come at a tremendous cost to the environment, with inefficient, excessive
resource use and high levles of pollution (Liu and Diamond, 2008). In this process,
state-owned enterprises inextricably make an enormous contribution, given that 80
percent of the total national value of gross industrial output came from state-owned
enterprises by the end of the 1970s and they accounted for 70 percent of the total
national assets (NBSC, 1999). During the current decade, state-owned enterprises still
account for 40 percent of the total national assets (NBSC, 2015).
The fast rate of GDP growth has been attributed to the so-called Chinese-style
federalism (Montinola et al., 1995; Xu, 2011). A key feature of this type of federalism
is that it combines fiscal decentralization with performance-based personnel control.
The decentralization of fiscal authority, combined with a fiscal transfer system, allows
regional governments (provincial, municipal, county and township level governments)
to have primary control over economic issues, including firms in their jurisdiction,
while the central government typically owns natural resources and possesses the right
to set pollution targets. In this system, short-term economic growth is rewarded with
promotions of the eligible people at each level of the political hierarchy; by contrast,
longer-term environmental issues such as resource degradation and pollution do not
4
negatively affect the likelihood of being promoted.
In Chapter III, Managerial Incentives for Environmental Protection in Chinese-
Style Federalism (with Wolfgang Habla and Jintao Xu), we explain the strategic con-
sequences of a manager of an SFE in this context, as an agent who faces two principals
– the jurisdictional sub-national government that is in charge of the state-owned en-
terprise, and the State Forestry Administration on behalf of the central government,
which is the owner of the forests that the agent manages. Career concerns by managers
of state-owned enterprises that manage natural resources, and asymmetric information
between managers and their superiors regarding the enterprises’ environmental perfor-
mance, are sources of environmental degradation. As well as needing to meet ecological
targets imposed by the national government, a manager wants to be promoted into the
ranks of the sub-national government. We develop three hypotheses based on a the-
oretical model with two principals and one agent. We then empirically test these
hypotheses for the case of China’s northeastern state-owned forests, combining satel-
lite imagery data on deforestation with economic survey data. Our findings suggest
that managers of state-forest enterprises that have a larger area are more difficult to
monitor with respect to ecological targets, log more timber, and are more likely to
deforest. The same holds true for managers who share a larger percentage of profits
with the local government. We find that the latter increases the likelihood that the
managers will get promoted.
4 Policy Implications
The results of this thesis have important policy implications along at least three di-
mensions. First, the two papers on the Chinese second-order devolution reinforce the
importance of focusing on the quality of resource management reforms. The quality of
reforms depends on deepening the recognition of tenure rights, as well as the in-depth
knowledge of a great number of factors that determine individual managerial incentives,
so that managers of devolved forest will not “fell and run”. The effects of devolution
and improved tenure on increased private investment and short-run forest conditions
enhance our understanding that the effectiveness of devolution relies on well-defined
and protected property rights for households. Emerging forestland rental markets are
facilitating forestland allocation among households. The households that need more
forestland, and that manage forests efficiently, rent in more land. A long-run, compre-
5
hensive effect on sustainable forest management will benefit from policy efforts that
intensify tenure rights – in particular, transferability rights to more households – and
that strengthen the protection of tenure rights and households’ perception of tenure
security.
Second and more specifically to China, although the results in Chapter II alleviate
concerns about agglomeration of forestland to wealthy or powerful households or those
already are large landholders, follow-up policies should still be vigilant on fighting con-
solidation to households of these kinds in the long run. Yet, agglomeration to more
efficient users is not bad. In particular, rapid urbanization and transformation from
rural to urban economic activities have demanded more rural-to-urban labor migration.
More and easier factor mobilization is required. In this case, strengthening local insti-
tutions to improve farmers’ access to factor resources and credit is necessary to address
these considerable constraints. In parallel, policies to improve rural households’ knowl-
edge and skills in forest management, and to help build transparent, well-functioning
forestland exchange platforms/markets that are easily accessible by villagers, could
reduce transaction costs and encourage forestland rental market development.
Finally, the findings about the lack of managerial incentives to protect the environ-
ment call for systematic change. The scale of China’s forests already makes it difficult
for national or regional environmental authorities to measure and monitor forest de-
pletion; an additional challenge comes from the fact that forest managers and local
governments have a mutual interest in maximizing revenue from forests. The latter
worsens the problem of deforestation even more, because of local governments’ “protec-
tive umbrella” effect. Policy suggestions are directed toward making the measurement
of forest activities easier – for example, by using monitoring technologies such as real-
time satellite imagery. More systematic solutions include transferring authority over
forests to local governments and basing promotions on the performance of environmen-
tal protection as well as revenue generation, or designing contracts for SFE managers
that make their interests compatible with both the local governments and the central
or regional environmental authority.
To reiterate, as of 2015, the world’s total forests still have 76 percent under gov-
ernmental ownership (FAO, 2015). Potential policy reforms, when it comes to forest
management, can be both bottom-up and top-down. In addition, there are a number
of potential steps from well-functioning government management to individual or pri-
vate or household management. Increasing evidence reveals a great deal of interest in
6
reform by the customary owners of their land and resources in African, Latin Amer-
ican, and Southeastern Asian countries where deforestation is still problematic. It is
my hope that this thesis – and the data and analyses – will make a small contribution
to increased awareness of this global trend of devolution, and to our understanding of
policy options for effective natural resource management.
References
[1] Baland, J.-M., P. Bardhan, S. Das, and D. Mookherjee. 2010. Forests to the people:
decentralization and forest degradation in the Indian Himalayas. World Develop-
ment 38 (11): 1642-1656.
[2] Coleman, E. A. and F. D. Fleischman. 2012. Comparing forest decentralization and
local institutional change in Bolivia, Kenya, Mexico, and Uganda. World Develop-
ment 40 (4): 836-849.
[3] FAO. 2015. Global Forest Resources Assessment 2015. Food and Agriculture Orga-
nization of the United Nations, Rome, 2015.
[4] Foster A.D., M. R. Rosenzweig, and J.R. Behrman. 2002. Population Growth, In-
come Growth and Deforestation: Management of Village Common Land in India.
University of Pennsylvania.
[5] Kaimovitz, D., C. Vallejos, P. B. Pacheco, and R. Lopez. 1998. Municipal govern-
ments and forest management in lowland Bolivia. The Journal of Environment and
Development 7 (1): 45-59.
[6] Liu, J. and J. Diamond. 2008. Revolutionizing China’s environmental protection.
Science 319 (5859): 37-8.
[7] Meyfroidt, P. and E. F. Lambin. 2008a. Forest transition in Vietnam and its envi-
ronmental impacts. Global Change Biology 14 (6): 1319-1336.
[8] Meyfroidt, P. and E. F. Lambin. 2008b. The causes of the reforestation in Vietnam.
Land Use Policy 25 (2): 182-197.
[9] Montinola, Gabriella, Yingyi Qian, and Barry R. Weingast. 1995. Federalism, Chi-
nese style: the political basis for economic success in China. World Politics 48 (01):
50-81.
7
[10] Nagendra, H., S. Pareeth, S. Bhawna, C. M. Schweik, and K. R. Adhikari. 2008.
Forest fragmentation and regrowth in an institutional mosaic of community, gov-
ernment and private ownership in Nepal. Landscape Ecology 23 (1): 41-54.
[11] National Bureau of Statistics of China (NBSC). 1999, 2015. China Statistical Year-
book 2015. China Statistics Press, Beijing, China (2015).
[12] Rights and Resources Initiative. 2014. What Future for Reform? Progress and
Slowdown in Forest Tenure Reform since 2002. Washington, D.C. 2014.
[13] State Council, P.R. China. 2003. Document No. 9: The Resolution on the Devel-
opment of Forestry. Beijing.
[14] State Forestry Administration, P.R. China. 2011. Forest Resource Statistics of
China. (Beijing, China).
[15] Xu, Chenggang. 2011. The fundamental institutions of China’s reforms and devel-
opment. Journal of Economic Literature 49 (4): 1076-1151.
[16] Xu, Jintao, Andy White, and Uma Lele. 2010. China’s Forest Tenure Reforms:
Impacts and Implications for Choice, Conservation, and Climate Change. Rights
and Resources Initiative. Washington, D.C., USA
8
Chapter I

Forest Devolution Reform in China:
A Trigger for Investment or Deforestation?*
Yuanyuan Yi„1
1Department of Economics, University of Gothenburg, Sweden
Abstract
I investigate whether and how the devolution of forestland to households in China
triggered investment in forestland, and its effect on forest resource conditions. The in-
vestment analysis is based on a panel dataset of a two-round survey of 3,000 households
in eight provinces before and after the implementation of the forest devolution reform,
while the analysis on resource conditions is based on satellite imagery on forest cover and
vegetation during 2001-2012. Using a difference-in-difference propensity score matching
model, I find that the devolution reform resulted in more investment, in terms of annual
labor input days and value of silvicultural treatments per area unit. At the county level,
more forestland under household management is found to improve forest conditions dur-
ing the time period studied. I also investigate the investment effect through two channels:
(i) the effect of tenure security, i.e., holding a forestland certificate and (ii) the reallocation
effect from obtaining more forestland resources. The effects of devolution and improved
tenure on increased private investment and resource conditions provide evidence that
well-defined and protected property rights for households offer an effective alternative to
common-pool resources management in small-scale forestry in China.
JEL Classification: O13, Q23, Q24, Q58
Keywords: forest devolution; household management; forest investment; deforestation;
China
*I am grateful to Yonas Alem, Peter Berck, Allen Blackman, Fredrik Carlsson, Martin Dufwenberg, H˚akan
Eggert, Pamela Jagger, Gunnar Ko¨hlin, Billy Pizer, Thomas Sterner, Jintao Xu, and seminar participants at
the University of Gothenburg for valuable comments and suggestions. I also acknowledge generous financial
support from the following organizations: the Ford Foundation, the National Science Foundation of China
(NSFC), State Forestry Administration of China, Richard C Malmstens Foundation and the Swedish Inter-
national Development Cooperation Agency (Sida) at the Environmental Economics Unit at the University of
Gothenburg and the World Bank PROFOR project.
„Vasagatan 1, SE-405 30 Gothenburg, Sweden, Email: yuanyuan.yi@economics.gu.se
1 Introduction
Devolution of forestland management and its implications for deforestation have con-
cerned both economists and policy-makers for a long time. Devolution means a transfer
of the control over a resource from the central government to lower levels such as com-
munities or households. First-order devolution in forest management means transferring
full responsibility for state forests to local collective management, while second-order
devolution means a transfer to household management. The overall question is whether
or not devolution of authority over natural resources is an efficient policy. Many propo-
nents argue that devolution to communities or households is a good idea because they
have better knowledge of local conditions and the unique characteristics of the resource.
In addition, internalizing protection costs and reducing transaction costs increase ef-
ficiency (Baland and Platteau, 1996; Agrawal and Ribot, 1999; Whitaker and Time,
2001; Kelleher and Yackee 2004; Hyde, 2015). Yet, one main concern with devolution
is that recipients of devolved rights might simply cut down the forest for short-run
profit. There are also concerns about equality and distribution of welfare, in that local
actors may be ill-equipped and lack capabilities to respond to local needs (Kenyon and
Kincaid, 1991; Peterson, 1996) or the poorest residents may not benefit from devolu-
tion. The literature provides no consistency in empirical findings regarding the effects
of forest devolution on user investments and forest conditions and focuses mostly on
forest devolution of the first-order. For example, in a selection of Asian and African
countries as well as South America, positive evidence is found by Ko¨hlin and Amacher
(2005), Engel and Palmer (2006), Bray et al. (2008), Qin and Xu (2013), Holden et
al. (2013a, 2013b), Yi et al. (2014), and Xie et al. (2016), and inconclusive findings
are found by, e.g., Kaimowitz et al. (1998), Baland et al. (2010), and Coleman and
Fleischman (2012).
In this study, my interest is to empirically evaluate the second-order devolution
of forest management. This has been in place in rural China since 2003 through the
Chinese Collective Forest Tenure Reform (State Council, 2003). This devolution reform
allows villages to reallocate village-owned forestland to rural households’ management
and certifies the households’ rights. The expectation has been that a local forest owner’s
decision would incorporate long-term forest returns so that investment in forestland is
incentivized, rather than clear-cutting trees and changing land use to a less sustainable
use.1 The incentives influencing the land-use decisions by households are crucial for
the outcome of forest cover and quality. With devolved forestland and secure tenure,
1 In this study, I use “owners” to represent those who possess tenure rights during the contract period.
I do not distinguish between the concepts of property rights and tenure rights, because in China all land is
owned collectively or by the country as a whole, while management/tenure rights are guaranteed by China’s
Property Law.
1
forest owners could: 1) conserve the standing forests by waiting and not harvesting,
2) invest in the forestland by undertaking silvicultural treatments such as thinning
and gap-filling, and 3) replant trees after harvest. All of these are examples of forest
investments. However, if owners are uncertain of the long-term stability of the reform,
they could immediately reap the “windfall” timber benefits of the mature forest stand
and convert it into other land uses such as cropland, or even abandon the land.
I provide a microeconomic perspective on the impact of second-order devolution
on a forest owner’s propensity to invest or deforest, and forest growth or decline as a
consequence. This reform was implemented by formally devolving forestland to house-
hold management, issuing forestland certificates, and reallocating more village-owned
forestland to households. I use two rounds of survey data from 2,940 randomly selected
households in eight Chinese provinces from south to north, covering the period 2000-
2010. I analyze a total of 10,860 forest plots managed by farmer households. They are
located in 258 villages, where the village committees decided on whether and how to
implement second-order devolution. To assess the devolution effect on investment, the
forest plots are divided into the treated group and the control group. The control group
comprises forest plots with no reform before 2010, and the treated group comprises the
plots reformed between 2006 and 2010. The reform is not completely exogenous to
the villages, the forest plots, or the households, because villages self-selected into the
reform and made many decisions related to implementation.
I adopt a difference-in-difference matching approach based on propensity scores.
This approach controls for the fact that villages self-selected into the reform, and for
the pre-existing observed and unobservable heterogeneity between the treated and the
control groups of forest plots, to identify the average treatment effect of the devolution
reform on investment. Investment is measured by days of labor input and annual silvi-
cultural investments (in Chinese Yuan, or CNY) for per area unit of each plot. I find
a large and statistically significant effect of forest devolution on both labor input and
value of investment. Next, I match the satellite MODIS land cover and the Enhanced
Vegetation Index (EVI) data with the 44 counties involved in the surveys. By instru-
mented fixed effects (FE-IV) models, I find short-run positive impacts of devolution on
forest cover and forest quality. In addition, I investigate two channels through which the
devolution reform triggered investment, i.e., the tenure security effect (through holding
a forestland certificate) and the reallocation effect of obtaining more forestland during
the reform.
The remainder of the paper is structured as follows. Section 2 reviews the literature
on forest devolution and identifies main problems of the mixed findings in previous
studies. Section 3 introduces the history of forest management institutions in China,
and generates two main hypotheses to test. Section 4 describes the data and Section 5
2
presents the empirical strategy to test the hypotheses. Section 6 discusses the results
and the final section concludes.
2 Evidence of Devolution and Forest Conditions
An increasing number of countries have adopted first-order devolution in forest man-
agement. In Asia, these include India (Somanathan, 1991; Agrawal, 2001; Foster et
al., 2002; Ko¨hlin and Amacher, 2005; Behera and Engel, 2006), Indonesia (Engel and
Palmer, 2006), Vietnam (Meyfroidt and Lambin, 2008a; Meyfroidt and Lambin, 2008b),
Nepal (Nagendra et al., 2008), and the Philippines (Dalmacio et al., 2000). In Africa,
these include Uganda (Coleman and Fleischman, 2012), Tanzania (Meshack et al.,
2006; Robinson and Lokina, 2011); in Latin America, Mexico (Antinori and Rausser,
2007; Bray et al., 2008), Bolivia (Kaimowitz et al., 1998; Andersson 2004; Coleman and
Fleischman, 2012) and others (Lynch and Talbott, 1995; Andersson 2004; Agrawal et
al, 2008). However, the effects of such first-order devolution on whether forest is more
sustainably managed are inconclusive in the literature.
For example, in Bolivia, where forest management was devolved to local municipal
governments in the mid-1990s, some studies find no evidence that either the forest
or indigenous people benefited from devolution (Kaimowitz et al., 1998; Coleman and
Fleischman, 2012). Other studies find that effective forest devolution depends on the
connectivity amongst actors and local politicians’ interests in forestry (Andersson,
2004; Andersson et al., 2006). Palmer and Engel (2007) investigated logging and forest
cover in forest-dependent communities in Indonesia before and after decentralization in
1999. They found logging significantly increased, but they did not evaluate the effect on
forest quality. Foster and Rosenzweig (2003) and Baland et al. (2010) both investigated
Indian communities but the results are different. Foster and Rosenzweig (2003) used
national census data and 1971-99 satellite images and found that the effect of common
ownership is positive on forest area and negative on biomass. Baland et al. (2010)
found community forest management leads to no improvements in either forest area or
biomass. Coleman and Fleischman (2012) assessed the effects of forest devolution on
forest conditions in Bolivia, Kenya, Mexico, and Uganda, and found that community
management was effective only in Mexico, with insignificant impact in the others.
Given the inconclusive effectiveness of the first-order forest devolution, a key issue
is whether to take a further step, by devolving forest management to the household
level. A large body of research has provided evidence on the success of second-order
devolution of agricultural land ownership, tenure or use; see, for example, Holden et
al., 2013a; Jacoby et al., 2002; Feder et al., 1988; Bandiera, 2007; Rozelle and Swin-
3
nen, 2004; Goldstein and Udry, 2008. Though the success of the devolution reforms in
cropland tenure or titling is not universal, lessons include how the establishment and
maintenance of land use rights are arranged (Besley, 1995; Carter and Olinto, 2003;
and Deininger and Jin, 2006). Moreover, the devolution of agricultural land in China
in 1978 doubled agricultural productivity in a period as short as six years and led to a
five-fold increase in rural household income in real terms in two decades (NBS, 2014).
Little evidence exists on second-order forest devolution and its effect on forest cover,
and that evidence points in different directions. For example, Meyfroidt and Lambin
(2008a, 2008b) found in Vietnam that first-order devolution of forestland before 1994
had no effect on forest regrowth, but second-order devolution post-1994 was positive
and statistically significant. Nagendra (2007) qualitatively studied a portion of Nepal’s
Chitwan Valley district and compared forest changes among community, government
and private ownership, using Landsat satellite-image data for 1989 and 2000. Com-
munity forests were found more stable than the other two, and the privately owned
forestland was cleared or had fragmented forests.
There are several reasons for the inconclusive findings. First of all, many papers
are case studies, or studies with small sample size and a narrow geographical focus.
Second, reforms are often endogenous to local circumstances and devolution tends
to emerge when local forests become degraded (Baland et al., 2010). This in turn
brings about a greater level of forest improvement than in previously less-degraded
forests.2 Third, except for satellite imagery data, regional-level or subjective estimates
on forest conditions may suffer from systematic measurement error and inconsistencies
in definitions across contexts, or extrapolations from outdated surveys, or other dubious
estimation techniques (Rudel et al., 2005). The fourth reason lies in the lack of focus
on individual behavior. The regional outcome is an aggregation of behaviors at the
micro level. Any variation among the following could affect the impact of devolution
on good forest management: the actors involved in the devolution process and their new
powers, the powers and resources transferred, the accountability of local authorities,
the amount of information, financial and human resources, and the degree of public
participation (Agrawal and Ribot, 1999; Etoungou, 2003; Andersson, 2004; Andersson
et al., 2006; Ribot and Agrawal, 2006).
More recently, studying the same forest devolution reform that is of interest in this
paper, Qin and Xu (2013), Holden et al (2013b), Yi et al. (2014), Huang (2015) and Xie
et al. (2016) documented the reform’s effectiveness in increasing owner investment and
village-level forestation. However, these studies rely on regional data and most of them
2 In the cases where financial aid is available, this positive effect can be expected to be very significant,
because this is usually a companion to devolution to communities and local governments, with the purpose of
protecting forests.
4
focus only on Fujian province. My paper uses a panel dataset of comprehensive, two-
round household and village surveys in eight provinces from south to north China. The
data allows me to address the aforementioned problems. To my knowledge, this paper
is the first study analyzing forest-plot-level panel data from a large-scale household
survey to evaluate the impacts of a second-order forest devolution reform in terms
of private investment and resource conditions. It will provide evidence on how well-
defined and protected property rights for households can be an effective alternative to
common-pool resources management.
3 The Chinese Collective Forest Tenure Reform
The collective forest tenure reform that I study in this paper is the most extensive
devolution of communal forests to households ever seen. By 2008, the reform had
devolved 62 million hectares of the total 100 million hectares of forests from collective
village ownership to individual households (Xu et al., 2010). The reform has involved
600 million people in rural areas in more than twenty provinces of China (State Forestry
Administration, 2011). In this section, I review the background to this reform, and the
features of the reform compared to earlier institutional changes in forest management
in this context.
Since 1954, when all private forests were collectivized, China’s collective forests have
undergone a number of tenure system reforms. What followed in the early 1960s was
returning trees around homesteads to individual households’ control. In the early 1980s,
inspired by the successful Household Responsibility System reform, which contracted
agricultural land to households in the late 1970s, the tenure rights of collectively owned
forests were allowed to be devolved to villagers within the village communities on a large
scale. This is known as the “Three Fixes” policy, in the policy document “Resolution on
Issues Concerning Forest Protection and Development”, announced by the State Council
of China in 1981.
By 1986, roughly 60-70 percent of collectively owned forests was under household
management (Xu et al., 2010). The “Three Fixes” had three characteristics – the
forests were still under collective ownership, there was a lack of clearly defined borders
and use rights, and the implementation was uneven. Excessive timber harvest and
extensive deforestation were perceived as rampant outcomes, especially in association
with the Chinese government’s attempts to liberalize trade control in the mid-1980s. By
1987, the government increased its control over forest management again, along with
a logging quota system. In the 1990s, villages took back forestland from households
and put it under collective control. In the meantime, some forestland was subject to
5
market transactions through, for example, auction of use rights. This marketization
process created opportunities for large-scale private forest management, but the large
poor rural population could not afford to participate in this market and still had no
full property rights to claim returns from the community forests (Hyde et al., 2003).
Soon after the millennium, a new wave of forest tenure reform was initiated in
Fujian province. This time, the focus was on devolution of forest management rights
to households. The main features of the devolution were reallocation of village col-
lectively owned forestland to households, and formal acknowledgement of household
tenure rights to these forest plots. This was very successful and encouraged the cen-
tral government to implement a few other pilot cases starting in 2003 (State Council,
2003). The reform spread to other provinces quickly, and was finally promulgated by
the central government in the policy document “Collective Forest Tenure Reform in the
Southern Collective Forest Areas in China” in 2008.
This round of forest devolution had the following features. First, based on votes
in the village – via village assembly or representative meetings – the village commit-
tee decided on the implementation of forest reallocation. Second, ambitious measures
were adopted to strengthen tenure security. Forestland certificates for each forest plot,
with clearly specified contract terms, were issued to owners. For instance, the tenure
for plots previously called “family plots” was given a clear duration in this round of
reform, ranging from 30 to 70 years, and some of these plots received certificates with
a “long-term” contract duration. The terms specified in the certificates were more
complete compared to the earlier, simpler contracts. Furthermore, the new certificates
often extended the rights to include production and harvest decisions, such as rights to
convert forestland to cropland, select tree or plant species, interchange different forest
types, harvest non-timber forest products, and even abandon plots, as well as transfer-
ability rights to other villagers or outsiders, and the right to use forestland as collateral.
Third, forestland rental/transfers markets were encouraged in the new devolution re-
form. The rights to transfer forestland to people within or even outside villages, and to
mortgage forestland as collateral, were acknowledged, which was unprecedented. The
transferability rights and rental markets are expected to mobilize production factors
and also to provide incentives for owners to invest in forestland in order to transfer it
with a higher price.
With these features, the devolution reform aimed to incentivize forest owners to
make long-term decisions on forestland uses. According to the policy documents, the
expectations included increases in investment and reforestation. Investment can take
one of the following forms: (i) waiting a longer period before harvesting,3 or (ii) un-
3 Waiting for a longer period before harvesting is regarded as an investing behavior in an asset with
expectation of future interest.
6
dertaking silvicultural treatments, such as thinning trees and reforesting after harvest.
Such investments protect the forest and foster forest growth. As discussed earlier, the
opposite could also occur, where households would reap the “windfall” timber benefits
of the mature forest stand and convert it into other land uses, such as cropland, or
even abandon the land.
Given the mixed evidence in the literature on the impacts of devolution in other
countries, and the past challenges in devolving forest management in China, I am
interested in analyzing the impact of the collective forest tenure reform in terms of
investment and forest quality. In order to assess these outcomes of the reform, I will
test the following hypotheses:
Hypothesis 1 . Forest investment increases due to the devolution reform.
Hypothesis 1a . The devolution reform increases forest investment through enhanced
tenure security.
Hypothesis 1b. The devolution reform increases forest investment through reallocation of
forestland. That is, households that receive more forestland during the reform conduct
more forest investment per area unit.
Hypothesis 2 . More forestland devolved to household management contributes to better
forest conditions.
4 Data and Descriptive Statistics
The data used in this study come from two sources. Data regarding forest invest-
ments come from a unique, comprehensive two-round survey of households and vil-
lages in eight Chinese provinces, conducted by the Environmental Economics Program
in China, based at Peking University. After a pilot survey in two counties in Fujian, the
first-round survey was conducted in 45 other counties during the period from March
2006 to August 2007. These counties are located in eight provinces: Fujian, Jiangxi,
Zhejiang, Anhui, Hunan, Liaoning, Shandong, and Yunnan. A stratified random sam-
pling rule was applied to survey 10-20 households in each of the 258 villages from 128
townships in the 45 counties. The second-round survey, in 2011, revisited the same
households. Figure 1 depicts the distribution and survey time of the samples.4
In each round, a household level questionnaire collected information on the past
year on forestland management practices for each forest plot, as well as households’
farming activities, with costs and outputs, and non-farm work and income. Village
4 Two counties in Fujian province were selected for pilot surveys in March 2006, so in the second round
of survey they were not followed up. The pilot samples are thus excluded from the analysis in this paper. In
total, 45 counties were included in the analysis. In the second survey, some observations are missing because
the houshold representative could not be surveyed due to temporary absence such as being in the hospital, or
busy at work, or long-term absence because of migration or death, etc.
7
leaders were asked about the decision-making and implementation of the Collective
Forest Tenure Reform and the community socio-economic characteristics during the
period 2000-2010.
< Figure 1 here >
To measure forest conditions, i.e., forest cover rate and forest quality, the second
data source is spatial data – the MODIS land cover (MOD12Q1) and the Enhanced
Vegetation Index (EVI) data (in MOD13Q1) from satellite images of NASA’s Terra
spacecraft. The data provides spatial resolution up to 250 meters and covers the period
from 2001 to 2012. I match them for the 45 counties with two year lags to the survey,
i.e., in 2002, 2007, and 2012, to allow for the forest management decisions to be better
captured in the satellite images.5 Figures 2 and 3 illustrate the changes in forest cover
and EVI of China as an overview from the satellite.
< Figures 2 and 3 here >
Variable Definitions and Descriptive Statistics
I drop the forest plots that were already reformed in 2005, because of no data about
them prior to the reform. The samples in the pilot survey are also excluded because
they were not followed-up. In order to compare the investment change in the reformed
forest plots with those not reformed, I divide the sample into control and treated
groups: the control refers to forest plots with no reform before 2010, and the treated
group includes the plots reformed between 2006 and 2010. Table 1 summarizes the
investment and reform variables for each group and period.
< Table 1 here >
Forest investment is represented by yearly days of labor input in taking care of the
forestland and the CNY value of other investments such as silvicultural treatments
or/and regeneration efforts. The reformed forest plots have higher levels of investment
in 2010, in both per-plot and per-mu terms.6 The difference seems to exist as early
as 2005, but not to the same level as in 2010. Interestingly, in 2010 a larger share
of plots have positive investment in the treatment group compared with the control
5 Huang (2015) evaluated the same reform’s effect on forest conditions using village shapefiles. In their
study, half of the village maps were not available, so the author created buffers based on the coordinates of
village centers and village land areas. In this paper, I use the administrative county maps to improve the
accuracy and variation in forest cover and quality.
6 Mu is an area unit used in rural China, with 1 mu equal to 1/15 hectare.
8
group, while the opposite was true in 2005, suggesting a positive effect of the reform
on investment.
In the treated group, 58 percent of the plots received their forestland certificates
and 72 percent were in households that obtained more forestland during the reform.
Whether a household received a certificate was due to administrative time and financial
constraints, and variations in resource endowments. I also observe that 27 percent of the
plots in the control group were rented in, so that their households’ access to forestland
increased, too.
< Table 2 here >
Table 2 lists the characteristic variables of households, forest plots and villages that
affect investment incentives. The mean values of demographic characteristics such as
household size, household head’s age, gender, and education are not different between
the two groups in either period. However, differences exist in other household, plot,
and village characteristics between the treated and control groups. The differences
indicate a methodological concern that the pre-treatment differences should be taken
into account, and also suggest differences in characteristics that may correlate with a
village selecting the reform, households’ incentives to invest, and the consequences for
forests.
In general, the treated group has wealthier households than the control group, given
by their higher livestock value, per capita income, and value of house(s). Furthermore,
these households have easier access to credit by closer distance to a local bank, and
better connections to village leadership. The forest plots in the treated group are lo-
cated farther from households’ homes and from the closest paved main road, and are
larger in size, with flatter slope and worse irrigation conditions, less timber and more
so-called “economic” forests such as fruit trees and medicine plantations. The resource
status of the stands looked similar in 2005 in the control and treated groups, i.e.,
3.94, implying near-mature forests. In 2010, the reformed forest plots looked better,
with 4.17, meaning closer to maturity, than 3.85 in the comparison group. So, if better
forestland is reformed first, this fact rather than the reform could drive the change in
investments. Therefore, I control for these factors in the identification.
Importantly, harvesting activities reflect how forest owners realize their property (or
tenure) rights, even though they may harvest due to poor tenure in the short run. In
2005, around 7 percent of the forest plots were harvested. In 2010, almost 20 percent
of the plots were harvested, with a higher percentage in the treated group than in
the control group. Harvest with improved tenure should differ from harvest because of
poor tenure (e.g., insecurity), in that the former would be carried out with regeneration
efforts, such as replanting trees or nurturing the harvested plots. I will look at this by
9
comparing the difference in investment between harvested and non-harvested plots in
the treated group, and the difference in investment in the harvested plots between the
treated and the control groups.
In addition, in parallel with the influence of harvest on investment incentives and
forest outcomes, there is a small number of forest plots with a special role. These are
the forest plots involved in the Sloping Land Conversion Program (SLCP), which are 5
percent of the sample.7 These plots were converted from cropland to forestland, based
on households’ decisions. They accepted governmental payment and in return promised
to protect and not harvest the standing trees. Although the SLCP is independent of
the forest tenure reform, I take into account any spillover effect to the new round of
devolution reform with respect to reform decisions and household investment.
At the village level, population, income, and commercial timber price increased over
time, as well as average precipitation and the number of households with telephones (an
indicator of development). The labor market became more developed, as measured by
the percentage of labor engaging in off-farm work. The overall changes in the economy
are relevant to the increased interest in forestry, possibly driving a village to reform.
In contrast to the control group, the villages that implemented the reform are
located in somewhat more remote areas, farther from paved roads and from the closest
county center. In those respects, they were less developed, but they grew quickly, as
indicated by the population having telephones and the increasing timber prices. They
were larger in size (in terms of the total number of households), had higher per capita
income, and were endowed with more forestland. By contrast, the villages that decided
not to reform had a smaller share of forestland, and households already managed over 80
percent of the forestland in 2005, though without legal acknowledgement. Also, villages
with a less developed labor market, in terms of a smaller proportion of off-farm labor
out of the total labor force, were more likely to select the reform; specifically, less than
30 percent of labor was engaged in off-farm work in the reform villages, compared to
61 percent in villages with no reform.
< Table 3 here >
For the county-level analysis on forest conditions, Table 3 reports the summary
statistics of the variables of interest for the years 2000, 2005 and 2010. From the
National Geomatics Center of China, I obtain the 44 administrative maps for the 45
7 The Sloping Land Conversion Program in China is one of the first and most ambitious payment for
ecosystem services programs in China (Bennett, 2008). The program started in 1999. It encourages farmers
to convert cropland to forests and uses a public payment scheme. The farmer households participating in this
program are obliged to take care of the converted forests and are limited in harvest.
10
counties involved in our survey.8 Forest conditions are computed using the MODIS
products data of 2002, 2007 and 2012. I use two dependent variables – one for forest
cover and one for forest quality.
The forest cover indicator includes evergreen (deciduous) needle leaf (broad leaf)
forests and mixed forests, as defined in the IGBP classification of land cover types in the
standard MODIS product MCD12Q1, at 500 meter spatial resolution in a sinusoidal
projection. The dependent variable, forest cover, is the percentage of land in a county
that is covered by forests, and it is observed to rise in this 10-year period, from 32 to
38 percent.
The forest quality variable is the average value of the vegetation index for each pixel
in the polygon that represents a county region, using the Enhanced Vegetation Index
(EVI) at 250 meter resolution in the MOD13Q1 product data. Overall, I observe a rise
in vegetation indices from 2002 to 2007, and then a decline to a level lower than in the
beginning in 2012.
The county socio-economic variables are generated as the county mean of the sur-
veyed villages weighted by village land area out of the county total. The representa-
tiveness of this approach is justified by the stratified randomness of sampling at each
level of the survey. The degree of forest devolution in one county is measured by the
percentage of area of forestland managed by households. This percentage increased
from 36 to over 43 in the study period. In addition to the institutional dimension –
forest devolution – I take the following factors of economic and social development into
account: per capita income, commercial timber price, daily labor wage, average village
size, and population having telephones.
Finally, the weather data come from the China Meteorological Data Sharing Service
System (CMDSSS) on daily precipitation and temperature.9 Matching with the closest
weather station recorded by the CMDSSS, I compute for each county the annual rainfall
average and its variability. For temperature, because of its nonlinear effect on plants
(Schlenker and Roberts, 2009), I use two aggregates for trees: the effective growing-
degree days of 0-35 degrees Celsius (GDD) and the harmful degree days of >35 degree
Celsius (HDD).10 In this decade, effective GDD is stable but the rainfall average and
variation and the HDD all increased. For example, the average precipitation increased
by 20 percent, from 37.4 mm to 44.9 mm during 2000-2010.
8 In total, 44 counties are matched, because two counties (Taierzhuang and Shanting Districts) belonged
to the same polygon as one county in Zaozhuang City of Shandong province.
9 http://data.cma.cn/
10 GDD is a measure of effective cumulative heat for days with 0-35 degree Celsius ( ◦C). GDD =∑N
i Ti,a − Ti,base, where Ti,a is the daily average temperature for day i, and Ti,base is the base tem-perature below which vegetation ceases to be biologically active (here, 0 ◦C is selected for trees). HDD mea-
sures harmful cumulative heat for the days with temperature higher than 35 ◦C, calculated by HDD =∑N
i (Ti,max − 35 ◦C)/Ti,max, where Ti,max is maximum temperature for day i.
11
5 Empirical Strategy
5.1 Estimating the Impact of Devolution on Forest Investment
To test Hypothesis 1, I estimate the impact of the forest devolution reform on own-
ers’ investments in forestland. As a natural experiment, a valid measure of impact
evaluation should compare outcomes in the plots that received the reform to what the
outcomes would have been if there were no reform. But two challenges exist in the iden-
tification of such an impact. One is that the counterfactuals are unobservable, while
a difference-in-difference estimator compares the outcomes based on observable differ-
ences. The other challenge is induced by unobserved characteristics that drive targets
to self-select into the experiment and also correlate with the outcome of interest, thus
biasing the impact estimate even if the pre-experiment characteristics are controlled
for (Heckman, 1990; Heckman et al., 1998; Heckman and Navarro-Lozano, 2004).
Because villages self-selected into the reform, village resource endowments and
socio-economic development might drive these reform decisions and also correlate with
households’ incentives to invest. However, the fact that the panel data consist of both
a control group with no reform and a treated group before and after the reform allows
me to apply the difference-in-difference propensity score matching approach. With this
approach, I first construct a plausible comparison group by matching the reformed
plots with similar non-reformed ones, based on a rich set of covariates. The covariates
include the plot-, household- and village-level characteristics that potentially influence
reform status and investment incentives. Second, taking advantage of the panel data
setting, changes in investments before and after the reform in the treated group are
compared to the change in investments in the sample of controls between periods. This
process removes possible unobserved time-constant differences between the treated and
the control group (Heckman et al., 1998; Heckman and Navarro-Lozano, 2004). These
differences include household risk and time preferences that are believed to be stable
in the long run and may influence incentives for forest investment.
For this approach, I follow the procedures formalized in Heckman et al. (1997, 1998),
Smith and Todd (2001, 2005), and Gilligan and Hoddinott (2007), and estimate the
average impact of the treatment on the treated (ATT) with the panel dataset:
ATT = E(∆|X,D = 1)
= E(Y 1 − Y 0|X,D = 1)
= E(Y 1|X,D = 1)− E(Y 0|X,D = 1)
= E(Y 1t − Y 1τ |Xτ , D = 1)− E(Y 0t − Y 0τ |Xτ , D = 1). (1)
12
where the superscripts 1 and 0 stand for “treatment” status, with 1 if the village
adopted the reform between 2006 and 2010, and 0 if the village did not adopt the
Reform by 2010. Y 1 = Y 1t − Y 1τ is the outcome, i.e., change in investment, from
2005 (τ ) to 2010 (t), of a forest plot receiving the reform between 2006 and 2010,
and Y 0 = Y 0t − Y 0τ is the change in investment if the forest plot did not receive the
reform. Investment (Y ) is measured by annual, per-mu labor days or the CNY value of
silvicultural investment conducted on each plot. Xτ is a vector of covariates, including
forest plot, household and village characteristics. D is an indicator that an observation
is in the “treatment” group, equal to 1 if in the treated group and 0 if in the control
group.
Because only Y 1 or Y 0 can be observed for each observation, E(Y 0t −Y 0τ |Xτ , D =
1) is not observable. The propensity score matching method allows me to match a
number of similar non-treated to the treated, and to estimate the counterfactual out-
come for the treated observations (Rosenbaum and Rubin, 1983). To match, I let
p(Xτ ) = Pr(D = 1|Xτ ) be the probability of a forest plot being in the treated group,
so that a reform recipient plot is statistically matched to a group of non-reformed forest
plots with similar values of p(Xτ ). To put it differently, the propensity scores, p(Xτ ),
are obtained as the fitted values from estimating the likelihood of receiving the reform,
by using a probit model that includes pre-reform observable characteristics, Xτ . Xτ in-
clude potential determinants of a village selecting the reform and factors affecting forest
investment. Based on the propensity scores, I match the treatment and control obser-
vations using kernel-based matching (KBM), which matches all treated observations
with a weighted average of all controls. The weights in KBM are inversely proportional
to the distance between the propensity scores of treated and control groups. Then, the
average of the differences of each matched pair is computed as ATT . The standard
errors for the impact are estimated by a bootstrap strategy.
The validity of this approach stands on two assumptions: the “conditional mean in-
dependence” (CIA) and common support condition. The former requires that E(Y 0|X,
D = 1) = E(Y 0|X,D = 0), implying that, conditional on the covariates, the obser-
vations in the control group have the same mean outcomes as the treated observations
would have had if they had not been treated. The latter condition, 0 < p(Xτ ) < 1,
requires that valid matches of p(Xτ ) can be found for all values of Xτ .
Next, in order to test Hypotheses 1a and 1b, regarding how the reform triggered
investment through tenure security and through reallocation, I estimate two separate
treatment effects of the reform, interacted with improved tenure security and with re-
allocation, respectively. In other words, I regard a household’s perception of security of
tenure over a specific forest plot due to the reform as one treatment, and a household
receiving more forestland during the reform as another treatment. As discussed in Sec-
13
tion 3, the tenure security effect may occur because the household received a forestland
certificate or formal acknowledgement of tenure rights. In addition, households could
receive more forestland from the administrative reallocation by village committees or
from the forestland rental markets that have been encouraged by the reform. Invest-
ments are therefore expected to be incentivized because of increased access to more
production assets, and through stronger tenure security and resulting gains-from-trade
(Deacon, 1994; Mendelsohn, 1994; Besley, 1995).
5.2 Estimating the Effect of Devolution on Forest Conditions
To test Hypothesis 2, I assess the impact of devolution on forest cover and quality in
a reduced-form regression model:
Fit = α+ θ1sit + θ2Pit + θ3Vit + ωi + µt + εit. (2)
where Fit is forest cover or EVI index of county i at time t; sit, extent of forest
devolution, i.e., percentage of forest under household management, see Table 3; and
Pit and Vit, respectively, are vectors of prices of input factors (e.g., timber price and
off-farm labor wage) and observed geographical and socio-economic characteristics.
These factors drive forest degradation through economic and population growth, and
agricultural expansion, that increase the demand for forest products. In spite of this,
forest could rehabilitate and increase in area and quality via increased output price
because of forest scarcity and/or reduced agricultural expansion, due to increased labor
moving to off-farm work (Foster and Rosenzweig, 2003; Rudel, 1998; Rudel et al., 2005).
ωi represent the unobservable, time-invariant factors; µt is year trends, and εit is the
error term.
The extent of forest devolution, sit, may be endogenous because some unobservable
factors in the error term, εit, may correlate with both sit and Fit. These factors may be
time-constant or varying. Let me take the exogenous geographical conditions first, as
in ωi. Their influences can be controlled by panel fixed effects (FE) models, assuming
there are permanent differences, such as soil conditions, between villages adopting and
not adopting the reform. However, some other unobservable, time-varying influences
cannot be removed by FE estimations. They stem from historical resource changes and
regional socio-economic development and structure. Such influences – including the
regional reliance on forestry, forestland productivity or ability factors, and villagers’
overall bargaining power in calling for the reform – could lead to the variation in
reform implementation amongst villages and counties. To put it simply, these factors
explain why a region is endowed with better institutions and more sustainable resource
14
management. Omitting them in FE estimations would result in biased point estimates.
Therefore, I use 2SLS fixed effects (FE-IV) estimators. The following instrumental
variables (IVs) are used: average and standard deviation of rainfall, the helpful and
harmful cumulative heat measures (GDD and HDD), and the population share with
telephones. Weather conditions may correlate with the general environment for forestry
and forestland productivity. A larger population with telephones suggests better infor-
mation access and social development, so that villagers may have stronger bargaining
power in calling for the reform and thus more forest may be devolved.11 Therefore, the
IVs are relevant (Cov(IV, s) 6= 0), and in the meantime they do not directly affect
forest cover, fulfilling the exclusion restriction (Cov(IV, εit) = 0). In addition, because
vegetation depends on weather conditions as trees grow, I remove the weather variables
from the excluded instruments and include them in the second-stage estimation.
6 Results
The empirical results of impact evaluation involve two sets of estimations: the matching
difference-in-difference estimators of the investment effect of the reform, and the fixed-
effects estimators of the impact of devolution on forest cover and quality.
6.1 The Investment Effect of the Forest Devolution Reform
The impact analysis of the reform is preceded by a specification of the propensity scores
for the treatment variables used to match the treatment recipients and non-recipients.
I analyze three treatments specifically: the reform and its two possible triggers, which
are the enhanced tenure security effect and the forestland reallocation effect. A probit
model is regressed on a broad set of covariates (for 2005) for each treatment variable to
predict the probability of being treated, the results of which are presented in Table 4.
The selection of the covariates is based on the desirability of over-parameterizing the
probit model for the best possible match, conditional on factors highly associated with
the treatment variable and the outcome. Also, the individual parameter estimates from
the model should not possess a causal interpretation, but only association (Heckman
and Navarro-Lozano, 2004; Lee, 2013).
< Table 4 here >
By a series of t-tests on the covariates across the treated and control groups, I check
that the CIA condition is satisfied and the groups are well balanced. Figure 4 distributes
11 Access to the phone grid can also be an indicator of cohesion and political clout, which would affect
tenure reform.
15
the propensity scores between the treated and control observations for each treatment.
Each figure shows that the common support condition is well satisfied by the substantial
overlap of the density distributions of the estimated propensity scores for the treated
and control groups. On the propensity to be treated, the significant coefficients provide
evidence that the reform decisions are associated with socio-economic development and
natural conditions in a village, as well as plot and household characteristics. Column 2
shows that, as expected, receiving a forestland certificate and a higher score of perceived
tenure rights are significantly and positively associated with more secure perception
of ownership of a forest plot.12 In Column 3, the significant estimates suggest that,
in addition to village selection into the reform, household characteristics such as work
experience in a forestry department, as well as plot conditions, correlate with the
likelihood of receiving more forestland during the reform.
< Figure 4 here >
Given the proper matches based on the distribution of the propensity scores, the
effect of the reform and its two triggers on forest owner’s investment is estimated with
a KBM algorithm with 0.01 bandwidth. Table 5 reports the estimates of the average
impacts of the reform on investments (in Columns 1-3). As in Equation (1), the level
of investment, in terms of annual per-mu labor-days or CNY value of silvicultural
investment, is Y 1 = ln(Y 12010/Y 12005) for the treated forest plots and Y 0 = ln(Y 02010/Y 02005)
for the control plots. The results reveal a large and statistically significant effect of
the reform on forestland owners’ investment, both in terms of labor input and value
of investments. Investments increase over time in both the treated and the control
groups, but the increase in the treated group is higher. The ATT estimates for the
reform treatment, 0.119 and 0.328, imply that the reform increases both labor inputs
and investment in a forest plot. The effect is equivalent to a 13 percent increase in the
ratio of average forest labor investment in the reformed plots to that in the matched
non-reformed plots over this period, and a 39 percent increase in the CNY value of
investment in the reformed plots compared to the controls. These findings support
Hypothesis 1.
< Table 5 here >
In Columns 2 and 3, the ATT estimates represent strong average impact of the re-
form through enhanced tenure security and through more forestland reallocated. Lack
of protection for ownership (or tenure) is a common problem in developing countries
12 The nine tenure rights are: production rights such as converting forestland to cropland, selecting tree
or plant species, interchanging different forest types and different land use purposes, using non-timber forest
products, and gains-from-trade rights such as forestland transfers, inheritance, and mortgaging.
16
that jeopardizes the incentives of landowners to invest. Many economists see security as
a prerequisite to successful forest investment (Deacon, 1994; Mendelsohn, 1994; Besley,
1995). In addition to legal acknowledgement of private tenure and rights, the Chinese
forest devolution reform issued forestland certificates, which strengthened tenure se-
curity as perceived by forestland owners, as found in Column 2 of Table 4. Evidently
supporting Hypothesis 1a, the statistically significant estimates (0.348 and 0.526) sug-
gest that perception of security of future ownership encourages a 42 percent increase
in labor input and a 69 percent increase in the CNY value of investment.
Besides the devolution of ownership and improved tenure, the reform encourages
reallocation of village-owned forestland to households and participation in forestland
rental markets. Increased access to production assets enables a household to conduct
more investment. In Column 3, the ATT estimates show that the reform’s reallocation
effect is sizable too: 0.246 (0.294), implying a 28 percent (34 percent) increase in the
labor (value) of investment on a forest plot where its owner received more forestland
during the reform than if it had not been received, supporting Hypothesis 1b.
Further Exploration: Forest Investment as Regeneration Efforts after Harvest
As discussed previously, the right to harvest empowers owners to realize their rights to
claim returns on investment. There also exists opposition to forest devolution, with the
fear that harvest would be conducted without regeneration efforts such as nurturing
and replanting. Due to this concern, I check whether harvesting brought about forest
investment in reformed forest plots (Column 4 of Table 5). The investment variables
(labor input and CNY value of investment) allow us to check this because they are
aggregated, excluding those associated with harvesting activities. The labor and other
investments associated with harvesting are more often regarded as harvest costs. By
contrast, for the variable, an investment increase in plots where harvest was carried
out in the study period reflects post-harvest regeneration efforts or investments. So, I
also compare the difference in investment in the harvested plots between the treated
and the control groups (Column 5).
First of all, as shown in Columns 2 and 3 of Table 4, previous harvests do not un-
dermine households’ perception of tenure security, but may encourage them to acquire
more forestland. On the other hand, harvest activities on the reformed plots are an
outcome of self-selection, where an owner may invest more because he or she wants to
harvest, given the perception of tenure security and other household and resource char-
acteristics. So, I use harvest as a treatment variable and the same set of covariates for
the first-stage propensity estimation and then report the second-stage ATT estimates
in Column 4 of Table 5. The ATT estimates show that the difference in investment in-
17
crease between the devolved plots harvested and the matched, devolved non-harvested
plots is statistically significant. The increase in labor input will be 8 percent higher on a
forest plot after harvest than if no harvest activities were undertaken, and the value of
investment will be 35 percent higher. These findings alleviate the concern that harvest
would be carried out without regeneration efforts on devolved forestland. Furthermore,
in Column 5, the estimated difference in investment between the matched pairs of har-
vested plots with the reform and the harvested plots with no reform is negative for
labor and positive for investment value, though statistically insignificant.13
6.2 Forest Devolution and Forest Conditions
In light of the evidence supporting Hypotheses 1, 1a and 1b, devolution of forest man-
agement to households increases forest investment, including regeneration efforts. As a
result, reduced deforestation and better forest outcomes can be anticipated. Table 6 re-
ports the FE and FE-IV estimation results of Equation (2), on the impact of devolution
on forest cover and forest quality (EVI). Columns 1-2 and 4-5 are FE estimators, where
year trends are included in Columns 2 and 5. Adding the year dummies influences the
estimate of devolution on forest volume change, as expected, as forest foliage naturally
grows with time in all regions. Following the earlier discussion on the endogeneity of
the degree of devolution, the variable of percentage of household-managed forestland
in the FE-IV estimates in Columns 3 and 6 are instrumented separately for forest cover
and EVI. The first-stage results are reported in the last column. The significant coef-
ficient of the population share having telephones and the strong F -statistic satisfy the
condition that the selected instruments are correlated with the endogenous variable.
The Sargan test cannot reject the null that all instruments are uncorrelated with the
error term, satisfying the exclusion restriction.
< Table 6 here >
The FE estimates in Columns 2 and 4 suggest that all forestland devolved to house-
hold management from communal management will lead to an increase of 23 percentage
points in forest cover, and of 6 percentage points in vegetation volume, ceteris paribus.
This effect indicates that households of the average levels of characteristics in the sam-
ples, in the case of obtaining 100 hectares of forestland, will enlarge the regional forest
cover by 23 hectares of forest. By waiting a longer period before harvesting, they will
13 I also employ a simple probit model to investigate the factors that explain why a forest plot had no
investment after harvest. I find plots with stronger tenure rights, greater household wealth and credit access,
and less time spent on off-farm work, are more likely to have post-harvest investment. Plot characteristics seem
to drive post-harvest investment, e.g., plots with fruit trees are more likely to have post-harvest investment
than those with timber trees. The results are not reported in the paper but are available upon request.
18
also improve forest conditions by undertaking more silvicultural treatments; as an ex-
pected result, forest volume will be improved, with a 6 percentage point increase in
vegetation index values.
The difference in the FE versus FE-IV estimates suggests a downward bias of the
FE estimates of devolution’s impact on forest conditions. The FE-IV estimate of the
devolution variable in Column 3 of Table 6 indicates that full devolution from com-
pletely collective management will lead to a 51.8 percentage points increase in forest
cover. Vegetation will have an increase of 7 percentage points from the mean level. In
other words, taking the 2010 sample mean of devolution rate (0.431, in Table 3) for
example, an increase in devolution by one standard deviation (0.194) will lead to an
increase in forest cover by 0.100 (or 26.7 percent of the mean level of forest cover, 0.377)
and an increase in vegetation by 0.014 (or 5.7 percent of the mean indexed vegetation,
0.245). These findings support Hypothesis 2 by suggesting a short-term impact of forest
devolution on forest cover and quality, given that the variables of forest conditions are
measured in two years.
I acknowledge that the devolution effects on forest conditions are identified only
for forestland under household management, instead of the overall marginal effect of
the new round of reform. Household management in the studied area comes from two
sources – the new round of devolution reform and the earlier Sloping Land Conversion
Program.14 I also recognize that the forest cover and EVI are constructed as averages
of the pixels inside a county and do not exclude urban land; therefore, the estimates
may be taken as a lower bound effect on forest conditions.
I also find interesting evidence on economic development and forest increase in area.
Economic growth would be expected to bring about forest decline due to increased
demand for forest products, but would also promote tree-planting because of forest
scarcity (Rudel, 1998; Foster and Rosenzweig, 2003; Rudel et al., 2005). I find, on
the one hand, that timber price increases would drive more deforestation to supply
more timber products and thus reduce forest cover, but the effect is nonlinear. On
the other hand, income growth increases forest cover and the increasing effect slows
down as income grows higher, though the marginal effect estimated by both the FE
and FE-IV models is trivial. A high labor wage attracts rural labor to engage more
in off-farm work; thus, less labor would be spent on enlarging forest cover and taking
14 Due to data availability, I cannot exclude the influences on forest increase of the Sloping Land Conver-
sion Program starting from 1999, which compensated farm households that converted agricultural land into
forest plots and refrained from cutting standing trees before they are mature. Some counties, in particular in
Shandong province, followed the devolution trend by allowing villages to devolve forest plots with ecological
forests, with the aim of encouraging better forest management by devolution. In these two cases, the owners
of forest plots promise not to cut the standing trees, at least before the trees are mature. Though the earlier
investment analysis controlled for the SLCP effect, the estimates of household management effect on forest
conditions would encompass the SLCP’s contribution to forest increase.
19
care of forests, and, as a result, forests would decline in both area and quality. Thus,
I do not find similar evidence as in Rudel et al. (2005) that the labor market moves
rural labor from agricultural expansion by a higher wage, which eases the pressures on
forests, allowing for their spontaneous regeneration.
7 Conclusion
In spite of natural disasters, forest degradation is more often caused by human activities
such as cattle ranching, logging, mining and oil extraction, construction of dams and
roads, and so on. Population and economic growth have been identified as two main
underlying driving factors of forest decline by a number of regional and country-level
studies, e.g., Rudel (1998) and Rudel et al. (2005), Barbier and Burgess (2001a, 2001b),
Foster and Rosenzweig (2003), and Coleman and Fleischman (2012). Weak institutions
and market failures due to non-existing prices for forest goods or services and bad
policies also contribute to forest degradation. For instance, centralized land ownership,
thence poorly defined or non-existing ownership and land tenure arrangements, are the
institutional factors of governance weakness leading to forest declines (Ostrom, 1990;
Deacon 1994; Mendelsohn, 1994).
The trends encouraging privatization and devolving collaborative resource manage-
ment emerged from the concern about the prediction of the tragedy of the commons by
Garrett Hardin (1968). Ostrom (1990) contributed to collective action and common-
pool-resources institutions with her famous eight design principles. Practically, the
findings on the first-order devolution effect on forest conditions are largely mixed in
the existing literature. The second-order devolution, to household management, has
undeniably improved the welfare of the local population compared to the previous sys-
tem of collective management of forestland in China (Xu et al., forthcoming). China
started devolution due to previously weak institutional management of public assets,
where non-cooperation and forest degradation were generally observed under collective
management.
This paper provides evidence of the positive effects in terms of investment and for-
est cover of the Chinese second-order devolution, implying that private management of
forests can be an alternative to common-pool resource management. I studied house-
holds’ behaviors as responses to the devolution of forestland management. I combined
repeated household surveys and satellite imagery data. The econometric results indi-
cate that privatizing forest management to households triggered more investment, and
had a positive impact on forest conditions. But we have to also keep in mind that there
does not exist a unique institutional arrangement that fits all. For China, the success of
20
forestland devolution can to a large degree be attributed to the reallocation of forest-
land and improved tenure security. By reallocation of forestland, farmers’ access to
resources and production factors is improved. With forestland certificates being issued
to farmer households, the contract terms are clearly specified. Ownership security is
strengthened to a degree unprecedented in China, which guarantees the right to claim
future returns on investments. Moreover, private ownership of forests has a longer his-
tory in industrial countries, where private property rights have been well protected by
institutions and rules of law. For developing countries, the path to forest transitions
from deforestation and degradation is still long.
References
[1] Agrawal, A. 2001. State formation in community spaces? Decentralization of control
over forests in the Kumaon Himalaya, India. The Journal of Asian Studies 60 (01):
9-40.
[2] Agrawal, A. and J.C. Ribot. 1999. Accountability in decentralization: a framework
with South Asian and West African cases. Journal of Developing Areas 33: 473-502.
[3] Agrawal, A., A. Chhatre, and R. Hardin. 2008. Changing governance of the world’s
forests. Science, 320 (5882): 1460-1462.
[4] Andersson, K. P. 2004. Who talks with whom? The role of repeated interactions in
decentralized forest governance. World Development 32 (2): 233-249.
[5] Andersson, K.P., C.C. Gibson, and F. Lehoucq. 2006. Municipal politics and forest
governance: comparative analysis of decentralization in Bolivia and Guatemala.
World Development 34 (3): 576-595.
[6] Antinori, C. and G. Rausser. 2007. Collective choice and community forestry man-
agement in Mexico: an empirical analysis. The Journal of Development Studies 43 (3):
512-536.
[7] Baland, J. M., and J. P. Platteau. 1996. Halting Environmental Degradation. Is There
a Role for Rural Communities? Oxford: Oxford University Press.
[8] Baland, J.-M., P. Bardhan, S. Das, and D. Mookherjee. 2010. Forests to the people:
decentralization and forest degradation in the Indian Himalayas. World Development
38 (11): 1642-1656.
[9] Bandiera, O. 2007. Land tenure, investment incentives, and the choice of techniques:
evidence from Nicaragua. World Bank Economic Review 21 (3): 487-508.
21
[10] Barbier, E. B. and J. C. Burgess. 2001a. Tropical deforestation, tenure insecurity,
and unsustainability. Forest Science 47 (4): 497-509.
[11] Barbier, E. B. and J. C. Burgess. 2001b. The economics of tropical deforestation.
Journal of Economic Surveys 15 (3): 413-433.
[12] Behera, B. and S. Engel. 2006. Institutional analysis of evolution of joint forest
management in India: a new institutional economics approach. Forest Policy and
Economics 8 (4): 350-362.
[13] Bennett, M. T. 2008. China’s sloping land conversion program: institutional inno-
vation or business as usual? Ecological Economics 65 (4): 699-711.
[14] Besley, T. 1995. Property rights and investment incentives: theory and evidence
from Ghana. Journal of Political Economy 103 (5):903-37.
[15] Bray, D.B., E. Duran, V. H. Ramos, J-F. Mas, A. Velazquez, R. B. McNab, D.
Barry, and J. Radachowsky. 2008. Tropical deforestation, community forests, and
protected areas in the Maya Forest. Ecology and Society 13 (2).
[16] Carter, M., and P. Olinto. 2003. Getting institutions “right” for whom? Credit
constraints and the impact of property rights on the quantity and composition of
investment. American Journal of Agricultural Economics 85 (1): 173-186.
[17] Channan, S., K. Collins, and W. R. Emanuel. 2014. Global mosaics of the stan-
dard MODIS land cover type data. University of Maryland and the Pacific Northwest
National Laboratory, College Park, Maryland, USA.
[18] Coleman, E. A. and F. D. Fleischman. 2012. Comparing forest decentralization
and local institutional change in Bolivia, Kenya, Mexico, and Uganda. World De-
velopment 40 (4): 836-849.
[19] Dalmacio, M. V., E. S. Guiang, B. R. Harker, and W. F. Hyde. 2000. Secure forest
tenure, community management, and deforestation: a Philippine policy application.
In Economics of Forestry and Rural Development: An Empirical Introduction from Asia.
The University of Michigan Press.
[20] Deacon, R. T. 1994. Deforestation and the rule of law in a cross-section of countries.
Land Economics 70 (4): 414-430.
[21] Deininger, K., and S. Jin. 2006. Tenure security and land-related investment: evi-
dence from Ethiopia. European Economics Review 50 (2006): 1245-1277.
22
[22] Engel, S., and C. Palmer. 2006. Who owns the right? The determinants of commu-
nity benefits from logging in Indonesia. Forest Policy and Economics 8 (4): 434-446.
[23] Feder, G., Y. Chalamwong, T. Onchan, and C. Hongladarom (Eds.). 1988. Land
Policies and Farm Productivity in Thailand. Baltimore and London: Johns Hopkins
University Press.
[24] Foster, A. D., and M. R. Rosenzweig. 2003. Economic growth and the rise of
forests. The Quarterly Journal of Economics 118 (2): 601-637.
[25] Foster A.D., M. R. Rosenzweig, and J.R. Behrman. 2002. Population Growth, In-
come Growth and Deforestation: Management of Village Common Land in India. Uni-
versity of Pennsylvania.
[26] Friedl, M. A., D. Sulla-Menashe, B. Tan, A. Schneider, N. Ramankutty, A. Sibley,
and X. Huang. 2010. MODIS collection 5 global land cover: algorithm refinements
and characterization of new datasets. Remote Sensing of Environment 114 (1): 168-
182.
[27] Gilligan, D. O. and J. Hoddinott. 2007. Is there persistence in the impact of
emergency food aid? Evidence on consumption, food security, and assets in rural
Ethiopia. American Journal of Agricultural Economics 89 (2): 225-242.
[28] Goldstein, M. and C. Urdy. 2008. The profits of power: land rights and agricultural
investment in Ghana. Journal of Political Economy 116 (6): 980-1022.
[29] Hardin, Garrett. 1968. The tragedy of the commons. Science 162 (3859): 1243-48.
[30] Heckman, J.J. 1990. Varieties of selection bias. American Economic Review 80: 313-
318.
[31] Heckman, J. J., H. Ichimura, J. Smith, and P. Todd. 1997. Matching as an econo-
metric evaluation estimator: evidence from evaluating a job training programme.
The Review of Economic Studies 64 (4): 605-654.
[32] Heckman, J. J., H. Ichimura, J. Smith, and P. Todd. 1998. Characterizing selection
bias using experimental data (No. w6699). National Bureau of Economic Research.
[33] Heckman, J. and S. Navarro-Lozano. 2004. Using matching, instrumental variables,
and control functions to estimate economic choice models. Review of Economics and
Statistics 86 (1): 30-57.
23
[34] Holden, S. T., K. Otsuka, and K. Deininger. 2013a. Land Tenure Reform in Asia
and Africa: Assessing Impacts on Poverty and Natural Resource Management. Palgrave
Macmillan.
[35] Holden, S. T., Y. Yi, X. Jiang and J. Xu. 2013b. Tenure security and investment
effects of forest tenure reform in China, in Holden, S., Otsuka, K., and Deininger,
K. (eds), Land Tenure Reform in Asia and Africa. pp. 256-282. Palgrave Macmillan.
[36] Huang, Zhumei. 2015. Forest Tenure Reform and Its Impact on Forest Growth. Mas-
ter’s Thesis (in Chinese), College of Environment Science and Engineering, Peking
University.
[37] Hyde, W.F., B. Belcher and J. Xu (Eds.). 2003. China’s Forests: Global Lessons from
Market Reforms. Washington, D.C.: Resources for the Future Press.
[38] Hyde, William F. 2015. Whereabouts devolution and collective forest manage-
ment? Working paper, presented at the International Conference on New Frontiers
of Forest Economics (NeFFE), Peking University, Beijing, China, August 17-23,
2015.
[39] Jacoby, H. G., G. Li, S. Rozelle. 2002. Hazards of expropriation: tenure insecurity
and investment in rural China. American Economic Review 92 (5): 1420-47.
[40] Kaimovitz, D., C. Vallejos, P. B. Pacheco, and R. Lopez. 1998. Municipal govern-
ments and forest management in lowland Bolivia. The Journal of Environment and
Development 7 (1): 45-59.
[41] Kelleher, C. A. and S. W. Yackee. 2004. An empirical assessment of devolution’s
policy impact. Policy Studies Journal 32 (2): 253-270.
[42] Kenyon, D. A., and J. Kincaid (Eds.). 1991. Competition Among States and Local
Governments: Efficiency and Equity in American Federalism. Washington, DC: Urban
Institute Press.
[43] Ko¨hlin, G. and G. S. Amacher. 2005. Welfare implications of community forest
plantations in developing countries: the Orissa social forestry project. American
Journal of Agricultural Economics 87 (4): 855-869.
[44] Lee, W. 2013. Propensity score matching and variations on the balancing tests.
Empirical Economics 44 (1): 47-80.
[45] Lynch, O. J. and K. Talbott. 1995. Balancing Acts: Community-Based Forest Man-
agement and National Law in Asia and the Pacific. World Resources Institute.
24
[46] Mendelsohn, R. 1994. Property rights and tropical deforestation. Oxford Economic
Papers Vol. 46: 750-756.
[47] Meshack, C. K., B. Ahdikari, N. Doggart, and J. C. Lovett. 2006. Transaction
costs of community-based forest management: Empirical evidence from Tanzania.
African Journal of Ecology 44 (4): 468-477.
[48] Meyfroidt, P. and E. F. Lambin. 2008a. Forest transition in Vietnam and its
environmental impacts. Global Change Biology 14 (6): 1319-1336.
[49] Meyfroidt, P. and E. F. Lambin. 2008b. The causes of the reforestation in Vietnam.
Land Use Policy 25 (2): 182-197.
[50] Nagendra, H., S. Pareeth, S. Bhawna, C. M. Schweik, K. R. Adhikari. 2008. Forest
fragmentation and regrowth in an institutional mosaic of community, government
and private ownership in Nepal. Landscape Ecology 23 (1): 41-54.
[51] National Bureau of Statistics of the People’s Republic of China (NBS). 2014. China
Statistical Yearbook 2014, China Statistics Press.
[52] Ostrom, E. 1990. Governing the Commons: The Evolution of Institutions for Collective
Action. Cambridge: Cambridge University Press.
[53] Peterson, P. 1996. The Price of Federalism. Washington, DC: Brookings Institution
Press.
[54] Qin, P., and J. Xu. 2013. Forest land rights, tenure types, and farmers’ invest-
ment incentives in China: an empirical study of Fujian Province. China Agricultural
Economics Review 5 (1): 154-170.
[55] Robinson, E. J., and R. B. Lokina. 2011. Efficiency, enforcement and revenue trade-
offs in participatory forest management: an example from Tanzania. Environment
and Development Economics 1-19.
[56] Rosenbaum, P., and D.B. Rubin. 1983. The central role of the propensity score in
observational studies for causal effects. Biometrika 70: 41-55.
[57] Rozelle, S. and J.F.M. Swinnen. 2004. Success and failure of reform: insights from
the transition of agriculture. Journal of Economic Literature 42 (2): 404-56.
[58] Rudel, T. 1998. Is there a forest transition? Deforestation, reforestation, and de-
velopment. Rural Sociology 63 (4): 533-552.
25
[59] Rudel, T. K., O. T. Coomes, E. Moran, F. Achard, A. Angelsen, J. Xu and E.
Lambin. 2005. Forest transitions: towards a global understanding of land use change.
Global Environmental Change 15 (1): 23-31.
[60] Schlenker, Wolfram and Michael J. Roberts. 2009. Nonlinear temperature effects
indicate severe damages to U.S. crop yields under climate change. Proceedings of the
National Academy of Sciences of the United States of America 106 (37): 15594-98.
[61] Smith, J. A. and P. E. Todd. 2001. Reconciling conflicting evidence on the per-
formance of propensity-score matching methods. The American Economic Review 91
(2): 112-118.
[62] Smith, J. A. and P. E. Todd. 2005. Does matching overcome LaLonde’s critique
of nonexperimental estimators? Journal of Econometrics 125 (1): 305-353.
[63] Somanathan, E. 1991. Deforestation, property rights and incentives in the Central
Himalayas. Economic and Political Weekly 26: 37-46.
[64] State Council, P.R. China. 2003. Document No. 9: The Resolution on the Development
of Forestry. Beijing.
[65] State Forestry Administration, P.R. China. 2011. Forest Resource Statistics of China.
(Beijing, China).
[66] Whitaker, I. P. and V. Time. 2001. Devolution and welfare: the social and legal
implications of state inequalities for welfare reform in the United States. Social
Justice 28 (1 (83)): 76-90.
[67] Xie, L., P. Berck, and J. Xu. 2016. The effect on forestation of the collective forest
tenure reform in China. China Economic Review 38: 116-129.
[68] Xu, J., K. Deininger, and W. Hyde. Forthcoming (Eds.). China’s Collective Forests:
Global Lessons from Tenure Reform. Washington: World Bank.
[69] Xu, Jintao, Andy White, and Uma Lele. 2010. China’s Forest Tenure Reforms: Im-
pacts and Implications for Choice, Conservation, and Climate Change. Rights and Re-
sources Initiative, Washington, D.C.
[70] Yi, Y., G. Ko¨hlin, and J. Xu. 2014. Property rights, tenure security and forest
investment incentives: evidence from China’s Collective Forest Tenure Reform. En-
vironment and Development Economics 19: pp 48-73.
26
Figure 1: Distribution of survey samples.
Survey Time Province County Township Village Household Forest plots
March-April 2006 Fujian 8 23 46 419 1385
June 2011 8 24 46 281 1013
May 2006 Jiangxi 5 15 30 279 884
March 2011 5 15 30 195 771
Oct-Nov 2006 Zhejiang 6 18 34 299 1339
July 2011 6 18 35 286 1739
April 2007 Anhui 5 15 30 283 1067
July 2011 5 15 30 249 1526
April 2007 Hunan 5 15 27 105 105
May-June 2011 5 15 29 202 689
May-June 2007 Liaoning 5 15 29 263 642
July 2011 5 14 28 254 756
May-June 2007 Shandong 5 15 30 263 482
August 2011 5 15 30 236 573
August 2007 Yunnan 6 12 30 485 1391
August 2011 6 12 30 502 1942
1st round Total 45 128 256 2396 7295
2nd round 45 128 258 2205 9009
Notes: The figures in the table exclude the two counties surveyed as pilots in Fujian province,
because they were not followed up in the second round of surveys, nor are the pilots included
in the analysis. The increase in the number of villages was because some administrative villages
split during the research period. Other missing samples were due to temporary absence, such as
being in a hospital, being at work, migration or death, etc. The first round survey skipped over
2006 and 2007 due to the unavailability of financial resources.
27
Figure 2: Forest cover in China, 2002-2012.
(a) Forest cover in 2002
(b) Forest cover in 2007
(c) Forest cover in 2012
Source: Channan et al. (2014); Friedl et al. (2010) − MODIS land cover type data
28
Figure 3: Trend in the Enhanced Vegetation Index (EVI), China, 2002-2012.
(a) EVI in 2002
(b) EVI in 2007
(c) EVI in 2012
Source: Channan et al. (2014); Friedl et al. (2010) − MODIS land cover data on EVI
29
Figure 4: Propensity score distribution and common support for
difference-in-difference propensity score matching estimation.
Notes: “Reformed (Treated) 2005-2010”, indicated by the red bars, means the
observations in the treated group that have a suitable comparison in the “No-Reform
(Control)” – the blue bars. “Treated: Off support”, indicated by the green bars,
means the observations in the treated group that do not have a suitable comparison.
30
Table 1: Summary statistics for investment and reform measures
2005 2010
Control Treated Control Treated
Labor input, days
Labor input on plot (days)∗ 13.89 20.10 11.53 19.89
(29.89) (38.23) (17.51) (41.38)
Labor input per mu (days)∗ 5.54 7.25 5.95 7.99
(9.13) (11.73) (9.63) (13.62)
If labor input is nonzero (1 if >0; 0 if otherwise) 0.547 0.491 0.414 0.461
(0.498) (0.500) (0.493) (0.499)
Value of investment, CNY
Value of investment on plot (CNY)∗ 725.49 698.79 549.80 855.87
(1650.99) (1484.64) (1444.26) (1855.36)
Value of investment per mu (CNY)∗ 166.29 170.94 213.45 342.00
(251.82) (234.94) (386.44) (709.47)
If investment is nonzero (1 if >0; 0 otherwise) 0.308 0.271 0.285 0.290
(0.462) (0.444) (0.452) (0.454)
Reform dummy, 1=yes, 0=no 0 0 0 1
(0) (0) (0) (0)
Received Forestland Certificate, 1=yes, 0=no 0 0 0 0.581
(0) (0) (0) (0.493)
Obtained more forestland, 1=yes, 0=no 0 0 0.271 0.722
(0) (0) (0.445) (0.448)
Number of observations 406 4229 649 5854
Notes: * The values are sample means among plots with nonzero labor input and nonzero investment,
respectively. Mu is a land area unit used in rural China, with 1 mu equal to 1/15 hectare. The control
group is defined as the forest plots with no reform in the village before 2010; the treated group includes
forest plots where the village selected the reform between 2006 and 2010. The table summarizes differences
in investments between the two groups of forest plots.
31
Table 2: Sample mean: household, forest plot, and village characteristics
2005 2010
Control Treated Control Treated
Household characteristics
Household size 4.877 4.342 4.143 4.249
Gender of household head (1=male, 0=female) 0.953 0.966 0.959 0.951
Age of household head 51.61 48.97 53.21 52.23
Household head education, years 5.488 5.971 5.550 5.890
Household head is Communist (1=yes, 0=no) 0.155 0.208 0.251 0.213
Household head is village leader (1=yes, 0=no) 0.0788 0.102 0.0647 0.0699
Work experience in forestry department (1=yes, 0=no) 0.00739 0.0175 0.0709 0.0687
Credit access (distance to local bank, km) 9.713 8.414 15.50 11.48
Per-labor nonfarm work days 108.4 103.1 112.9 112.7
Per capita cropland size (in mu) 1.849 2.113 1.709 1.636
Household livestock value (CNY) 789.3 1,585 1,491 3,970
Per capita income (CNY) 5,820 6,595 9,770 12,038
House value (10,000 CNY) 4.443 4.923 12.35 13.77
Forest plot characteristics
Forest plot size, mu a 4.191 10.42 4.495 8.381
Distance to home, km 0.971 1.862 0.987 1.769
Distance to main road, km 1.271 1.593 0.804 1.194
Irrigation condition, 1-4 b 3.246 3.579 2.928 3.301
Slope (1: <25◦; 2: 25-45◦; 3: >45◦) 2.151 1.595 2.066 1.520
Sloping Land Conversion Program (1=yes, 0=no) 0.0123 0.0419 0.0139 0.0516
Timber forest (1=yes, 0=no) 0.727 0.452 0.750 0.648
Economic forest (1=yes, 0=no) 0.0985 0.167 0.106 0.293
Bamboo forest (1=yes, 0=no) 0.0911 0.106 0.159 0.125
Current resource status (1-6) c 3.946 3.945 3.849 4.175
If harvested (1=yes, 0=no) 0.0813 0.0641 0.173 0.190
Bundle of rights score, 0-9 d 6.628 6.088 5.581 5.676
Perception of ownership risk (1=secure) 0.919 0.955 0.871 0.953
Village characteristics
Village distance to paved road, km 0.610 6.914 0.277 1.487
Village distance to closet county center, km 26.64 36.32 30.85 35.58
Village total forestland area, ha 5,043 16,685 5,241 18,004
Village forestland area share 0.561 0.709 0.572 0.703
Proportion of village forest managed by household 0.869 0.804 0.836 0.769
Village total population 1,278 1,503 1,082 1,601
Village size, number of households 283.2 383.7 276.6 425.5
Village average per capita income (CNY) 2,146 2,588 3,467 3,859
Proportion of households having telephone 0.779 0.724 0.871 0.930
Development of labor market (Off-farm labor/total) 0.311 0.292 0.610 0.257
Commercial timber price, CNY 500 422 580 579
Precipitation: annual average, mm 36.87 35.32 39.81 44.29
Precipitation: standard deviation, mm 113.4 93.33 111.6 108.3
Effective cumulative heat, 0-35◦C 6,101 5,871 5,953 6,014
Harmful cumulative heat, >35◦C 8.137 4.472 8.859 5.037
Number of observations 406 4229 649 5854
Notes: a: 1 mu =1/15 hectare.
b: Irrigation condition takes four values: 1 means the best and 4 means no irrigation.
c: Current resource takes 6 values, 1 if scattered trees, 2 if very young trees, 3 if young trees, 4 if near-mature
forest, 5 if mature or old forest, and 6 if bamboo, or other “economic” forest such as fruit trees and medicine
plantations.
d: The rights scores are summed from the nine rights: production rights such as converting forestland to
cropland, selecting tree or plant species, interchanging different forest types and different land use purposes,
using non-timber forest products; and gains-from-trade rights such as forestland transfers, inheritance, and
mortgaging.
32
Table 3: Summary statistics of forest conditions and county characteristics
2000 2005 2010
Mean Std. Dev. Mean Std. Dev. Mean Std. Dev.
Forest cover in 2 years 0.321 0.239 0.355 0.252 0.377 0.257
Forest quality (EVI) in 2 years 0.261 0.156 0.266 0.159 0.245 0.147
Proportion of village forest of household management * 0.359 0.195 0.418 0.201 0.431 0.194
Per capita income, CNY * 1,217 695.0 1,592 1,044 2,372 1,469
Commercial timber price, CNY * 193.8 102.4 256.4 123.5 523.9 601.3
Forestland as percentage of village land area * 0.150 0.102 0.154 0.102 0.162 0.105
Labor daily wage, CNY * 19.95 10.76 27.75 15.57 44.40 26.40
Average village population * 865.5 607.2 902.7 662.4 974.2 698.4
Household size * 2.426 1.134 2.374 1.124 2.296 1.010
Population share with telephone * 0.186 0.114 0.455 0.202 0.595 0.245
Precipitation: annual average, mm 37.44 11.86 38.40 11.84 44.90 15.25
Precipitation: standard deviation, mm 100.3 25.28 109.4 34.10 118.0 30.63
Growing degree-days (GDD, 0-35◦C) 6,036 1,174 6,098 1,191 6,141 1,241
Harmful degree-days (HDD, >35◦C) 2.861 3.072 5.396 5.566 6.241 6.527
Number of observations 44 44 44
Notes: Forest cover includes evergreen (and deciduous) needle leaf (and broad leaf) forests and mixed forests, as defined in the
IGBP classification of land cover types in the standard MODIS product MCD12Q1. Forest quality is computed as the average
value of the Enhanced Vegetation Index (EVI) in the MOD13Q1 product data for each county, and the values are divided by
1000 for scaling purpose.
*: Variables are generated as mean of sample villages for each county.
33
Table 4: Probit estimates of the propensity to treatment
Treatment 1 Treatment 2 Treatment 3
Reform Security Reallocation
Log village population 0.229*** 0.268*** 0.131*
(0.081) (0.073) (0.068)
Log village average income 0.174 0.165 -0.134
(0.221) (0.171) (0.170)
Proportion of households having telephone 2.946*** 2.721*** 2.375***
(0.835) (0.577) (0.732)
Development of labor market (Off-farm labor/total) -1.269*** -0.541 -1.277***
(0.406) (0.343) (0.339)
Log village total forestland area 0.006 0.004 -0.002
(0.045) (0.040) (0.039)
Proportion of village forest managed by household -0.567* -0.630*** -0.070
(0.307) (0.225) (0.227)
Log distance to closet county, km 0.017 -0.015 0.009
(0.119) (0.085) (0.089)
Log timber price 0.142 -0.047 -0.007
(0.158) (0.121) (0.119)
Precipitation: annual average, mm 0.026*** 0.012* 0.025***
(0.009) (0.007) (0.007)
Precipitation: standard deviation, mm -0.001 -0.002 -0.002
(0.005) (0.004) (0.003)
Effective cumulative heat, 0-35◦C 0.000*** 0.000*** 0.000**
(0.000) (0.000) (0.000)
Harmful cumulative heat, >35◦C -0.033 -0.039** -0.008
(0.022) (0.018) (0.018)
Log household size -0.395*** -0.376*** -0.071
(0.110) (0.108) (0.132)
Gender of household head (1=male, 0=female) -0.204* -0.028 -0.274**
(0.113) (0.181) (0.138)
Household head age 0.880*** 0.637*** 0.408***
(0.133) (0.134) (0.133)
Household head education years 0.141** 0.133** 0.040
(0.059) (0.053) (0.062)
Household head is Communist (1=yes, 0=no) -0.262*** -0.209*** -0.228***
(0.075) (0.074) (0.088)
Household head is village leader (1=yes, 0=no) -0.272*** -0.125 -0.188*
(0.089) (0.095) (0.102)
Work experience in forestry department (1=yes, 0=no) 0.669*** 0.365*** 0.600***
(0.158) (0.123) (0.140)
Log per-labor nonfarm work days -0.068 -0.054 -0.005
(0.052) (0.040) (0.052)
Log per-capita cropland size 0.046 0.056 0.073
(0.094) (0.065) (0.091)
Log household livestock value 0.071*** 0.052*** 0.055***
(0.011) (0.011) (0.009)
Log per-capita income 0.173*** 0.129*** 0.135***
(0.031) (0.029) (0.028)
Log house value 0.251*** 0.154*** 0.187***
(0.037) (0.032) (0.035)
Log distance to local bank 0.018 0.073 0.015
(0.092) (0.067) (0.072)
Log distance to home 0.071 0.026 0.053
(0.065) (0.061) (0.069)
continued on next page
34
Table 4 continued
Log distance to main road -0.205*** -0.235*** -0.202***
(0.056) (0.055) (0.063)
Irrigation condition -0.118*** -0.108*** -0.107***
(0.026) (0.024) (0.024)
Slope (1: <25◦; 2: 25-45◦; 3: >45◦) -0.087** 0.009 -0.044
(0.037) (0.032) (0.039)
Sloping Land Conversion Program plot (1=yes, 0=no) -0.032 -0.165 0.001
(0.115) (0.110) (0.108)
Timber forest (1=yes, 0=no) 0.267*** 0.129* 0.158***
(0.075) (0.068) (0.055)
Economic forest (1=yes, 0=no) 0.355*** 0.170** 0.264***
(0.100) (0.086) (0.077)
Current resource status 0.005 -0.016 -0.003
(0.027) (0.021) (0.022)
If harvested (1=yes, 0=no) 0.407*** 0.311*** 0.277***
(0.088) (0.078) (0.080)
Bundle of rights score, 0-9 -0.006 -0.038*** -0.008
(0.014) (0.014) (0.012)
Perception of ownership risk (1=secure; 0 otherwise) 0.803*** 2.169*** 0.512***
(0.183) (0.225) (0.171)
Receiving forestland certificate 1.861***
(0.134)
Observations 10,860 10,860 10,860
Log Lik -4310 -3486 -5121
Pseudo R-squared 0.427 0.536 0.292
Notes: The probit models regress on all village and household variables of 2005. Robust standard
errors in parentheses. Significance is denoted: *** p<0.01, ** p<0.05, * p<0.1
35
Table 5: DID estimates of the impact of the Reform on forestland owner investment
(1) (2) (3) (4) (5)
Treatment 1 Treatment 2 Treatment 3 Reformed: Harvested:
Harvest vs. Reform vs.
The Reform Tenure Security Reallocation No harvest No reform
Labor input per mu, days
Average in the treated 0.492 0.514 0.581 0.508 0.449
Average in the control 0.373 0.166 0.335 0.427 0.470
Difference in average outcomes (ATT) 0.119* 0.348*** 0.246*** 0.081* -0.020
(0.070) (0.043) (0.038) (0.045) (0.122)
Value of investment per mu, CNY
Average in the treated 0.942 0.977 1.065 1.058 0.988
Average in the control 0.614 0.450 0.771 0.76 0.812
Difference in average outcomes (ATT) 0.328*** 0.526*** 0.294*** 0.298*** 0.176
(0.117) (0.100) (0.077) (0.082) (0.247)
Notes: Column 4 reports the ATT estimates of post-harvest investment in the reformed plots, by comparing the difference in
investment between the matched, harvested plots and non-harvested plots using the treated group. Column 5 reports the ATT
estimates of the reform’s effect on investment if harvested, by comparing the difference in investment between the matched
pairs of harvested plots with the reform and the harvested plots with no reform. Standard errors of the ATT estimates are in
parentheses, bootstrapped using 200 replications. Significance is denoted: *** p<0.01, ** p<0.05, * p<0.1.
36
Table 6: Devolution impact on forest conditions (FE and FE-IV estimators)
Forest Cover Forest Quality (EVI) IV 1st Stage
(1) FE (2) FE (3) FE-IV (4) FE (5) FE (6) FE-IV % devolved
Proportion of household-managed forestland 0.243*** 0.233*** 0.518*** 0.061** 0.025 0.070*
(0.053) (0.058) (0.125) (0.023) (0.024) (0.041)
Household size -0.022 -0.018 -0.010 0.023** 0.025** 0.023** 0.045
(0.025) (0.025) (0.027) (0.011) (0.010) (0.010) (0.051)
Commercial timber price -0.750* -0.831** -0.427 -0.016 0.014 -0.014 -1.094
(0.405) (0.407) (0.430) (0.178) (0.164) (0.162) (0.792)
Timber price squared 2.661** 2.892** 2.308* 0.130 0.080 0.137 1.899
(1.268) (1.271) (1.363) (0.556) (0.513) (0.508) (2.481)
Labor daily wage -0.001 -0.001 -0.002** -0.000 -0.000 -0.000 0.003*
(0.001) (0.001) (0.001) (0.000) (0.000) (0.000) (0.002)
Labor wage squared 0.000 0.000 0.000 -0.000 0.000 -0.000 -0.000
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Average village population -0.000 -0.000 0.000 0.000 0.000 0.000 0.000
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Population squared 0.000 0.000 0.000 -0.000 -0.000 -0.000 -0.000
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Per capita income 0.000*** 0.000 0.000*** -0.000 -0.000 -0.000 -0.000***
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Per capita income squared -0.000** -0.000 -0.000*** 0.000 0.000 0.000 0.000
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Forestland share -0.039 -0.086 -0.574 -0.257* -0.234* -0.273** 1.280**
(0.294) (0.295) (0.380) (0.129) (0.119) (0.134) (0.542)
Rainfall: average 0.001 0.001 -0.001 -0.000 -0.001 0.001
(0.001) (0.001) (0.000) (0.000) (0.000) (0.002)
Rainfall: variation 0.000 0.000 0.000 -0.000 0.000 0.000
(0.000) (0.000) (0.000) (0.000) (0.000) (0.001)
GDD 0.000 -0.000 -0.000*** -0.000*** -0.000*** -0.000
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
HDD 0.003* 0.002 0.001 0.000 0.001* -0.004
(0.002) (0.002) (0.001) (0.001) (0.001) (0.003)
Population share with telephone 0.014 0.011** 0.342***
(0.013) (0.005) (0.068)
Year trends NO YES NO NO YES NO NO
Observations 132 132 132 132 132 132 132
R-squared 0.574 0.588 0.380 0.445 0.547 0.444 0.540
IV validity (LM -statistic) 26.15 22.77 25.49
Sargan statistic P -value 0.2518 / /
Notes: Instrumented: Proportion of forestland devolved to household management. Excluded instruments: population share possessing telephones,
for both forest cover and EVI estimation (Column 3 and 6); average and deviation of precipitation, effective and harmful cumulative heat variables
(GDD and HDD) are instrumented only in Column 3. Results of the first stage are reported in the last column.
The LM -statistic of IV(s) rejects the null hypothesis that excluded instruments are orthogonal, implying the instruments are relevant. The
overidentification test on multiple instruments in Column 3 (Sargan statistic) cannot reject the joint null that all instruments are valid, i.e.,
uncorrelated with the error term, and that the excluded instruments are correctly excluded from the estimated equation.
Standard errors are in parentheses, bootstrapped with 100 replications. Significance is denoted: *** p<0.01, ** p<0.05, * p<0.1.
37
Chapter II

Allocative Efficiency or Agglomeration?
The Emergence of Forestland Rental Markets and the Forest
Devolution Reform in China*
Yuanyuan Yi„1
1Department of Economics, University of Gothenburg, Sweden
Abstract
This paper evaluates whether the devolution reform of forestland to household man-
agement had an effect on allocative efficiency and household welfare through partici-
pation in forestland rental markets. Using a household panel dataset from three Chi-
nese provinces, I find that the emerging forestland rental markets improved allocative
efficiency in terms of factor equalization. With the reform, forestland is transferred to
forestland-constrained and labor-rich households, and to households with higher levels
of productivity in forestry. I do not find any support for agglomeration of forestland to
land-richer, wealthier, larger or powerful households. Participation in forestland rental
markets increases household per-capita income and decreases the likelihood of having an
income below the poverty line.
JEL Classification: D61, O12, Q15, Q23
Keywords: allocative efficiency, household welfare, forestland rental markets, collective
forest tenure reform, China
*I am grateful to Yonas Alem, Fredrik Carlsson, H˚akan Eggert, Wolfgang Habla, Elina Lampi, Mitesh
Kataria, Gunnar Ko¨hlin, Elizabeth Robinson, Thomas Sterner, Jintao Xu, and seminar participants at the
University of Gothenburg for valuable comments and suggestions. I also acknowledge generous financial support
from the following organizations: the Ford Foundation, the National Science Foundation of China (NSFC),
State Forestry Administration of China, Richard C Malmstens Foundation, Swedish International Development
Cooperation Agency (Sida) at the Environmental Economics Unit at the University of Gothenburg and the
World Bank PROFOR project.
„Vasagatan 1, SE-405 30 Gothenburg, Sweden, Email: yuanyuan.yi@economics.gu.se
1 Introduction
Land reform, by means of changing land tenure arrangements, is an institutional in-
strument for reallocation of factors of production to and among farm households. Prop-
erty (tenure) rights are an important precondition for economic growth (North, 1981;
Barro, 1991; Besley, 1995). In China, the Household Responsibility System (HRS) in
1978 successfully transformed collectively owned agricultural land to household tenure.
Following that reform, agricultural production increased by 225 percent in six years and
rural household income increased five-fold in real terms by 2000 (NBS, China, 2014).
Following the success of the HRS reform, in the early 2000s the Collective Forest Tenure
Reform started to formally devolve forestland owned by village collectives to individual
households (hereinafter, the Reform). As in earlier land reforms, the Reform applied
an egalitarian rule to reallocate forestland on a per capita basis, leading to concerns
that fragmented and small operational sizes of forest plots might result (Wen, 1989).
As an attempt to address potential problems in fragmentation of land, the Reform en-
courages forestland rental markets by giving formal transfer rights to the households to
which forestland was devolved, with the aims of sustainable management of collective
forests and growth of forestry productivity and household income.1 The reform also al-
located more forestland to household management and improved tenure security. The
combination of increased land resources under private management, secure tenure, and
transfer rights, in the presence of heterogeneity in managerial ability, allows households
to maximize the value of forestland by providing both willing out-renters and in-renters
for land markets (Zimmerman and Carter, 1999).
In light of Besley (1995)’s seminal work on the link between investment and property
rights, a land owner with property rights is incentivized to invest more when given
transfer rights, which makes it easier to rent (or sell) land. Earlier in Chapter I of this
thesis, I investigated the impact on investment when households obtained tenure rights
to forestland. The objectives of this chapter are to explore whether the Reform also
encouraged forestland rental markets to emerge and improved the allocative efficiency
of forestland; and to assess the impact of renting on household welfare. More productive
farmers are expected to rent in forestland, and less productive farmers to rent it out.
In case of an equal level of managerial ability, households with a larger area of forest
are expected to be willing to rent out, and those in the opposite situation to rent
in. In addition, farmers may use forestland as collateral, making it easier to obtain
credit for investment purposes. Opportunity costs are important to rural households;
those who rent out forestland due to relatively lower ability in forest management and
1 The overall objectives were stated in the State Council’s announcement of “The Promotion of the Collec-
tive Forest Tenure Reform” on June 8, 2008. (http://www.gov.cn/jrzg/2008-07/14/content_1044403.htm)
1
higher ability in off-farm jobs would be able to obtain more income through engaging
in off-farm work. Hence, participating in the rental market could improve household
welfare and meanwhile, facilitate rural structural transformation by shifting rural labor
to nonagricultural sectors.
This is the first study that uses empirical data to assess the performance of forest-
land markets in a context where forest management is individualized. The Reform
involved two-thirds of the total forestland area of China (i.e., 62 million hectares of
the total 100 million hectares) and 70 million rural households, though forestland mar-
kets are thin (Xu et al., 2010; Siikama¨ki et al., 2015). In the literature on agricultural
land use and rental markets, small operational size is likely to affect investment incen-
tives and sustainable management (Rozelle et al., 2002). Inefficiency in land markets
concerns both policy makers and development economists, in that small farmers and
land-scarce or landless workers tend to be rationed out, so that land is agglomerated
into the hands of richer and more powerful people, or those who already are large
landowners (Otsuka, 2007). In addition, rapid growth of off-farm industries is fostering
rural-urban labor migration, which calls for a flexible system for facilitating factor and
resource reallocation. Farm households in places like rural China could benefit from a
land rental system, not only through optimizing factors allocation, but also by holding
forestland as a safety-net asset. Forestland is different from farmland in that standing
trees store wealth, require less intensive labor, and are subject to a greater extent of
uncertainty and production risk because forests growth takes a longer time than crops
on farmland. As a unique model of forest devolution – private management accom-
panied by forestland markets – the lessons of China will provide options for effective
common-pool resources management in other contexts.
The analysis uses a panel dataset of 1,192 randomly surveyed households in three
Chinese provinces: Fujian, Jiangxi, and Yunnan. These provinces were the earliest
adopters of the forest devolution reform, and data is available as a panel with two
waves. The data include households’ farming and non-farming work and income, in
particular forestland management activities in 2005 and 2010; villages’ socio-economic
characteristics, resources, and natural conditions; and whether, when and how the Re-
form was implemented in 2000, 2005 and 2010. Controlling for the fact that village
characteristics correlate with the adoption of the forest devolution reform, I exam-
ine the determinants of forestland rental participation, i.e., how a household chooses
among rent-in, rent-out, and no-rent. I focus on two factors: a household’s endowment
of forestland per unit of labor and managerial ability in forestry. The managerial abil-
ity is obtained as a composite residual term from a Cobb-Douglas forestry production
function for each household. A multinomial logit model is applied on the determinants
of forestland rental participation. I then analyze the welfare effects of rental partici-
2
pation in terms of income and poverty by propensity score matching (PSM) methods
because the renting decision is not random.
I find positive effects of forestland rental market participation on household factor
equalization, income and poverty reduction, suggesting efficiency and household welfare
improvement. Forestland is transferred from land-rich, labor-constrained households to
households with relatively abundant labor and less forestland and to more productive
households. Little evidence is found for agglomeration in forestland to richer house-
holds, households that already have large landholdings, or more politically powerful
households. Consequently, participation in forestland rental markets increases house-
hold per-capita income and reduces the likelihood of absolute poverty. The results are
highly consistent with alternative estimation procedures (multinomial logit and tobit
models and PSM with a number of matching algorithms).
The rest of this article is structured as follows. Section 2 reviews a selection of the
literature related to allocative efficiency, rural land institutions, and land rental/sales
markets. Section 3 introduces the background and emerging forestland rental markets
in China. Section 4 discusses a conceptual model, followed by four testable hypotheses
and the associated methodological approach in Section 5. Section 6 describes the data,
and the next section presents the estimation results. Finally, I conclude with policy
implications in Section 8.
2 Studies on Agricultural Land Rental Markets and Efficiency
This chapter focuses on the performance of forestland rental markets with regard to al-
locative efficiency and rural household welfare. Standard economic theory started with
very strong assumptions – an identical, constant-returns-to-scale production function
for all households, with clearly defined and perfectly functioning markets – says that
the initial resource endowment should not be important for allocative efficiency (Bliss
and Stern, 1982; Feder, 1985; Bardhan and Urdy, 1999; Deininger 2003). Under these
assumptions, managerial ability does not differ, land-labor ratios will be efficiently al-
located across farm households, and output per factor unit will be optimized. However,
these assumptions are often not fulfilled in reality, and, particularly in developing coun-
tries, the outcome of a rental market could be that land or other forms of resources are
prevalently captured by a small group of people. Moreover, parts of the markets might
be imperfect or missing, due to the presence of transaction costs such as lack of access
to credit.
Imperfect land rental markets and transaction costs have long been recognized in
the literature. In addition to unclear ownership, one main source of transaction costs
3
arises from asymmetric information between renting partners in the process of making
transactions. Such costs might include searching, bargaining, enforcing bargains, and
monitoring tenants (Alston et al., 1984; Otsuka and Hayami, 1988; Carter and Yao,
2002; Holden et al., 2009; Kimura et al., 2011). Another type of transaction costs
include varying degrees of regulation on land transfer rights, which result in missing and
imperfect labor, credit and insurance, and land markets (Skoufias, 1995; Deininger and
Jin, 2005; Pender and Fafchamps, 2006; Deininger et al. 2008).2 The overall conclusion
is that these costs cause the discrepancy between optimal and actual transactions,
and the latter systematically agglomerate land to wealthier, land-rich, or powerful
households (Deininger and Jin, 2008).
Empirical evidence on the performance of rental markets is mainly about farmland,
with a considerable variation in geographical and cultural conditions, and with mixed
findings. For example, the following studies found evidence for inefficient farmland
markets: Andre and Platteau (1998) in Rwanda, Zimmerman and Carter (1999) in
Burkina Faso, and Deininger et al. (2011) and Holden et al. (2011) in Ethiopia. The
others, such as Migot-Adholla et al. (1994) in Ghana, Gavian and Fafchamps (1996)
in Niger, Yamano et al. (2009) and Jin and Jayne (2013) in Kenya, Lunduka et al.
(2009) in Malawi, Deininger and Mpuga (2009) and Nkonya et al. (2009) in Uganda,
Deininger et al. (2008) in India, Deininger and Jin (2008) in Vietnam, and Carter and
Yao (2002), Deininger and Jin (2005), and Kimura et al. (2011) in China, found that
farmland markets equilibrate factor ratios among households. A key factor determining
rental participation is, according to many studies, transaction costs.
Forestland markets differ from farmland markets in a number of aspects. First
of all, the nature of forests means longer-term uncertainty, and the public benefits
provided by standing trees mean that forest plots often are undervalued. Moreover,
relatively larger land size, compared to farmland, requires a considerable amount of
rental payment. Because of these two facts, cash constraints and credit access would be
associated with imperfect forestland rental markets, raising the costs of rental contracts
of forestland and thus reducing households’ willingness to rent. On the other hand,
forestland functions as a safety-net asset, with standing trees which store wealth with
merely a small amount of labor, so that renting-out makes it easier for rural labor to
engage in off-farm work while still enjoying the benefits of owning the forestland. Given
all these features and the large scale of the Reform3, this study adds to the literature
2 Pender and Fafchamps (2006) find that transaction costs do not necessarily affect rental market partici-
pation and efficiency in a negative way in Ethiopia.
3 In Chapter I of the thesis, a review was presented in Section 3. According to government statistics, the
reform had devolved 62 million hectares of the 100 million hectares of forests from village-owned to household
tenure, involving 600 million people in rural areas in more than twenty provinces of China (Xu et al., 2010;
State Forestry Administration, 2011).
4
as one of the first empirical evaluations of forestland markets in terms of determinants
of participation and consequences on household welfare.
3 China’s Forest Devolution Reform and Emerging Forestland
Rental Markets
China’s forestland tenure system started to reform as early as 1954. However, it under-
went privatization and collectivization, revoking privatization and then redistributing
forestland through contracts to households in the late 1980s and 1990s (Xu et al.,
2010). By the end of the 1990s, all the forestland in southern collective forest areas was
legally owned by village collectives. The success of the agricultural HRS reform in the
late 1980s provided lessons for improving agricultural productivity and the welfare of
the rural population, and for the introduction of devolution of forestland management
to rural households. Eventually, since the beginning of the 2000s, village committees,
on the basis of majority voting in village assemblies or representative meetings, decided
on whether to implement the reform.4 In mid-2003, a state council resolution (Docu-
ment No. 9, The Resolution on the Development of Forestry) was issued, formally calling
for deeper reform in the collective and state forest sectors.
In villages adopting the Reform, the village committees decided on whether and
how much of the village-owned forestland to devolve to household management by
means of administrative reallocation. Measures were taken to strengthen and secure
tenure rights. These included formally documenting rights in production, transferability
and use as collateral, and issuance of forestland certificates. Besides forest devolution,
together with local institutional support on land value assessment, for example, efforts
have been directed toward reducing transaction costs in the forestland transfer market,
with the hope of encouraging rental transactions.
Land reforms by the Chinese government are based on egalitarian rules for equity
purposes. The administrative reallocation of forestland is not an exception. Land was
thus distributed amongst village members on a per capita basis. Earlier studies have
shown that this type of reallocation results in land fragmentation and small house-
hold operational size (e.g., Wen, 1989). Small-scale management dis-incentivizes land
investment because of the low level of expected returns. Fragmentation hampers ac-
quisition of timely information about markets because of the large cost of acquiring
4 The village committee is the lowest administrative organization in China, elected by member households
in both administrative and natural villages. (Natural villages mean villages in place for a long period as
village clusters, which trace their history back in time to at least before 1949. Administrative villages refer to
a structural integration of neighboring villages during the early Mao years, and comprise two or more natural
villages.) Village committee membership ranges from 3 to 7 people, as required by China’s Organic Law of
Villagers Committees.
5
information relative to the small quantity of goods produced on a small plot. Therefore,
the egalitarianism in land (re)allocation has been subject to problems of inefficiency
and high transaction costs. In addition, attributed to concentrated power in the hands
of village leaders, it may create rent-seeking opportunities and preferences for friends
and relatives (Rozelle et al., 2002; Deininger and Jin, 2005). Moreover, forest fragmen-
tation concerns enviornmental scientists due to potential negative impacts on forest
quality such as biodiversity (Hill et al., 2011). With an aim to address fragmentation
concerns, the Reform was designed to also encourage forestland holders to participate
voluntarily in the land rental market. The main purpose in this paper is to explore
how the recipients of devolved rights to forestland respond in reallocating forestland
among themselves.
The rural population in collective forest areas lives on agricultural, forestry, and
livestock production, as well as off-farm income. Their heterogeneous skills in (forest)
land management and off-farm occupations call for a flexible system that optimizes the
allocation of resources and factors. In China, the rapid growth of off-farm industries has
induced rural-to-urban labor migration, resulting in a need to reallocate land resources
among rural households (Deininger et al., 2014). Well-functioning land rental markets
bring land to more productive users, and allow owners working off-farm to continue
enjoying the benefits of owning land (Otsuka and Hayami, 1988; Jin and Deininger,
2009).
Agricultural land rental transactions increased nationwide following the HRS re-
form, with 2.3 percent of the total number of households renting out in 1992. By 2013,
the total rented area of agricultural land accounted for 26 percent of agricultural land
in the nation. Still, compared to other developing countries, this number is very low
(Otsuka, 2007; Deininger and Jin, 2005). In the meantime, along with the recent forest
devolution reform, forestland rental transactions are emerging among households. The
three provinces studied in this paper – Fujian, Jiangxi and Yunnan – were among the
earliest adopting the Reform. In these provinces, the percentage of households engaged
in the forestland rental market increased from 4 percent in 2005 to 12 percent in 2010,
as shown in Table 1.
< Table 1 here >
Forestland rental participation increased over the period: the percentage of house-
holds renting in forestland increased from 2.1 to more than 7 in 2005-2010; the more
than threefold increase mainly comes from households that were affected by the Reform
during the study period. The percentage of households renting out forestland increased
from 2.9 to 5.5 in the same period. Pre-rental, households affected by the Reform had a
larger mean area of forestland compared to non-Reform households. Among households
6
not subject to the Reform, 4.4 percent had informal forestland transfers in 2005, im-
plying lax regulation in some villages. When villages implemented the Reform, former
informal rental contracts for forestland, if still valid, were presented to the leaseholders
and tenants for their re-assessment of whether to rent again or keep the forestland for
own use. So, the 4.42 percent of the non-reform households who rented out in 2005
do not necessarily appear among the 5.58 percent of the rent-out households with the
Reform in 2010. In addition, compared to the non-reformed households, the reformed
households had a larger area of forestland rented-in, and their rented-in forest plots
accounted for a larger proportion of the endowment level of forestland.
4 A Conceptual Model of Forestland Rental Market Partici-
pation
A representative household in the study area is endowed with some size of forestland
area (A¯), family labor (L), managerial ability (α), household characteristics such as
risk preferences (ρ), and credit and cash constraints (C). Other household characteris-
tics (H) include age and years of education of the household head, who is usually the
decision-maker in a rural household, as well as the household head’s professional knowl-
edge of forestry. In household forestry, the optimal size of the operational forestland
area, A∗, is a function of the household’s forestry managerial ability (α), its character-
istics (A¯,L,ρ,C,H), and village-level exogenous factors (V ), such as natural and social
endowments, market access, precipitation and temperature. The household’s utility is
maximized when the difference between the household’s actual and desired land size is
minimized.
This utility maximization can be achieved through rental market participation,
taking transaction costs into account. When a forestland rental market exists but there
are transaction costs (T ), the household decides to participate in the rental market
based on the rental payment and opportunity cost of rental participation (R), e.g.,
potential income from managing cropland, off-farm work, and livestock. That is, the
household finds itself in one of the three rental states:
Rent-out (A∗ < A¯) if MB(A¯) + εi ≤ R− T out;
No-rent (A∗ = A¯) if R− T out < MB(A¯) + εi < R+ T in;
Rent-in (A∗ > A¯) if MB(A¯) + εi ≥ R+ T in.
(1)
where MB(A¯) + εi stands for the marginal benefit or marginal value of the forest and
non-forest production of managing an additional area unit of forestland. The marginal
7
returns are evaluated at the level of forestland endowment, and depend on α,ρ,C,R,T,H
and V. εi is an error term containing non-observable factors and is assumed to follow
a standard normal distribution.
One type of transaction cost arises from imperfect factor markets, resulting in credit
constraints for some households. Firstly, the lack of finance for rental payments, and,
secondly, the possibility that wealthier farmers will have the opportunity to monopolize
land, greatly impede market transactions in rural areas of developing countries, as in
Skoufias (1995), Rozelle et al. (2002), Deininger and Jin (2005), and Deininger et al.
(2008). Rozelle et al. (2002), Deininger and Jin (2005) and Deininger et al. (2008)
also found that varying degrees of regulation on land transfer rights result in high
transaction costs that reduce the efficiency of land rental markets. Deininger and Jin
(2005) developed a model for transaction costs affecting rental decisions, so I follow
them to denote T in and T out as the transaction costs associated with rent-in and rent-
out transactions. In the study area, the rights to rent forestland to any other villagers
and to any outsiders, the right to use a forest plot as collateral, and the right to change
forest type, i.e., among timber, fruit, bamboo, and other types of trees, vary across
villages.
In the forest tenure reform in China, a household’s ownership of one specific forest
plot is guaranteed for 30 to 70 years, and in some cases even longer (“long term” or
“forever”). The length of ownership is clearly specified in the forestland certificate for
each forest plot. Thus, a rental contract for any forest plot can be as long as the length
of the ownership. Besides the security brought by a certificate, the household’s attitude
towards production risk and future ownership risk, reflected by the perception of risk
of losing the plot, factors into the household’s willingness to participate in the rental
market. In addition, a cash-constrained household, with cropland size too small to meet
its own consumption needs, may rent out some forestland to obtain some cash.
5 Testable Hypotheses and Empirical Strategy
According to the conceptual model, a rural household’s decision to rent in, rent out, or
do neither, and the implications of its decisions for allocative efficiency, are driven by
factor endowment and managerial ability in forestry, ceteris paribus. To improve the un-
derstanding of the Reform’s impact on forestland rental markets and the determinants
and welfare effects of household participation, I am interested in testing four hypothe-
ses, as follows. I take three steps in the empirical analysis. The identification strategies
are explained later in this section, taking into account the factors of ownership and
tenure security, policy restrictions on transfer rights, and market access.
8
Hypothesis 1 . The forest devolution reform improves allocative efficiency through forest-
land rental markets, which contribute to the equalization of forestland-labor ratios among rural
households. In other words, forestland is transferred from households with less labor and
more forestland to households with more labor and less forestland, holding other factors
constant.5
Hypothesis 2 . Forestland rental markets improve allocative efficiency by transferring
forestland from less productive to more productive households.
Hypothesis 3 . Rental market participation has a positive income effect on rent-in house-
holds through higher production value in forestry, and on rent-out households through more
income from off-farm work.
Hypothesis 4 . Rental market participation has a positive impact on poverty alleviation.
That is, participating in forestland rental market reduces the likelihood of household
per capita income falling below the absolute poverty line.
5.1 Household Forestry Production and Managerial Ability
First, I estimate a measure of a household’s managerial ability in forestry, which is not
directly observable. To obtain the measure, for each household, I compute an annual
incremental value of household forestry production – the value of marginal benefit of
leaving the trees to grow for an extra year – from standing timber, economic forests
(including fruit trees, herbal medicine and seedling nursery plantations), or bamboo
forests on each forest plot,6 and then aggregate a Cobb-Douglas forestry incremental
5 A focus on equalization of factor ratios is warranted if the quality of factor inputs are homogenous or at
least similar. While this is commonly assumed, I try to keep closer to reality by controlling for households’
different quality of labor (managerial skill). This means that equalization of forestland-labor ratios is only
an indication of allocative efficiency. In agriculture it is therefore more common to analyze changes in actual
productivity. But in my case it is too early after the reform to analyze that (particularly for timber outputs).
Given earlier research and an assumption of not too different within-household managerial skill in land and
labor (in agricultural production, e.g., Gavian and Fafchamps, 1996; Deininger and Jin, 2005), I therefore
focus on forestland-labor ratios.
6 This value of marginal benefit of leaving the standing trees to grow for an extra year is computed in
a manner similar to the decision to invest the stumpage for an interest return: for timber forest, the self-
estimated standing timber volume is multiplied by the village-reported timber price and annual interest rate;
for economic forests – fruit trees, herbal plantations, etc. – it is the annual production value; and for bamboo
forest, the value is a summation of annual production value of bamboo shoots and the bamboo fiber volume
times output price and interest rate. I acknowledge that this value function differs from the forest valuation
function in which the (discounted) present value of forest rent is maximized based on a decision of optimal
rotation in the long run. One major issue is the unavailability of precise data from the surveys on optimal
(or household perceived optimal) rotation years for all the tree species. So the current measure for forest plot
productivity, by using the annual incremental value, is not able to evaluate a longer-term impact of the forest
tenure reform on forest (forestry) productivity. In addition, it is still too early for us to assess that, due to
the lack of knowledge about the recently planted trees – the rotation decisions and long-run uncertainty, for
example. So this paper is devoted to the short-term effects on the responses of forestland owners, how they
adapt to improved tenure and transfer rights in small-scale forest management, and whether their responses
would imply more efficient forest production overall.
9
value function for a household i in village j at time t, as in Equation (2).
logyijt = γ1logLijt +γ2logAijt +γ3logKijt +γ4Xijt +ηt +ϑj +ϑjt +µi + ijt. (2)
where yijt, Lijt, Aijt, and Kijt are, respectively, annual incremental value of household
forestry production, family labor, total forestland area, and value of inputs. Xijt is a
set of household control variables (ρ,C,R,H) that could potentially affect households’
forestry production. Time trend (ηt), province and village fixed effects (ϑj) and village
year effects (ϑjt) are controlled for. ϑjt is expected to also capture a few particular
events. First, some provinces and villages were affected by the 2008 global financial
crisis, and these effects might have lasted until 2010. Second, in the year when villages
implemented the Reform, there might have been some immediate effects encouraging
efforts in production.
Equation (2) is estimated by a household-level fixed effects linear regression (FE)
model on a two-period panel dataset. I obtain the measure of forestry managerial abil-
ity, αi, as the composite term of the time-invariant constant (µi) and the stochastic
error (ijt): αˆi = µˆi + ˆijt. Including the latter takes into account the randomness of,
for example, weather and social influences in the productivity of a piece of land with
standing trees, which is a feature of the measurement of total factor forestry produc-
tivity. Therefore, the measure of αi differs from several previous studies of agricultural
land that use only the household fixed effect µi (Deininger and Mpuga, 2009; Jin and
Deininger, 2009; Jin and Jayne, 2013; Chamberlin and Ricker-Gilbert, 2016).
5.2 Determinants of Forestland Rental Market Participation
In the next step, I use the derived estimate of total factor forestry productivity, αˆ,
as an explanatory variable in a model of the determinants of forestland rental market
participation in Equation (3):
τijt = δ1αˆi + δ2(AL )ijt + δ3FTRjt + δ4Tijt + δ5Xijt + δ6Vjt + εijt. (3)
τijt < 0 when A∗ < A¯ (so the household rents out forestland), τijt > 0 when A∗ > A¯
(so the household rents in forestland), and τijt = 0 when A∗ = A¯ (so the household
does not rent). I will treat τijt in the following two ways. First, I use a categorical
response of τijt ∈ {0, 1, 2} to indicate {Rent out, No rent, Rent in}, with the category
0 representing a case when A∗ < A¯, category 1 representing a case when A∗ = A¯,
and category 2 representing a case when A∗ > A¯. Second, τijt also can be continuous,
meaning the difference between the desired and the endowed forestland, A∗− A¯, which
is the amount of rented area. The three variables – the household’s managerial ability
10
in forestry (αˆi), the pre-rental area of forestland per unit of labor ((AL )ijt), and the
length of time with the Reform (FTRjt) – allow me to test Hypotheses 1 and 2 by the
signs and marginal effects of the coefficients δ1∼3.
FTRjt estimates the Reform’s effect in encouraging rental transactions. An imme-
diate effect is expected to be an increase in (AL )ijt once a village decided to reform
and reallocate some its forestland. Furthermore, forestland rental participation might
increase over time with the Reform.7 To identify the effect of the Reform, potential
endogeneity has to be addressed, because the reform decision, driven by village natural
endowments and socio-economic characteristics, may also correlate with households’
willingness to rent forestland. In particular, the villages with more forestland involved
in informal transfers may be more likely to reform. I define FTRjt = zjξ′+υjt, where
zj includes a vector of village characteristics before the Reform (zj0) and in the current
period (zjt). εijt = βυjt + eijt leads to a biased estimate of the Reform’s effect on
forestland rental participation, i.e., δ3 in Equation (3). I use the control function (CF)
approach to correct the selection bias. Then, Equation (3) is adjusted to:
τijt = δ1αˆi + δ2(AL )ijt + δ3FTRjt + δ4Tijt + δ5Xijt + δ6Vjt + βυˆjt + eijt. (4)
and υˆjt = ˆFTRjt − zj ξˆ′ as a first-stage step to estimate Equation (4). The two
main advantages of using the CF approach are that (1) the null of the endogeneity
of FTRjt can be tested simply by the statistical significance of the parameter β and
(2) the average partial effects are easy to compute when Equation (4) is estimated by
a nonlinear (and multinomial response) model (Petrin and Train, 2006; Wooldridge,
2010).
In interpreting αˆi and its parameter estimate δ1, I follow several other studies such
as Jin and Jayne (2013), and acknowledge that αˆi also captures some other things that
do not vary with time, for example, soil quality. Soil quality of plots with standing
trees is believed to be even more stable, that is, time-invariant, than that of plots with
crops. Given the assumption that a household is more likely to rent out (forest) land
with lower soil quality than to maintain such land for its own use, δˆ1 is more likely to
be downward biased than to over-estimate the effect of productivity on renting in or
out forest plots.
Xijt is the same set of household-level factors as in Equation (2), including percep-
tions of risk (ρ), cash constraints and credit access and ability to obtain credit (C),
opportunity costs of managing more forest (R) (reflected by cropland size, income
7 For example, it takes time to arrange forestland transfers; to obtain professional services, such as the
assessment of the value of land with standing trees, which is important for establishing market rates; and to
develop information platforms. Further, owners’ confidence in tenure security may take time to develop.
11
from off-farm work, and livestock production value), and other household characteris-
tics (H): household head’s gender, age, education, political connections, and experience
working in a forestry department.
Tijt is a vector of variables representing transaction costs affecting rental decisions.
Two examples in the vector include holding a forestland certificate, which guarantees
the security of tenure rights, and the perceived rights to change forest types, to transfer
forestland to other villagers and outsiders, and to mortgage forestland, which are based
on policy regulations. More forestland transactions at an earlier time may imply better-
functioning markets and lower transaction costs. Other village-level characteristics in
Vjt include demographic and resource endowment, and distances to cement-paved road,
to closest county center, and to local bank. Precipitation and temperature are controlled
for because they are important for forest growth.
Equation (4) is estimated by multinomial logit regressions on farmers’ choices among
rent-in, rent-out, and no-rent, using the same two-period panel data. I also unravel the
determinants of the extent of rental participation by random effects tobit models,
where the dependent variables are continuous – the amount of area and the percentage
of forestland rented-out (rented-in) compared to the pre-rental level. Tobit estimators
allow us to account for the corner solution for a large number of households that do
not participate in the forestland rental market. The results of the binary response of
rent-out (rent-in) estimated by probit models are similar to the results of multinomial
logit models and are not reported in the paper. Standard errors are bootstrapped for
models, including the two estimates αˆi and υˆjt that are generated from the two-step
estimation process.
5.3 Welfare Impacts of Forestland Rental Market Participation
The third step of the empirical analysis is to quantify the impact of rental partici-
pation on household welfare, in terms of income and income distribution, in order to
test Hypotheses 3 and 4. I estimate a set of models on the average treatment effect
of rental participation on a set of welfare indicators: i) a household’s per capita in-
come; ii) household’s income share from off-farm work; iii) household’s income share
from forestry; and iv) probability of income falling below the poverty line. The treat-
ment, a household’s renting-in (renting-out) forestland (Di in Equations (5) and (6)) is
not randomly assigned. Some factors driving households to self-select to renting may
also correlate with their welfare outcomes. Propensity score matching (PSM) methods
correct the self-selection bias so as to identify the causal impact of renting, based on
the observable pre-treatment differences between the treated and the control groups. I
elaborate the conceptual foundation and procedure of the PSM method as follows.
12
Following Becerril and Abdulai (2010), a random utility framework is used to model
the difference between the utility from renting (in or out) (U1) and the utility from
not renting (U0), defined by a latent variable, D∗i . Di is observed as households’ actual
forestland rental behavior given a set of observable household and forestland charac-
teristics (Zi) according to the decision function (5):
D∗i = g(Zi) + ui, Di =


1 if D∗i = U1 − U0 ≥ 0
0 if D∗i = U1 − U0 < 0
. (5)
BecauseDi is by self-selection instead of random assignment, I use the following welfare
Equation (6) to illustrate how PSM is carried out to identify the average renting effects:
Ii = p(Zi) + θDi + σi. (6)
where Ii refers to one of the aforementioned welfare indicators. p(Zi) = Pr[Di = 1 |
Zi] = F{h(Zi)} and F{·} can be normal or logistic cumulative distribution; σi is
a normal error term. As suggested by Rosenbaum and Rubin (1983), counterfactual
situations for those who rented (i.e., the treated, ifDi = 1) will be statistically matched
with those who did not rent (i.e., the control,Di = 0) and who had the same probability
of assignment to the treatment, p(Zi). Conditional on the propensity score, p(Zi), the
coefficient of rental participation θˆ, will be unbiased and estimated as the average
treatment effect on the treated (ATT). The average welfare effect of forestland rental
participation is given by Equation (7):
θ = ATT = E{E[Ii1 | Di = 1, p(Zi)]− E[Ii0 | Di = 0, p(Zi)] | Di = 1}. (7)
The estimation of the treatment effects using propensity score matching requires two
assumptions. First, the Conditional Independence Assumption (CIA) assumes that, for
a given set of covariates, participation is independent of potential outcomes. Second is
the condition that the ATT is only defined within the region of common support. The
propensity scores, p(Zi) are conditional on a vector of observable pre-treatment char-
acteristics, suggesting that the selection of the Zi covariates is crucial for the matching.
My selection of Zi includes all the observed covariates determining a household’s forest-
land rental status, as used in the estimation of Equation (4).
The last step is to match the “neighbors” of a rent-in (or rent-out) household with
a propensity score very close to the propensity scores of the no-rent households. I
apply the most commonly used matching methods: the nearest neighbor (NNM) and
the kernel-based (KBM) matching. The NNM and KBM differ in that the former
13
matches each treated observation with the control observation that has the closest
propensity score. The latter matches all treated observations with a weighted average
of all controls. The weights in KBM are inversely proportional to the distance between
the propensity scores of treated and control groups. Then, the average of the differences
of each matched pair is computed as the ATT (θ).
6 Data
I use a two-period panel dataset from over 1,000 randomly surveyed households in
three Chinese provinces: Fujian, Jiangxi, and Yunnan. The survey team, from the
Environmental Economics Program in China at Peking University, collected household
data on demographic and social-economic characteristics, labor allocation, agricultural
production, forest management activities at plot level, income and assets, basic social
relationships, and (forest) land use practices and rights, in the years 2005 and 2010.
The village-level data is for three years (2000, 2005 and 2010) and covers natural and
resource endowments, socio-economic characteristics, and whether, when, and how the
Reform was implemented in the village. The survey is nationally representative, in
eight provinces with collective forests from the south to north of China. In 2006, 10-20
households were selected in one village in each county following a stratified random
sampling rule, and were revisited in 2011. The three provinces are included in this
study because the surveys in the other five did not ask about rent-in and rent-out of
forestland plots in 2005. More importantly, the provinces included were the earliest to
adopt the Reform, and thereby provide a better understanding of the emergence and
evolution of forestland rental markets.
The analysis focuses on the 1,192 and 967 households who were interviewed in
106 villages in the two waves, respectively.8 This gives an attrition rate of 19 percent.
The variables used in the analysis are defined in Tables 2 and 3, for the surveyed
households in reformed and non-reformed villages in both periods and their village
characteristics in 2000-2010. The Reform had been implemented in 58 percent of the
households in 2005 and 98 percent in 2010. As Table 2 shows, sizeable differences exist
between the reformed and non-reformed households in landholdings, tenure rights,
and ownership security. Households in reformed places had a larger average size of
per labor forestland as an endowment than in non-reformed places (18 vs. 12 and
17 vs. 5 mu in 2005 and 2010, respectively). In general, households with the Reform
8 The first wave surveyed 480, 300, and 600 households in Fujian, Jiangxi and Yunnan respectively (1,380
in total). In the second wave, 325, 228, and 528 households were available to be re-surveyed (1,081 in total).
The sample size for the analysis comes to 1,192 and 967 excluding observations with missing and bad quality
data. The reason for attrition is that the respondent was not available. I believe this is not systematic and
will not affect the analysis. My reasoning is discussed below.
14
perceived stronger rights to change forest type and to transfer forestland within and
outside the village. The households receiving forestland certificates in reformed places
increased from 19 to 65 percent during 2005-2010, while the 2005 non-reform households
already perceived that they were secure about the risk of ownership/production. Other
household characteristics will be discussed later in Table 4 with regard to different
rental participation status.
< Table 2 here >
Table 3 reports the village characteristics of social, economic, geographic and weather
conditions during 2000-2010. Among the sampled villages, 4 villages where forestland
accounted for 90 percent of the total land area, decided, in a spontaneous and informal
way, to reform in 2000 and devolved 43 percent of their collective forests to house-
holds. In 2005, when the central government officially announced its encouragement to
proceed with the Reform, 73 villages followed. Villages with better forest endowment,
in terms of larger size and percentage of the village total land area, seem to be the
first-movers of reform. Also, the Reform was more likely to be selected by villages with
smaller population size and fewer households, those with higher average income, and
those facing a higher price of commercial timber. By 2010, the 105 reformed villages
had devolved on average 62 percent of the collective forests to households. The higher
percentage of forestland transfers in reformed villages suggests better rental markets
in villages with the Reform than in those without in 2000 (10 vs. 5 percent), 2005 (8
vs. 2 percent) and 2010 (almost 10 percent vs. 0).
< Table 3 here >
Weather data on rainfall and temperature is obtained from the China Meteorolog-
ical Data Sharing Service System (CMDSSS). The CMDSSS records daily minimum,
maximum and average temperatures, precipitation, and solar radiation in 820 weather
stations (with exact coordinates) distributed across all of China. Using village center
coordinates taken by the survey team with GPS devices, I match each village with
the closest weather station. As shown in the bottom of Table 3, I calculate average
and standard deviation of annual precipitation. Because temperature effects are non-
linear, a higher level of effective cumulative heat (0-35 degree Celsius, ◦C) is good for
plants, whilst the influence by harmful cumulative heat (>35 degree Celsius, ◦C) is
the opposite (Schlenker and Roberts, 2009). The reformed villages had higher average
rainfall and rainfall variation, as well as higher levels of both heat measures, than the
non-reformed at all times.
To illustrate the extent to which the hypotheses on forestland markets are borne
out, Table 4 presents key variables by household’s forestland rental status – rent-in,
15
rent-out, or no-rent – and by reform status and time. On average, in both 2005 and
2010, the rent-in households with the Reform were endowed with a smaller area of
forestland per unit of labor (17.61 and 12.88 mu) and the households who rented out
had a larger endowment (38.15 and 43.34 mu). Among the non-reformed households, the
gap is bigger, with 79.76 mu of forests in rent-out households and only 6 mu of forests
in rent-in households in 2005. In 2010, none of the 18 sampled households in the only
non-reformed village rented any forestland. The comparison between the endowed level
and the currently managed level of forestland per unit of labor among the three groups
suggests that rental participation decreases the gap in landholdings. Furthermore, this
gap is smaller in households with the Reform than in those without. In 2010, the
gap almost disappeared, when each member of the family available for labor managed
about 34 mu of forestland in both rent-in and rent-out households. These facts imply
a direction in which forestland markets transferred land from households richer in per
labor forest-landholdings to those with less forestland per labor unit.
< Table 4 here >
The rent-in households had the highest value of investment in forestry, compared
with the rent-out and no-rent households. As expected, the annual value of all forest
and non-forest products on plots is higher than that of the no-rent or the rent-out
households. Labor work time in off-farm jobs does not differ much by rental status.
Off-farm income was the highest in the rent-out households and the lowest in the
rent-in households in 2005, but in 2010 the rent-in households surpassed those renting
out. The difference in per capita income suggests a higher level of income achieved by
households who participated in forestland rental markets than by the no-renters.
I also observe that forestland is transferred from households with older heads to
those with younger heads. The renters had easier credit access (by the distance to
closest local bank) and stronger ability to obtain credit than the no-renters had.9 The
other variables, such as regulations on forestland use and transfer rights, do not differ
in a meaningful way, except that the rent-in households perceived stronger rights to
mortgage forestland to obtain loans. Note that the rent-in households are never the
richest, nor are the rent-out households the poorest, based on values of house and
livestock, for example. Further, the household head’s Communist Party membership
and village committee service do not indicate a consistent difference among the three
rental states. So, forestland agglomeration into the hands of wealthy and powerful
people should not be a major concern based on the descriptive evidence so far. The
9 Ability to obtain credit is measured by the answer to whether a surveyed household could successfully
borrow 500 CNY within one week’s time. 500 CNY is roughly the wage income of two weeks of off-farm work
in rural areas in the three provinces.
16
next section exploits econometric methods to control for endogeneity and other factors
affecting rental decisions, and examines the agglomeration or the allocative efficiency
effect in household forestry via rental markets.
Potential Attrition Bias
The second wave survey lost 225 observations (19 percent attrition) because of house-
hold heads’ temporary absences for business, sickness, or migration. Concerns exist
about potential attrition bias if the reasons that households were not re-surveyed are
non-random exist with respect to their forestland rental market participation. To cor-
rect the bias, I use the inverse probability weighting (IPW) method introduced in
Wooldridge (2010, p. 837). I first estimate a household’s probability of being re-surveyed
on their first-period observable characteristics (Pri,2010 = Φ(Xi,2005)), and then use the
inverse predicted probabilities as weights (IPWi = 1/Pri,2010) for each variable of 2010
in the estimation of Equation (4). The coefficient estimates from the IPWs-adjusted
specifications do not differ much from the basic results, suggesting that attrition bias
should not be a major concern, so I report the main results in which IPWs are not
included.
7 Results
In this section, I present results for three sets of regressions: first, a Cobb-Douglas pro-
duction function of household forestry, via a household FE model, aiming to derive the
measure of forestry productivity; second, the multinomial logit and tobit models on de-
terminants of forestland rental participation, with implications for allocative efficiency
and agglomeration concerns; and, finally, the welfare impacts of renting forestland,
estimated by the PSM method.
7.1 Household Forestry Productivity
Table 5 presents the FE estimates of the Cobb-Douglas production function. Results
indicate that the annual incremental value of household forestry is significantly higher
in households with larger forest landholding, higher value of investment, easier access
to credit (with shorter distance to local bank), farther distance from urban popula-
tion pressure (with longer distance to closest county center), and higher product price
17
(proxied by commercial timber price, due to unavailable prices for all products).10
Lower average and higher variation of rainfall, more days with temperature between 0
and 35 ◦C, and less days warmer than 35 ◦C are statistically significantly related to a
higher level of forestry production.
< Table 5 here >
As discussed previously, I use the time-invariant and the stochastic terms from these
results as a measure of forestry productivity: αˆi = µˆi + ˆijt for each household i. The
distribution of household forestry productivity is plotted in Figure 1. Overall, household
forestry productivity shows a normal distribution in all subsamples of provinces in
each year, but with various degrees of dispersion in different provinces and times.
Active rental markets and better factors mobility are believed to contribute to a more
concentrated normal distribution of managerial ability towards a zero mean; however,
more spikes and larger dispersion of the distribution of productivity would be due to
inactive land rental markets and less off-farm or migration opportunities, which impede
factors mobility. Similar to what Deininger and Jin (2005) found for household farming
ability, a more dispersed productivity distribution is found in Yunnan than in the other
two provinces due to less-developed factor markets.
< Figure 1 here >
7.2 Forestland Rental Market Participation and Allocative Efficiency
Table 6a reports the estimation results on the determinants of forestland rental market
participation (Equation (4)). I apply multinomial logit models with three rental states:
rent-out, no-rent, and rent-in (the reference case is no-rent), with standard errors clus-
tered at the household level. I estimate them firstly on pooled data (in column-set 1)
and then for 2005 and 2010 samples separately (in column-sets 2 and 3). The results
include the two generated regressors – αˆi representing the household forestry produc-
tivity and υˆjt representing the endogeneity of reform selectivity to rental participation,
with p-values based on 200 iterations of bootstrapping the standard errors. Omitting
υˆjt and then also αˆi yields results very similar to other coefficient estimates (not shown
and available upon request), indicating robustness to the addition. The coefficient of
υˆjt is statistically significant, and the model corrects the estimation bias of the reform’s
time effect.
10 The quadratic terms of family labor, forestland area and input value are added to investigate the non-
linearity of scale and investment increments, as well as that of education and price of timber. None of the
quadratic terms has significant coefficients. The coefficient of the quadratic term of endowed forest landholdings
is negative, suggesting the returns to scale could be diminishing. Results are available upon request.
18
< Table 6a here >
The coefficient estimates in Table 6a are consistent with previous literature in agri-
cultural land rental markets regarding the implications for welfare, in terms of allocative
or productive efficiency (e.g., Deininger and Mpuga, 2009; Jin and Deininger, 2009; Jin
and Jayne, 2013). In all specifications, household pre-rental, per-labor forestland has
a positive and statistically significant effect on the likelihood of renting out, and a
negative and significant effect on the likelihood of renting in. Stated otherwise, holding
other factors equal, households with a larger amount of forestland per unit of labor
are more likely to rent out, and those with a smaller amount of forestland per unit of
labor are more likely to rent in. The forestry productivity variable has positive and
statistically significant coefficients in all “rent-in” columns, suggesting that relatively
productive households are more likely to rent in forestland. A longer period with the
Reform in place increases the likelihood of renting in forestland, and, for 2005, increases
the likelihood of renting out. The findings support Hypotheses 1 and 2 that, via forest-
land rental markets, the forest devolution reform improves allocative efficiency. The
other covariates that significantly explain the variation in forestland rental partici-
pation have the expected signs. Their marginal effects are evaluated at the sample
means and presented in Table 6b, on the likelihood of rent-out, no-rent, and rent-in,
respectively.
Holding other factors constant, for a household where per labor forestland is 100
percent higher than the average (of 15 mu), the probability of the household renting
out some forestland will be 2.2 percentage points higher; the likelihood of renting in
will decrease by 1.9 percentage points. Given that the average percentage of renting-
out households is 7 and that of renting-in households is 5.4 (in 2010), the effects are
sizeable – about one-third. Hypothesis 1 is supported by the evidence that rental mar-
kets transfer forestland from the forestland-rich, labor-constrained households to the
labor-rich, forestland-constraint households. This points toward the equalization of
forestland-labor ratios among households. An increase in productivity, from the bot-
tom to the top in the sampled households, is estimated to increase the likelihood to
rent in by 9.5 percentage points. Households with younger (older) heads are more likely
to rent-in (rent-out), and they rent in more forestland, implying a net transfer from
less able to more able households. These findings suggest a productive efficiency effect
through the forestland rental market that transfers forestland from less productive to
more productive households, supporting Hypothesis 2.
< Table 6b here >
To examine the extent of rental participation – the determinants of amount and
19
percentage of rented forestland relative to the endowment level, I employ random effect
tobit models. Results for renting-in and renting-out are presented in Table 7. The
negative (positive) coefficients of the per-labor forest landholding, accompanied by the
positive (negative) coefficients of household forestry productivity in the area rented in
(rented out) are evidence of the allocative efficiency effect of these rental markets.
Little evidence is found on agglomeration in forestland to a specific kind of land-
holders, by either the size of household landholdings, political connections, or wealth.
Political connections such as membership in the Communist Party or village commit-
tee, or work experience in forestry departments, have no statistically significant impact
on rental market participation. An exception is that members of village committees
were less likely to rent forests out in 2005. Households with a higher asset value of,
e.g., houses, are more likely to rent in forests. But this does not necessarily mean ag-
glomeration to the rich, because there is no significant effect suggesting that poorer
households are more likely to rent out.
< Table 7 here >
In addition to forest landholdings and owner productivity, I find that access to
credit and ability to obtain credit affect forestland rental participation, as expected.
Households with a shorter distance to local banks are more likely to participate in
forestland rentals. Softened credit constraints, i.e., ease of borrowing 500 CNY within
a week, increases the likelihood of forest rental transactions (in Tables 6a and 6b), but
there is no significant effect on rental amount (in Table 7). As a tenure security effect,
households’ stronger perception of secure ownership or rights to production is found
to significantly encourage renting in forestland in 2005, while those perceiving a bigger
chance of losing plots were more likely to rent out.
Timber price, as the proxy variable for the price of forest products, indicates the
value of a forest plot on rental markets. A price increase would drive the market
supply up; thus, more rental transactions would be anticipated. As expected, I find
the likelihood and magnitude of renting out, and the area rented in, increasing with
an increase in timber price. Another important factor that would make the expected
returns to land differ is a restriction on forestland use – the right that allows owners
to change forest type amongst timber, bamboo, and economic forests.11 This right
enables owners to rationalize their use in order to maximize the profits from a forest
plot. Controlling for the right does not affect the driving effects of factor endowment
and productivity. Nevertheless, this right seems to have an encouraging effect on renting
in, as shown by the positive signs.
11 It is usually forbidden to change forestland to other uses, such as cropland or for construction. In places
with this right, any restrictions on harvest would be non-binding.
20
A household’s opportunity costs in view of agricultural and livestock production
and off-farm work do not have a statistically significant impact on rental participation.
Note that households with a smaller amount of per capita cropland tend to rent out
some forestland (significantly in 2005), probably due to cash constraints and the need
to fulfill their own consumption of food. As Table 7 shows, households with greater
per-capita cropland are more likely to rent in a larger amount of forestland, indicating
the complementarity of agricultural and forestry production. This is also supported by
the evidence that renting in and renting out are more likely in villages with a larger
share of cropland.
One more year with the Reform is estimated to increase the rent-in likelihood
by 2.7 percentage points, and also the likelihood to rent out in 2005. Apart from
the effect of perceived tenure security, other sources of transaction costs related to
rental markets, such as regulations, earlier transactions and forestland certificates, have
little statistical power. An exception is that the right to mortgage forestland is found
to encourage renting in (for 2005) and a stronger transfer right (the ability to rent
forestland to outsiders) encourages renting out (in 2010). Finally, households in villages
with a larger area of forests are more likely to rent. Larger population and social-
economic development, represented by average income, percentage of population having
telephones, and development of off-farm labor markets, entail high pressures on forests,
and are found to discourage households from participating in rental markets. Weather
conditions, on which forests are very dependent for growth, affect rental participation
in a significant way. Households are more likely to rent out forestland, and to rent
out a greater amount of forestland, in places with lower average rainfall and greater
variation, and are more likely to rent in where rainfall is higher on average but varies
less and where there are more days between 0 and 35 degrees Celsius.
7.3 Impact of Forestland Rental Participation on Household Welfare
Table 8 shows the differences in the main welfare indicators between the no-renters
(Column A), the rent-in households (Column C), the rent-out households (Column
D), and the households rented (Column B). Compared to the no-renters, the rent-in
and rent-out households have higher annual per capita income, ranging from 1,930
to 2,448 CNY, and more than 10 percent lower probability of income falling below
the absolute poverty line. The poverty line is defined as 2000 CNY for 2005 and 2300
CNY for 2010, in accordance with the World Bank’s global poverty standard, i.e., 1.25
Purchasing-Power-Parity (PPP) U.S. dollars a day in 2005. Households participating
in forestland rental markets have more off-farm income than the no-rent households, by
1,766-2,758 CNY a year. The difference in products value of per-mu forestland does not
21
significantly differ by rental status. This suggests that the average forestry productivity
of the incumbent managers is satisfactory. In support of this point of view, I also find
that, compared to the no-renters, the rent-out households have 6 percent less income
from forestry production, and this difference is statistically significant.
< Table 8 here >
In the light of the significant mean differences in welfare indicators, I undertake the
following procedures to identify average treatment effects for each treatment – renting
(including either rent-in or rent-out), rent-in and rent-out separately. The propensity
to be treated, i.e., p(Zi), is estimated on the basis of the observables from the earlier
results on rental market participation.12 In the estimation of a treatment effect, I use
a common support, assuming the treated and the control are distributed in the same
domain. Figure 2 provides a visual inspection of the density of the distribution of
the estimated propensity scores for the treated group and the control group, for each
treatment. The substantial overlap in the distribution of the propensity scores for each
treatment indicates that the common support condition is satisfied.
< Figure 2 here >
Table 9 presents the estimates of the average treatment effect of rental participation,
rent-in and rent-out, using the Epanechnikov kernel estimator (KBM) with a bandwidth
of 0.06. Standard errors of the estimated ATT are bootstrapped with 200 replications.
For sensitivity analysis, other matching algorithms, of KBM with bandwidth 0.1, and
nearest neighbor matching (NNM) with 1, 5 and 10 neighbors, are also applied and
produce similar results (not shown and available upon request).
Renting in forestland could increase household income through forestry production,
and renting out land could release labor for greater engagement in off-farm work. As
expected, participation in forestland rental markets has a positive impact on household
annual per-capita income, of more than 1,500 CNY, and reduces the probability of
income falling below the poverty line, by 9 percentage points, for either renting in
or renting out forestland. These results support Hypotheses 3 and 4 – namely, rental
market participation has positive impacts on household income and poverty alleviation.
I find that renting out forestland leads to higher off-farm incomes, with an average
treatment effect of 1,646 CNY a year, and 7 percent more income from off-farm work.
12 The logit models for propensity scores have Pseudo-R2 values of 0.133, 0.177 and 0.183. These values are
similar to the levels in recent literature using the PSM strategy, e.g., Becerril and Abdulai, 2010; Kassie et
al., 2011. By contrast, significant parameter estimates (t-test) or goodness-of-fit measures like Pseudo-R2 can
be misleading (Heckman and Navarro-Lozano, 2004; Lee, 2013). They only show association and are used to
balance the observed distribution of covariates across the treated and the control groups, in order to provide
good matches for the estimation of ATT .
22
The findings lend support to my contention that renting out forestland enables family
labor to engage better in off-farm jobs, given that off-farm working days per household
labor do not differ across rental status, as observed in Table 4.
Renting in forestland seems not to increase forestry production value in a statis-
tically significant way. Renting out forestland significantly reduces household forestry
production, by an average value of 78 CNY per mu, and reduces income from forestry
by almost 5 percent, compared to the case if the household does not rent out. Results
indicate that the no-rent households are doing well as forest managers, so that they
are willing to maintain the forestland for their own use rather than rent it out.
< Table 9 here >
8 Conclusion
I analyzed panel survey data and found that, through allowing more productive house-
holds to access forestland and equalize factor ratios, rental markets improve allocative
efficiency and household welfare. The number of transactions of forestland transfers
is still very small compared to the rental transactions in agricultural land. Obviously,
the devolution in agricultural land to households started twenty years earlier than the
devolution in forest management. With the experiences and lessons from agriculture,
policy makers anticipate the same efficiency gains as were achieved with forestland
markets.
The main findings include the importance of factor endowment and managerial
ability, which determine productivity in forestry, in driving the direction of resource
allocation. Policies should improve rural households’ knowledge and skills in forest
management, and should facilitate more productive individuals in obtaining their op-
timal operational size of forestland. Transparent, well-functioning land exchange plat-
forms/markets have been put in place at the municipal level by the Chinese central
government. Policies that promote the inclusion of forestland in these markets and the
extension of such exchange platforms to counties and communities could encourage
forestland rental market development.
I also propose that policies should be devoted to address constraints and strengthen
local institutions to improve farmers’ access to factor resources and credit. Effective
efforts should include securing property rights to forestland. I find, and so have previous
studies, that households’ perceptions on property rights (or tenure security) drive their
decisions on management and transfer. However, in 2010, seven years after the formal
announcement of the Collective Forest Tenure Reform, only two-thirds of households
had received forestland certificates in the earliest reformed provinces.
23
Finally, the current low participation rate in forestland rental markets implies po-
tential for a future market. Although no evidence was found of agglomeration of forest-
land to larger, wealthier, or more powerful landholders, follow-up policies should still
be vigilant on avoiding consolidation among such households. On the contrary, forest-
land agglomeration to more efficient users is not bad, especially in the context where
the growth of rural migration would demand reallocation of factor resources in rural
areas. Given the results that rental participation enables households to better engage
in off-farm work, forestland could provide a household not only with a safety-net asset
but also with improved off-farm ability. From an integrative perspective, a shift of labor
to the manufacturing and service industries may play an underpinning role in fostering
the structural transformation of rural development in China.
References
[1] Alston, J. Lee, Samar K. Datta, and Jeffery B. Nugent. 1984. Tenancy choice in a
competitive framework with transaction costs. Journal of Political Economy 92 (6):
1121-33.
[2] Andre, Catherine, and Jean-Philippe Platteau. 1998. Land relations under unbear-
able stress: Rwanda caught in the Malthusian trap. Journal of Economic Behavior
and Organization 34 (1): 1-47.
[3] Bardhan, Pranab K., and Chris Udry. 1999. Development Microeconomics. Oxford,
UK: Oxford University Press.
[4] Barro, Robert J. 1991. Economic growth in a cross section of countries. Quarterly
Journal of Economics 106 (2): 407-43.
[5] Becerril, J. and A. Abdulai. 2010. The impact of improved maize varieties on
poverty in Mexico: a propensity score-matching approach. World Development 38
(7): 1024-1035.
[6] Besley, Timothy. 1995. Property rights and investment incentives: theory and evi-
dence from Ghana. Journal of Political Economy 103 (5): 903-37.
[7] Bliss, C.J. and N.H. Stern. 1982. Palanpur: The Economy of an Indian Village. Oxford,
UK: Clarendon Press
[8] Carter, M.R., and Yang Yao. 2002. Local versus global separability in agricultural
household models: the factor price equalization effect of land transfer rights. Amer-
ican Journal of Agricultural Economics 84 (3): 702-715.
24
[9] Chamberlin J, and J. Ricker-Gilbert. 2016. Participation in rural land rental mar-
kets in Sub-Saharan Africa: who benefits and by how much? Evidence from Malawi
and Zambia. American Journal of Agricultural Economics 98 (5): 1507-1528.
[10] Deininger, K. 2003. Land Policy for Growth and Poverty Reduction. World Bank
Policy Research Report, World Bank and Oxford University Press.
[11] Deininger, K., D.A. Ali and T. Alemu. 2011. Impacts of land certification on tenure
security, investment, and land market participation: evidence from Ethiopia. Land
Economics 87 (2): 312-334.
[12] Deininger, K., and Songqing Jin. 2005. The potential of land rental markets in
the process of economic development: evidence from China. Journal of Development
Economics 78: 241-270.
[13] Deininger, K. and S. Jin. 2008. Land sales and rental markets in transition: evi-
dence from rural Vietnam. Oxford Bulletin of Economics and Statistics (70): 67-101.
[14] Deininger, Klaus, Songqing Jin, and Hari K. Nagarajan. 2008. Efficiency and eq-
uity impacts of rural land market restrictions: evidence from India. European Eco-
nomic Review 52 (5): 892-18.
[15] Deininger, K., S. Jin, F. Xia, and J. Huang. 2014. Moving off the farm: land insti-
tutions to facilitate structural transformation and agricultural productivity growth
in China. World Development 59: 505-520.
[16] Deininger, K., and Paul Mpuga. 2009. Land markets in Uganda: what is their
impact and who benefits? In The Emergence of Land Markets in Africa, eds. Stein
T. Holden, Keijiro Otsuka, and Frank M. Place, pp. 131-158. Washington, DC:
Resources for the Future.
[17] Feder, Gershon. 1985. The relationship between farm size and farm productivity:
the role of family labor, supervision and credit constraints. Journal of Development
Economics 18 (2-3): 297-313.
[18] Gavian, S. and Marcel Fafchamps. 1996. Land tenure and allocative efficiency in
Niger. American Journal of Agricultural Economics 78 (2): 460-471.
[19] Holden, S.T., K. Deininger and H. Ghebru. 2011. Tenure insecurity, gender, low-
cost land certification and land rental market participation in Ethiopia. The Journal
of Development Studies 47 (1): 31-47.
25
[20] Heckman, J. and S. Navarro-Lozano. 2004. Using matching, instrumental variables,
and control functions to estimate economic choice models. Review of Economics and
Statistics 86 (1): 30-57.
[21] Hill, J.K., M.A. Gray, C.V. Khen, S. Benedick, N. Tawatao and K.C. Hamer. 2011.
Ecological impacts of tropical forest fragmentation: how consistent are patterns in
species richness and nestedness? Philosophical Transactions of the Royal Society of
London B: Biological Sciences, 366 (1582): 3265-3276.
[22] Holden, Stein T., and Keijiro Otsuka. 2014. The roles of land tenure reforms and
land markets in the context of population growth and land use intensification in
Africa. Food Policy 48: 88-97.
[23] Holden, S.T., K. Otsuka and F. Place (eds). 2009. The Emergence of Land Markets
in Africa: Impacts on Poverty, Equity, and Efficiency. Washington, DC: Resources for
the Future.
[24] Jin, S., and K. Deininger. 2009. Land rental markets in the process of rural struc-
tural transformation: productivity and equity impacts from China. Journal of Com-
parative Economics 37 (4): 629-46.
[25] Jin, Songqing and T.S. Jayne. 2013. Land rental markets in Kenya: implications for
efficiency, equity, household income, and poverty. Land Economics, 89 (2): 246-271.
[26] Kassie, M., B. Shiferaw, and G. Muricho. 2011. Agricultural technology, crop in-
come, and poverty alleviation in Uganda. World Development 39 (10): 1784-1795.
[27] Kimura, S., K. Otsuka, T. Sonobe, and S. Rozelle. 2011. Efficiency of land allo-
cation through tenancy markets: evidence from China. Economic Development and
Cultural Change 59 (3): 485-510.
[28] Lee, W. 2013. Propensity score matching and variations on the balancing tests.
Empirical Economics 44(1): 47-80.
[29] Lunduka, R., S.T. Holden, and R. Øygard. 2009. Land rental market participation
and tenure security in Malawi. In The Emergence of Land Markets in Africa: Impacts
on Poverty, Equity and Efficiency, Chapter 6, eds. S.T. Holden, K. Otsuka, and F.M.
Place, 112-30. Washington DC: Resources for the Future Press.
[30] Migot-Adholla, S.E., G. Benneh, F. Place, S. Atsu. 1994. Land, Security of Tenure,
and Productivity in Ghana. J.W. Bruce, S.E. Migot-Adholla (Eds.), Searching for
land tenure security in Africa, Kendall/Hunt, Dubuque, Iowa (1994), pp. 169-98.
26
[31] National Bureau of Statistics of the People’s Republic of China (NBS). 2014. China
Statistical Yearbook 2014, China Statistics Press.
[32] Nkonya, Ephraim, John Pender, Samuel Benin, and Edward Kato. 2009. Land
rental markets and land management: Evidence from Uganda. In The Emergence of
Land Markets in Africa, eds. Stein T. Holden, Keijiro Otsuka, and Frank M. Place,
238-253. Washington, DC: Resources for the Future.
[33] North, Douglass C. 1981. Structure and Change in Economic History. New York:
Norton, 1981.
[34] Otusuka, K., 2007. Efficiency and equity effects of land markets, in Evenson, R.,
and Pingali, P. (eds.), Handbook of Agricultural Economics, 3: 2671-2703.
[35] Otsuka, K., and Y. Hayami. 1988. Theories of share tenancy: a critical survey.
Economic Development and Cultural Change 37 (1): 31-68.
[36] Pender, John, and Marcel Fafchamps. 2006. Land lease markets and agricultural
efficiency in Ethiopia. Journal of African Economies 15 (2): 251-84.
[37] Petrin, A. and K. Train. 2006. Control function corrections for omitted attributes
in differentiated product markets. Quantitative Marketing and Economics.
[38] Rosenbaum, P. R. and D. B. Rubin. 1983. The central role of the propensity score
in observational studies for causal effects. Biometrika 70 (1): 41-55.
[39] Rozelle, S., L. Brandt, L. Guo, J. Huang. 2002. Land rights in China: facts, fictions,
and issues. China Journal 47 (1): 67-97.
[40] Schlenker, Wolfram and Michael J. Roberts. 2009. Nonlinear temperature effects
indicate severe damages to U.S. crop yields under climate change. Proceedings of the
National Academy of Sciences of the United States of America 106 (37): 15594-98.
[41] State Forestry Administration, P.R. China. 2011. Forest Resource Statistics of China.
(Beijing, China).
[42] Siikama¨ki, J., Y. Ji, and J. Xu. 2015. Post-reform forestland markets in China.
Land Economics 91 (2): 211-234.
[43] Skoufias, Emmanuel. 1995. Household Resources, transaction costs, and adjust-
ment through land tenancy. Land Economics 71 (1): 42-56.
[44] Wen, G. J. 1989. The Current Land Tenure System and Its Impact on Long Term
Performance of Farming Sector: The Case of Modern China. (Doctoral dissertation,
University of Chicago, Department of Economics).
27
[45] Wooldridge, Jeffrey M. 2010. Econometric Analysis of Cross Section and Panel Data.
Cambridge, MA: MIT Press.
[46] Xu, Jintao, Andy White, and Uma Lele. 2010. China’s Forest Tenure Reforms: Im-
pacts and Implications for Choice, Conservation, and Climate Change. Rights and Re-
sources Initiative. Washington D.C., USA.
[47] Yamano, Takashi, Frank M. Place, Wilfred Nyangena, Juliet Wanjiku, and Keijiro
Otsuka. 2009. Efficiency and equity impacts of land markets in Kenya. In The
Emergence of Land Markets in Africa, eds. Stein T. Holden, Keijiro Otsuka, and
Frank M. Place, pp. 93-111. Washington, DC: Resources for the Future.
[48] Zimmerman F. J., and M. R. Carter. 1999. A dynamic option value for institutional
change: marketable property rights in the Sahel. American Journal of Agricultural
Economics 81 (2): 467-478.
28
Figure 1: Distribution of household forestry productivity in three provinces
Notes: The graphs depict kernel density estimates of the distribution of households’
forestry productivity as predicted from the household fixed-effects Cobb-Douglas
forestry production functions.
29
Figure 2: Propensity score distribution and common support
for propensity score estimation
(a) Rental Participation
(b) Rent-in
(c) Rent-out
Notes: For each of the three treatments: “rental participation” in (a), “rent-in” in (b),
and “rent-out” in (c), all treated observations have common support. That is, all the
“Treated” observations, i.e., each of the 180 rental participating households, each of
the 93 renting-in households, and each of the 87 renting-out households, has a
suitable comparison.
30
Table 1: Emerging forestland rental market, sample mean, by reform and time
2005 2010
No- With- No- With-
All Reform Reform All Reform Reform
Total forestland area (mu)*, endowment 48.99 36.59 57.89 52.21 12.43 52.97
% of households renting-in 2.10 1.41 2.59 7.03 0 7.17
% of households renting-out 2.85 4.42 1.73 5.48 0 5.58
Rented-in forestland, area (mu)* 0.854 0.274 1.269 4.229 0 4.309
Rented-out forestland, area (mu)* 2.070 4.189 0.549 2.146 0 2.186
Rented-in forestland, % to endowment 1.18 0.673 1.54 4.10 0 4.17
Rented-out forestland, % to endowment 1.32 1.84 0.942 1.99 0 2.03
Notes: * 1 mu=1/15 hectare.
Author’s computation based on representative survey data of 2005 and 2010 in Fujian, Jiangxi and
Yunnan provinces of China, conducted by the Environmental Economics Program of China at Peking
University, China.
31
Table 2: Household characteristics by reform and time (Mean)
2005 2010
No- With- No- With-
All Reform Reform All Reform Reform
Demographic characteristics:
Number of household members 4.758 4.721 4.784 4.851 5.611 4.837
Number of labor equivalents 3.398 3.239 3.512 3.347 3.556 3.344
Household head age (years) 47.74 44.70 49.91 50.82 54.67 50.75
Household head gender (1=male, 0=female) 0.949 0.920 0.970 0.946 0.667 0.952
Household head is Communist (1 if yes, 0 if no) 0.153 0.114 0.180 0.174 0.111 0.175
Household head is village leader (1 if yes, 0 if no 0.0629 0.0562 0.0677 0.0558 0 0.0569
Average education (years) 5.280 5.029 5.460 5.557 6.897 5.531
Work experience in forestry department (1=yes, 0=never) 0.0201 0.0141 0.0245 0.0496 0 0.0506
Forestland, cropland, assets and incomes:
Mu of forestland per labor 16.13 12.47 18.75 17.24 5.331 17.46
Mu of forest endowed to household 48.99 36.59 57.89 52.21 12.43 52.97
Capital investment in forestry (CNY) 647.9 796.7 541.1 1,353 706.7 1,365
Annual value of forestry production (CNY) 3,851 1,356 5,642 859.9 6.624 876.1
Mu of cropland per capita 1.644 2.037 1.361 1.931 0.552 1.957
Off-farm work days per labor 106.7 78.24 127.2 110.9 139.9 110.3
Total off-farm income (CNY) 12,257 8,197 15,171 25,879 28,067 25,838
Total livestock value (CNY) 1,358 2,450 574.7 3,829 343.3 3,895
Total house value (10,000 CNY) 3.379 3.244 3.476 11.48 44.69 10.85
Income per capita (CNY) 4,134 3,109 4,870 8,772 7,256 8,801
Risk perception, credit constraint, and restriction on forestland:
Holding a forestland certificate (1 if yes, 0 if no) 0.110 0 0.189 0.635 0 0.647
Ability to obtain credita 0.787 0.772 0.797 0.948 1 0.947
Credit access (distance to local bank, km) 8.209 10.63 6.475 9.165 0.311 9.333
Perception of ownership/production riskb 0.690 0.951 0.504 0.931 0.889 0.932
Right to change forest type (1 if yes, 0 if no) 0.878 0.867 0.885 0.807 0.556 0.811
Right to transfer forestland within village 0.693 0.645 0.727 0.687 0.500 0.690
Right to transfer forestland to outsiders 0.617 0.581 0.642 0.609 0.333 0.614
Right to mortgage forestland 0.430 0.406 0.447 0.309 0.361 0.308
Number of observations 1192 498 694 967 18 949
Notes: 1 mu=1/15 hectare.
a Ability to obtain credit measures household ability to obtain credit, with value equal to 1 if they can successfully borrow 500
CNY within a week, 0 if not.
b 1 if secure, 0.5 if unsure or do not know, 0 if insecure.
32
Table 3: Villages characteristics by reform and time (Mean)
2000 2005 2010
No- With- No- With- No- With-
Reform Reform Reform Reform Reform Reform
Length of period with the Reform (years) 0 1 0 2.195 0 6.581
Number of households in village 349.3 152.5 536.4 300.1 867 413.3
Population 1,504 642.8 2,457 1,224 2,471 1,650
Average per capita income (CNY) 1,640 2,800 1,290 2,374 3,213 3,022
Distance to paved road (km) 0.601 0 1.517 0.0629 0 0.541
Distance to closest county center (km) 36.08 65.75 39.76 36.08 0 35.03
Population share possessing telephones 0.286 0.497 0.583 0.751 1 0.928
Development of labor marketa 0.185 0.0724 0.211 0.284 0.731 0.301
Village cropland area share 0.111 0.0512 0.121 0.0884 0.0434 0.125
Village forestland area share 0.700 0.904 0.751 0.715 0.957 0.654
Village total forestland area (mu) 15,587 21,185 23,092 13,267 10,080 16,420
Proportion of household-managed forestland areab 0.525 0.427 0.546 0.608 0.00645 0.617
Proportion of forestland involved in transfers 0.0556 0.105 0.0159 0.0793 0 0.0958
Commercial timber price (10,000 CNY) 0.0269 0.0310 0.0406 0.0469 0.0208 0.0816
Annual precipitation: average (mm) 44.32 45.24 32.63 50.35 27.12 51.03
Annual precipitation: standard deviation (mm) 107.2 107.3 81.10 137.1 70.90 125.5
Effective cumulative heat, 0-35◦C, GDDc 6,687 7,606 6,403 6,962 5,281 6,888
Harmful cumulative heat, >35◦C, HDDc 2.794 6.352 0.00701 6.544 0 5.580
Number of villages 102 4 29 77 1 105
Source: Author’s computation based on representative survey data of 2005 and 2010 in Fujian, Jiangxi and Yunnan
provinces of China, conducted by the Environmental Economics Program of China at Peking University, China; the
village-level data on 2000 is recalled. Weather data is obtained from the China Meteorological Data Sharing Service
System (CMDSSS).
Notes: a Development of labor market is measured by the ratio of number of off-farm workers to the total labor force in
village.
b In 2000, the variable refers to the ratio of forestland area accessed by households over the village total forest area.
c Effective cumulative heat, or growing degree days (GDD), is a generated measure of days with 0-35◦C. Harmful
cumulative heat, or harmful degree days (HDD), is generated for days with temperature higher than 35◦C. GDD =∑N
i Ti,a − Tbase, where Ti,a is the daily average temperature for day i, and Tbase is the base temperature below which
vegetation ceases to be biologically active (here we select 0◦C for trees). HDD = ∑Ni (Ti,max − 35◦C)/Ti,max, whereTi,max is maximum temperature for day i.
33
Table4:Householdcharacteristicsbyrentalandreformstatusin2005and2010(Mean)
20052010
No-ReformWith-ReformNo-ReformWith-Reform
Rent-outNo-rentRent-inRent-outNo-rentRent-inNo-rentRent-outNo-rentRent-in
Demographiccharacteristics:
Numberofhouseholdmembers5.3644.6954.4294.7504.8103.8335.6114.6044.8365.029
Numberoflaborequivalents3.9553.2132.7143.8333.5262.7783.5563.2833.3263.603
Householdheadage(years)45.7344.6743.5751.8350.0942.2254.6753.6750.6849.35
Householdheadgender(1=yes,0=never)0.9090.9210.85710.96810.6670.8870.9540.971
HouseholdheadisCommunist(1=yes,0=never)0.1820.1110.1430.08330.1820.1670.1110.1890.1710.206
Householdheadisvillageleader(1=yes,0=never)00.05760.14300.0708000.05660.05310.103
Averageeducation(years)6.2844.9883.8976.0935.4455.5816.8974.8965.5465.844
Workexperienceinforestrydepartment(1=yes,0=never)0.04550.01070.14300.0256000.07550.04710.0735
Forestland,cropland,assets,andincomes:
Muofforestlandperlabor,endowed79.769.4126.22438.1518.4317.615.33143.3416.1812.88
Muofforestlandperlabor,currentlymanaged44.679.41213.1828.0718.4335.725.33133.3116.1834.64
Capitalinvestmentinforestry(CNY)481.0803.21,353525515.81,487706.7819.11,2653,005
Annualvalueofforestryproduction(CNY)69.471,27211,0142,4865,56910,4366.624915.8732.62,593
Muofcroplandpercapita2.4331.9953.6410.7891.3701.4320.5521.4732.0341.393
Off-farmworkdaysperlabor75.7378.4174.88110.1127.9112.9139.9111.5109.2123.4
Totaloff-farmincome(CNY)13,9657,9735,09727,27815,04211,84728,06729,72624,56538,310
Totallivestockvalue(CNY)1,9072,4891,509300.8592.0118.9343.32,8074,1182,037
Totalhousevalue(10,000CNY)3.7253.2511.2942.5633.5122.74944.6910.9810.2118.53
Incomepercapita(CNY)4,4783,0373,6798,0874,8005,3067,2569,3408,63210,443
Riskperception,creditconstraint,andrestrictiononforestland:
Holdingaforestlandcertificate(1ifyes,0ifno)0000.1670.1850.33300.8110.6440.559
Abilitytoobtaincredita0.9090.76210.8330.7940.889110.9420.971
Creditaccess(distancetolocalbank,km)3.15910.9710.794.2086.4787.8610.3117.3239.7216.168
Perceptionofownership/productionriskb0.8860.95310.5000.5040.5000.8890.9430.9300.941
Righttochangeforesttype(1ifyes,0ifno)0.7730.87010.8330.8890.7780.5560.7550.8100.868
Righttotransferforestlandwithinvillage0.2730.6630.6430.8540.7260.6670.5000.6890.6830.787
Righttotransferforestlandtooutsiders0.2500.5980.5000.6880.6420.6110.3330.6600.6030.706
Righttomortgageforestland0.1140.4160.6430.5000.4420.5830.3610.2360.3050.397
LengthofperiodwiththeReform(years)0002.6672.1972.83306.2456.0106.941
Proportionofforestlandinvolvedintransfersin20000.01460.01530.04470.1630.07020.06140.05850.07900.06840.104
Numberofobservations2246971266418185382868
Note:In2010,noneofthe18sampledhouseholdsinthenon-reformedvillagerentedanyforestland.
1mu=1/15hectare. a
Abilitytoobtaincreditmeasureshouseholdabilitytoobtaincredit,withvalueequalto1iftheycansuccessfullyborrow500CNYwithinaweek,0ifnot. b
1ifsecure,0.5ifunsureordonotknow,0ifinsecure.
34
Table 5: Household forestry: Cobb-Douglas production function (FE estimates)
Coefficient Std.Err. (P -value)
Log labor equivalents 0.139 (0.343)
Log forestland endowment (mu) 0.225 (0.120)*
Log value of investment in forestry (CNY) 0.180 (0.0412)***
Log household head age (years) 0.807 (0.731)
Household head gender (1=yes, 0=never) -0.261 (0.427)
Household head is Communist (1=yes, 0=never) -0.184 (0.549)
Household head is village leader (1=yes, 0=never) 0.417 (0.379)
Average education (years) 0.0346 (0.0779)
Work experience in forestry department (1=yes, 0=never) 0.141 (0.555)
Log cropland per capita (mu) 0.160 (0.251)
Log off-farm work days per labor 0.109 (0.0508)**
Log total livestock value (CNY) 0.00103 (0.0278)
Log total house value (10,000 CNY) 0.182 (0.114)
Holding a forestland certificate (1 if yes, 0 if no) 0.191 (0.308)
Ability to obtain credit (1=can borrow 500 CNY in a week) 0.0568 (0.286)
Perception of ownership/production risk (1 if secure) 0.467 (0.492)
Log distance to local bank (km) -0.748 (0.121)***
Log distance to closest county (km) 0.769 (0.0692)***
Log timber price (CNY) 0.263 (0.105)**
Annual precipitation: average (mm) -0.236 (0.0325)***
Annual precipitation: standard deviation (mm) 0.064 (0.00706)***
Effective cumulative heat, 0-35◦C, GDD 0.00239 (0.00134)*
Harmful cumulative heat, >35◦C, HDD -0.0595 (0.0205)***
R-squared 0.394
Number of households 1,266
Number of observations 2,155
Notes: Table shows coefficient estimates from household fixed effects regression. The dependent variable
is log of the annual incremental value of forestry. Model includes year, province year and village year
dummies (not shown). Standard errors are robust to clustering at the village level. Significance is
denoted by *** p<0.01, ** p<0.05, * p<0.1.
35
Table 6a: Determinants of forestland rental market participation
(Multinomial Logit Model estimates)
(1) Full sample (2) 2005 samples (3) 2010 samples
Rent-out Rent-in Rent-out Rent-in Rent-out Rent-in
αˆ (forestry productivity) 0.017 0.172*** -0.063 0.214** 0.058 0.195***
(0.051) (0.054) (0.082) (0.100) (0.070) (0.060)
Log forestland endowment 0.609*** -0.524*** 0.673*** -0.579** 0.471*** -0.573***
(0.132) (0.186) (0.196) (0.281) (0.167) (0.173)
Log household head age 0.019* -0.027** 0.016 -0.064*** 0.023 -0.017
(0.011) (0.012) (0.023) (0.023) (0.014) (0.016)
Household head gender (1=male) -0.501 0.043 1.073 -0.940 -0.775 0.616
(1.155) (2.010) (0.894) (0.853) (0.537) (0.730)
Household head is Communist (1=yes) 0.023 -0.229 -0.203 0.560 0.258 -0.347
(0.396) (0.428) (0.568) (0.715) (0.445) (0.421)
Household head is village leader (1=yes) -0.514 0.354 -13.887*** -1.224 -0.227 0.774
(3.004) (0.467) (0.555) (1.335) (0.802) (0.561)
Average education (years) -0.120 0.024 0.054 0.026 -0.227*** 0.048
(0.074) (0.063) (0.146) (0.111) (0.072) (0.072)
Work experience in forestry department (1 if yes) 0.152 0.361 0.252 0.638 -0.019 0.199
(0.637) (1.293) (0.880) (1.161) (0.688) (0.544)
Log cropland per capita -0.370 -0.152 -1.314** 0.548 -0.218 -0.161
(0.303) (0.305) (0.627) (0.597) (0.282) (0.324)
Log off-farm work days per labor 0.076 -0.019 0.025 -0.013 0.106 -0.034
(0.084) (0.075) (0.120) (0.140) (0.096) (0.087)
Log total livestock value 0.017 -0.014 0.117* 0.001 -0.026 -0.025
(0.042) (0.041) (0.070) (0.090) (0.048) (0.045)
Log timber price 0.638** 0.304 0.572 0.196 0.557** 0.231
(0.281) (0.267) (0.471) (0.533) (0.248) (0.269)
Right to change forest type (1 if yes) -0.185 0.138 0.370 0.170 -0.277 0.380
(0.331) (0.412) (0.518) (0.719) (0.380) (0.472)
Log total house value 0.029 0.276* -0.346 -0.172 0.184 0.382**
(0.158) (0.164) (0.281) (0.384) (0.170) (0.164)
Ability to obtain credit 0.982 0.630 0.023 1.156* 14.963*** -0.109
(2.416) (2.855) (0.721) (0.699) (0.651) (0.875)
Credit access (km to local bank, log) -0.593*** -0.680*** -1.516* -0.794** -0.559* -0.735***
(0.212) (0.214) (0.839) (0.398) (0.295) (0.225)
Risk perception (1=secure) -0.950 1.687 -4.689*** 28.410*** -0.100 1.433
(0.836) (1.294) (1.393) (2.515) (0.747) (1.069)
Holding a forestland certificate (1 if yes) 0.546 0.360 -1.052 0.274 0.758 0.374
(0.437) (0.370) (1.107) (0.644) (0.480) (0.372)
Right to transfer forestland within village -0.706 -0.211 0.099 -0.687 -1.304 -0.060
(3.241) (0.572) (1.031) (0.997) (0.676) (0.593)
Right to transfer forestland to outsiders 0.605 0.024 -0.111 0.217 1.165* 0.161
(3.209) (0.525) (1.055) (0.907) (0.705) (0.529)
Right to mortgage forestland -0.081 0.343 0.138 1.087** 0.036 0.128
(0.321) (0.290) (0.547) (0.515) (0.432) (0.315)
Earlier transfers in village (%) 0.154 0.028 3.542 -1.557 -0.602 0.269
(1.070) (1.032) (2.183) (3.142) (1.171) (0.963)
Distance to paved road (km, log) -0.086 0.184 -1.327 0.216 0.103 0.203
(0.217) (0.276) (0.826) (0.667) (0.224) (0.250)
Distance to closest county center (km, log) -0.057 -0.482* -0.690 0.336 0.263 -0.581**
(0.217) (0.249) (0.446) (0.647) (0.304) (0.269)
continued on next page
36
Table 6a continued
Village total forestland area (mu, log) 0.428* 0.613*** 0.859 1.132*** 0.326 0.757***
(0.226) (0.229) (0.615) (0.422) (0.210) (0.245)
Village cropland area share (%) 4.194*** 2.915* 11.975** 5.070 2.766** 3.212***
(1.557) (1.644) (5.191) (3.689) (1.253) (1.235)
Log number of households in village -0.808 -1.023 -4.088 -4.603*** -2.104 0.183
(1.552) (1.088) (3.110) (1.413) (2.339) (1.377)
Log village population 0.329 -0.242 2.584 3.373 1.417 -1.621
(1.512) (0.980) (2.826) (1.362) (2.378) (1.285)
Log village average pc income (CNY) -0.890 -1.259** -2.504*** 0.163 -1.143 -1.593***
(0.560) (0.584) (0.906) (1.301) (0.739) (0.516)
Village income growth 0.149 -0.236 -1.125 -1.699 0.482 0.042
(1.019) (0.774) (1.048) (1.064) (0.485) (0.685)
Population share possessing telephones -2.201 -3.767*** -4.589 -0.578 -2.436 -4.244***
(1.390) (1.452) (3.051) (3.519) (1.739) (1.342)
Development of labor market (%) -1.200 -0.439 -3.572* -0.956 -0.754 -0.998
(1.301) (1.005) (1.835) (3.583) (1.897) (1.119)
Annual precipitation: average (mm) -0.140*** 0.019 -0.265*** 0.335** -0.126*** -0.008
(0.042) (0.054) (0.097) (0.136) (0.048) (0.046)
Annual precipitation: standard deviation (mm) 0.036*** -0.007 0.035 -0.091** 0.040** 0.003
(0.013) (0.014) (0.036) (0.041) (0.016) (0.014)
Effective cumulative heat (days 0-35◦C) -0.000 0.001** -0.000 0.002*** 0.000 0.001**
(0.000) (0.000) (0.001) (0.001) (0.000) (0.000)
Harmful cumulative heat (days >35◦C) -0.036 -0.016 -0.155 -0.072 -0.021 -0.011
(0.032) (0.040) (0.095) (0.089) (0.044) (0.048)
Length of period with the Reform (years) 0.251 0.805** 1.644** 0.531 -0.192 0.685**
(0.320) (0.314) (0.678) (0.690) (0.363) (0.279)
υˆjt (Reform selectivity) -0.223 -0.578** -1.190** -0.533 0.088 -0.461*
(0.251) (0.263) (0.492) (0.574) (0.307) (0.241)
Number of observations 2,151 2,151 1,187 1,187 964 964
Log Lik -567.5 -567.5 -171.4 -171.4 -339.6 -339.6
Pseudo R-squared 0.230 0.230 0.367 0.367 0.236 0.236
Notes: The first-stage regression of length of period with the Reform is on a set of village characteristics for 2000 and a set in
the current year. F -statistic reports 57.92, and adjusted R-squared is 62%. Second-stage MNL regressions include year trends,
province fixed effects and province year effects (not shown because of insignificance). Robust standard errors are in parentheses,
using a bootstrapped method with 200 iterations in model 1, and household-clusters adjusted in models 2 and 3. Significance is
denoted: *** p<0.01, ** p<0.05, * p<0.1.
37
Table6b:Determinantsofforestlandrentalmarketparticipation
(MarginaleffectsfromMultinomialLogitModel,Delta-method)
Rent-outNo-rentRent-out
dy/dxStd.Err.dy/dxStd.Err.dy/dxStd.Err.
αˆ(forestryproductivity)0.0290.168-0.618***0.2360.588***0.168
Forestlandendowment(mu,1%)2.159***0.394-0.2620.674-1.896***0.555
Householdheadage(years)0.070*0.0360.0250.057-0.095**0.045
Averageeducation(years)-0.415*0.2370.3110.3100.1030.211
Totalhousevalue(1%)0.0540.461-0.9970.6540.944**0.466
Abilitytoobtaincredit(0−1)3.247*1.867-5.254**2.5002.0081.925
Creditaccess(kmlocalbank,1%farther)-1.912***0.6324.153***0.885-2.240***0.651
Riskperception(0−1)-3.509*2.024-2.4404.1145.949*3.599
Timberprice(1%)2.126**0.834-3.067***1.1360.9410.769
Villagetotalforestlandarea(mu,1%)1.361*0.664-3.400***1.0012.038***0.707
Villagecroplandareashare(%,0−1)13.825***3.885-23.164***6.1619.338**3.696
Earliertransfersinvillage(%,0−1)0.5222.954-0.5954.0140.0732.989
Villageaveragepcincome(1%)-2.831*1.6337.014***2.183-4.183***1.492
Distancetoclosestcounty(km,1%)-0.1180.6551.766*0.950-1.648**0.691
Populationsharepossessingtelephones(%,0−1)-6.895*3.82619.483***5.516-12.588***4.020
Annualprecipitation:average(mm)-0.479***0.1250.392**0.1840.0870.146
Annualprecipitation:standarddeviation(mm)0.124***0.041-0.093*0.053-0.0310.037
Effectivecumulativeheat(days0-35◦C)-0.0010.001-0.0020.0010.003***0.001
LengthofperiodwiththeReform(years)0.7260.939-3.452***1.2822.726***0.901
Notes:Tableshowsmarginaleffectsofaselectionofvariablesevaluatedatsamplemeans,basedontheestimationresultsofmodel1
intable6a.Delta-methodisusedforcomputation.Significanceisdenoted:***p<0.01,**p<0.05,*p<0.1.
38
Table 7: Determinants of forestland rent-in and rent-out (Random Effect Tobit Model)
Area % of Endowment
Rented-out Rented-in Rented-out Rented-in
αˆ (forestry productivity) -7.346 9.698** -0.030 0.053**
(5.502) (3.775) (0.022) (0.024)
Log forestland endowment 100.460*** -45.187*** 0.270*** -0.440***
(13.156) (10.906) (0.055) (0.073)
Log household head age 0.591 -2.754*** 0.003 -0.015**
(1.142) (0.947) (0.005) (0.006)
Household head gender -19.290 19.454 -0.177 0.110
(46.567) (56.747) (0.184) (0.346)
Household head is Communist -18.995 2.018 -0.042 0.062
(33.916) (25.083) (0.139) (0.160)
Household head is village leader 14.301 7.446 -0.240 0.079
(53.989) (35.224) (0.250) (0.221)
Average education years 3.201 5.434 -0.017 0.011
(6.464) (4.650) (0.026) (0.029)
Work experience in forestry department -64.985 44.009 -0.237 0.143
(55.705) (39.048) (0.239) (0.254)
Log cropland per capita -39.573 30.966* -0.257** 0.152
(28.189) (18.200) (0.119) (0.116)
Log off-farm work days per labor -4.093 -2.088 0.007 -0.000
(6.405) (5.089) (0.027) (0.032)
Log total livestock value 7.083* 0.467 0.022 0.006
(3.725) (3.059) (0.015) (0.019)
Log timber price 207.751 78.108* 0.853 0.530**
(170.712) (43.815) (0.539) (0.266)
Right to change forest type -13.210 24.213 -0.109 0.109
(30.223) (26.143) (0.119) (0.160)
Total house value -1.703 11.930 -0.029 0.041
(13.416) (9.442) (0.054) (0.060)
Ability to obtain credit 38.435 19.397 0.244 0.297
(46.191) (34.202) (0.189) (0.222)
Credit access (km to local bank, log) 30.106 -60.184 0.378 -0.434
(80.594) (55.697) (0.293) (0.340)
Risk perception -182.293*** 106.762 -0.618** 0.533
(65.340) (71.560) (0.268) (0.434)
Holding a forestland certificate -23.710 39.042 -0.013 0.161
(45.903) (25.472) (0.172) (0.162)
Right to transfer forestland within village -6.160 -33.379 -0.104 -0.262
(47.162) (39.701) (0.197) (0.241)
Right to transfer forestland to outsiders 40.768 40.259 0.167 0.180
(46.920) (35.783) (0.197) (0.221)
Right to mortgage forestland -8.161 26.027 0.004 0.141
(28.406) (19.667) (0.114) (0.122)
continued on next page
39
Table 7 continued
Earlier transfers in village -337.277** 110.926 -1.199* 0.530
(164.589) (151.280) (0.612) (0.887)
Distance to paved road (km, log) 40.378 2.413 0.184 0.074
(49.687) (31.115) (0.196) (0.193)
Distance to closest county center (km, log) -1,182.138* -180.106* -4.261* -1.102*
(608.678) (105.869) (2.201) (0.630)
Village total forestland area (mu, log) -174.739 -211.777* -0.359 -1.308*
(169.389) (124.114) (0.553) (0.756)
Village cropland area share -538.202 -45.326 -0.030 -0.244
(1,287.811) (185.803) (4.355) (1.215)
Log number of households in village -242.241 152.191 -0.682 0.892
(379.638) (185.403) (1.373) (1.119)
Log village population -2,091.388* 501.198 -6.695 3.618
(1,228.247) (458.668) (4.533) (2.833)
Log village average pc income -219.599 -75.460 -0.589 -0.354
(167.966) (61.154) (0.690) (0.362)
Village income growth 48.631 -6.638 0.188 -0.209
(59.590) (59.716) (0.239) (0.381)
Population share possessing telephones 141.047 -117.632 0.463 -0.930
(228.755) (123.791) (0.833) (0.751)
Development of labor market -920.215* -57.267 -3.012* -0.480
(535.685) (161.956) (1.714) (0.930)
Annual precipitation: average (mm) -48.438* 6.259 -0.168* 0.035
(26.257) (7.881) (0.088) (0.048)
Annual precipitation: standard deviation (mm) 14.159** -1.363 0.052*** -0.007
(5.751) (2.133) (0.019) (0.013)
Effective cumulative heat (days 0-35◦C) -0.992 0.240 -0.004 0.001
(0.689) (0.357) (0.003) (0.002)
Harmful cumulative heat (days >35◦C ) 83.899** -4.846 0.321** -0.080
(36.139) (14.063) (0.132) (0.085)
Length of period with the Reform -39.218 230.954*** -0.015 1.567***
(144.038) (88.390) (0.426) (0.536)
υˆjt (Reform selectivity) -27.266 -25.929 -0.143 -0.140
(32.337) (20.332) (0.126) (0.124)
Observations 2,151 2,151 2,151 2,151
Number of households 1,264 1,264 1,264 1,264
Log Lik -662.3 -684.3 -208.0 -236.1
Notes: Robust standard errors in parentheses, adjusted for clustering effect at the village level. Province,
village and year dummies and the interaction between year and province are included. *** Significant at
1%; ** significant at 5%; * significant at 10%.
40
Table8:Differencesinwelfareindicatorsacrossno-rent,rent-in,rent-outhouseholds(Samplemean)
No-rentRentingRent-inRent-out
householdshouseholdsDifferencehouseholdsDifferencehouseholdsDifference
Variable(A)(B)(B-A)(c)(C-A)(D)(D-A)
Percapitaincome,CNY6007.68455.22447.6***8939.52931.9***7937.41929.8**
Poverty:likelihoodunderpovertyline0.3050.156-0.149***0.129-0.176***0.184-0.117**
Off-farmincomepercapita,CNY3495.55773.72278.2***6253.32757.8***52611765.5***
Off-farmincomeshare0.5140.5680.054**0.5720.058*0.5640.05
Forestryproduction,CNY/mu162.492.6-69.8171.99.57.7-154.7
Forestryproductionshare0.1010.07-0.031**0.098-0.0030.04-0.061***
Numberofobservations19791809387
Notes:Author’sowncomputation.Column(B)includesboththerent-inhouseholdsandrent-outhouseholds.Povertylineisdefinedas2000CNY
for2005and2300CNYfor2010,inaccordancewiththe1.25Purchasing-Power-Parity(PPP)U.S.dollarsadayin2005astheWorldBank’s
globalpovertystandard.Significanceisdenoted:***p<0.01,**p<0.05,*p<0.1.
41
Table9:Welfareeffectsofforestlandrentalparticipation,PSMestimates
AverageTreatmentEffects(ATT,θ)
WelfareindicatorRentingStd.Err.Rent-inStd.Err.Rent-outStd.Err.
Incomepercapita1538.36(657.10)**1318.91(973.18)1849.48(986.91)*
Absolutepoverty-0.093(0.027)***-0.093(0.040)**-0.088(0.051)*
Off-farmincomepercapita1445.30(600.73)**1423.21(974.94)1645.87(812.37)**
Proportionofincomefromoff-farm0.035(0.028)0.027(0.040)0.070(0.040)*
Forestryincrementalvaluepermu1.47(78.25)76.86(150.72)-77.85(16.57)***
Proportionofincomefromforestry-0.020(0.014)0.020(0.023)-0.047(0.022)**
Notes:Tablereportsresultsofaveragetreatmenteffectinthetreated(ATT)usingthematchingalgorithm−the
Epanechnikovkernelestimator(KBM)withbandwidths0.06.Standarderrors,inparentheses,arebootstrappedwith
200iterations.Significanceisdenoted:***p<0.01,**p<0.05,*p<0.1. 42
Chapter III

Managerial Incentives for Environmental Protection in
Chinese-Style Federalism*
Yuanyuan Yi„1, Wolfgang Habla…2 and Jintao Xu§3
1Department of Economics, University of Gothenburg, Sweden
2Center for European Economic Research (ZEW), Mannheim, Germany
3National School of Development, Peking University, China
Abstract
China’s fast economic growth has come at the expense of environmental quality and
the degradation of natural resources such as forests. In this paper, we identify career
concerns by managers of state-owned enterprises that manage natural resources, and
asymmetric information between managers and their superiors regarding the enterprises’
environmental performance, as sources of environmental degradation. A manager of such
an enterprise is the agent of two principals: national and sub-national governments. As
well as needing to meet ecological targets imposed by the national government, a manager
wants to be promoted into the ranks of the sub-national government. We develop three
hypotheses based on a theoretical model with two principals and one agent. We then
empirically test these hypotheses for the case of China’s northeastern state-owned forests,
combining satellite imagery data on deforestation with economic survey data. Our findings
suggest that managers of state forests that have a larger area and volume, and are thus
more difficult to monitor with respect to ecological targets, log more timber and are more
likely to deforest. The same holds true for managers who share a larger percentage of
profits with the local government. In turn, we find that sharing more revenue with the
local government increases the likelihood of getting promoted.
JEL Classification: H77, O13, O44, Q23, Q56
Keywords: managerial incentives; environmental protection; deforestation; state-owned
forests; state-owned enterprises; Chinese-style federalism; two-principal, one-agent model
*We thank Peter Berck, Fredrik Carlsson, Jessica Coria, Gunnar Ko¨hlin, Thomas Sterner and seminar
participants at the University of Gothenburg for valuable comments and suggestions. We also acknowledge
generous financial support from the following organizations: the Ford Foundation, FORMAS through the
research program Human Cooperation to Manage Natural Resources (COMMONS), the National Science
Foundation of China (NSFC), the International Institute of Applied Systems Analysis (IIASA), the Swedish
International Development Cooperation Agency (Sida) through the Environmental Economics Unit at the
University of Gothenburg, the State Forestry Administration of China, and the World Bank PROFOR project.
„Vasagatan 1, SE-405 30 Gothenburg, Sweden, Email: yuanyuan.yi@economics.gu.se
…L7 1, DE-68161 Mannheim, Germany, Email: habla@zew.de
§No.5 Yiheyuan Rd., Beijing 100871, China, Email: xujt@pku.edu.cn
1 Introduction
China’s remarkable economic growth has come at a tremendous cost to the environ-
ment, with inefficient, excessive resource use and high levels of pollution (Liu and
Diamond, 2008). The growth and environmental consequences are closely linked to
so-called Chinese-style federalism (Montinola et al., 1995; Xu, 2011). A key feature
of this type of federalism is that it combines fiscal decentralization with performance-
based personnel control. The decentralization of fiscal authority and a fiscal transfer
system allow regional governments (provincial, municipal, county and township level
governments) to have primary control over economic issues, including firms in their ju-
risdiction, while the central government typically owns natural resources (through var-
ious agencies) and sets pollution targets. In this system, short-term economic growth
is rewarded with promotions; by contrast, longer-term environmental issues such as
resource degradation and pollution do not negatively affect the likelihood of being
promoted.
In this paper, we argue that a systematic cause for the neglect of the environment
is the fact that authority over the environment lies with the central government, while
local governments are interested in economic performance and use this as the basis for
promoting managers of state-owned enterprises (SOEs) to local government positions.
During the economic reforms of the 1970s and 1980s, SOE ownership was decentral-
ized from the central to sub-national governments at all levels. SOEs have enormous
importance to local governments: they provide public goods such as schools and hospi-
tals, and they are also a considerable source of revenue. In addition to tax payments,
the managers of SOEs can transfer part of their profits to the local government to
increase their chances to be promoted into the ranks of the local government. At the
same time, SOEs are monitored by central government agencies in terms of their en-
vironmental performance. For instance, the Ministry of Environmental Protection is
in charge of environmental regulation and its enforcement. Responsibility for natural
resources is with the Ministry of Land Resources, the Ministry of Water Resources,
and the State Forestry Administration (SFA). If managers are found to fail to meet
ecological targets, they can be fined. Although the central government authorities have
sub-level bureaus responsible for the enforcement of their respective ecological targets,
these local bureaus are not superior to local governments but work independently.
Overall, the managers’ incentives to protect the environment seem to be outweighed
by their career concerns.
We focus on state-owned forest enterprises (SFEs), which are owned by local govern-
ments but whose forests are owned and monitored by the central government through
the SFA. In this context, the difficulty of monitoring the SFEs’ and their managers’ en-
1
vironmental performance is due in part to the size and volume of the managed forests.
We interpret the complex relationship between an SFE manager, the local government
and the SFA as a two-principal, one-agent framework, in which the SFA has only limited
information on the manager’s environmental performance and the manager cares about
being promoted. Establishing a simple political career concern model with asymmetric
information, we derive three hypotheses. First, a larger degree of asymmetric informa-
tion between the SFA and the SFE, as measured by the forest area and volume, as well
as the area and volume share of the natural forest, increases the amount of logging and
thus leads to more deforestation. The reason that these variables are correlated with
the degree of asymmetric information between the SFA and the SFE is that it is more
difficult, and also more costly, to conduct inspections in larger forest areas and in areas
with more forest in terms of volume. Furthermore, in natural forests, the application
of selective logging makes it harder to detect whether too much cutting has been done.
Second, if a manager transfers a higher percentage of profits to the local government,
the logged volume increases (or decreases) and thus induces more (or less) deforesta-
tion. The intuition for why this effect is ambiguous in sign is that, on the one hand, a
higher transfer rate directly increases a manager’s chances of getting a promotion (an
assumption we confirm with the next hypothesis), which makes cutting more attractive
to the manager. On the other hand, if a manager transfers a higher percentage to the
local government, cutting can be reduced. This will, of course, decrease profits but the
corresponding drop in the chance to get promoted is offset by the higher transfer rate.
Third, the likelihood of an SFE manager getting promoted increases with the share
of profits transferred to the local government as well as with the absolute amount of
transferred profits. The last hypothesis sets out to test the assumptions that we have
made in the theoretical model.
Using data on state-owned forest enterprises (SFEs) in Northeast China over the
period 1980-2009 and combining it with satellite imagery data on vegetation growth
as a measure of deforestation, we test these hypotheses. We find support for the first
hypothesis and show that the logged volume increases with a higher rate of transfers
from the SFE to the local government (second hypothesis). Finally, we find evidence in
support of the first part of the third hypothesis (that the share of profits is a determi-
nant of a manager’s likelihood of getting promoted) but not in support of the second
part (that the absolute amount of transferred profits has a significant and positive
effect on this likelihood). Our results suggest that the information asymmetry with the
SFA is exploited by SFE managers and that higher transfers to local governments go
hand in hand with more deforestation. On the other hand, our findings also suggest
that it is rational for SFE managers to transfer a higher percentage of profits to the
SFE’s owner – the local government – as this does indeed increase their likelihood of
2
getting promoted.
This paper highlights the managerial incentive problem regarding environmental
protection in this Chinese-style fiscal federalism. It aims to explain how, in such
Chinese-style federalism, the SOE managers lack incentives to protect the environ-
ment. The paper draws extensively on the growing literature on managerial incentives
and their impact on performance and societal welfare (e.g., Besley and Coate, 1997,
2003; Acemoglu and Johnson, 2005; Xu, 2011). It also draws on the Chinese fiscal
federalism literature (Qian and Xu, 1993; Lin and Liu, 2000; Maskin et al., 2000; Jin
et al., 2005; Xu, 2011), in that career concerns not only motivate government officials
to compete among each other but also motivate the managers of SOEs to over-harvest
resources or to over-pollute.
The impact of fiscal federalism on economic growth has been studied intensively,
but not sufficiently with respect to the impact on environmental quality and resources.
Zhang et al. (2011) study this relationship for China, using provincial data on carbon
emissions from 1998 to 2008. They find that pollution is positively correlated with
fiscal decentralization. Li and Chan (2016) compare firms of different ownership types
and find that SOEs spend less on pollution abatement, but they do not find different
environmental consequences among SOEs and private or foreign-funded firms. Indeed,
the incentives that SOE managers face are very similar to those in Western corporations
(Murphy and Zimmerman, 1993; Weisbach, 1988). The most closely related paper
is Groves et al. (1995), who show that local governments rely mostly on economic
performance criteria for the selection of SOE managers.
Overall, we add to the large body of literature on managerial incentives and a vari-
ety of principal-agent models (e.g., Tirole, 1986; Thomas and Worrall, 1990; Holmstrom
and Milgrom, 1991, 1994; Fehr and Schmidt, 2004; Ward and Filatotchev, 2010; Dela-
cote et al., 2014) with a dual-principal, one-agent model with asymmetric information
between one principal and the agent, with the agent possessing information that one
of the two principals does not have.
The paper proceeds as follows. Section 2 introduces Chinese-style federalism, the
importance of state-owned enterprises, the system of environmental authority, and the
implications for the state forest sector. Our empirical analysis is grounded on this sec-
tor. In Section 3, we outline a simple theoretical framework that captures the complex
relationship between an SFE manager, his jurisdictional government and the SFA (we
use male gender because all managers in our surveyed SFEs are males). Section 4
formalizes testable hypotheses based on the theoretical predictions and specifies econo-
metric models. Section 5 describes the construction of the data sets and Section 6
presents the econometric results. Section 7 concludes.
3
2 Background
In the following sections, we explain the features of “Chinese-style” federalism, the
importance of state-owned enterprises, the system of the environmental policy-making
and enforcement, and the implications for the state forest sector.
2.1 Chinese-style Federalism
China’s economic reforms, which began in 1978, transformed China from a highly
centralized to a decentralized, market-oriented economy (Montinola et al., 1995; Xu,
2011). In the 1980s, authority over jurisdictional economies was decentralized to local
governments. In this process, sub-national governments – at provincial, prefecture,
county and township levels – were endowed with control rights over land, firms, financial
resources, energy, raw materials, etc. At the same time, all sub-national governments
obtained full fiscal autonomy, which secured tax revenues from “grabbing-hand” upper-
level governments.1 Based on this autonomy, a fiscal transfer system was established
which allows each sub-national government to contract with the next higher level.2
The decentralization in China is different from other countries for two reasons. First,
typical elements of Western federalism are absent: the protection of individual rights;
strong and explicit constitutional foundations; political freedom; and the right of rep-
resentation and democratization (Montinola et al., 1995; Qian and Weingast, 1997; Xu,
2011). Second, despite decentralization in economic terms, there is strong political cen-
tralization by means of personnel control. Through control over promotion decisions,
the central government provides provincial officials with incentives to adhere to cen-
trally proclaimed goals but, also, because promotion decisions are taken mainly based
on economic performance, with incentives to compete as if they were in a tournament.
However, as we will argue below, personnel control applies not only to officials but also
to the managers of state-owned enterprises, who can be promoted into the ranks of
local government officials. It has been argued that this “Chinese-style” federalism is
one of the key determinants of China’s economic success (Qian and Weingast, 1997).
1 This stands in stark contrast to some post-Soviet Eastern European countries, in which the central
authorities retained discretionary authority over the firms’ profits and bailed out firms with deficits; see
Montinola et al., 1997.
2 The fiscal contract system (1980-1993) and revenue assignment system (since 1994) allow each sub-
national government to contract with the next level up. They decide on the remittance amount (or share) of
tax and revenue for a certain number of years. The lower level keeps the rest and enjoys sole discretion to
allocate the retained revenues (World Bank, 1993).
4
2.2 Importance of State-Owned Enterprises
State-owned enterprises (SOEs) have played an important role in China’s development.
By the end of the 1970s, SOEs accounted for 80 percent of the total national value of
gross industrial output (NBSC, 1999). In 1998, SOEs managed almost 70 percent of
the total national assets, and in the 2010s this number is still at 40 percent (NBSC,
2015).
The SOEs were initially established to fulfill development and construction needs,
such as the production of timber, coal, cement and steel, agricultural machinery, au-
tomobiles and tractors (Xu and Zhuang, 1998). Managers were initially appointed by
government officials of, e.g., industrial bureaus. Prior to the economic reforms in 1978,
the central and regional planning commissions controlled the entire industrial system.
Managers of the SOEs had to follow the orders of the (upper level) political leaders
and to provide services to the government (Groves et al., 1995). Because the areas
in which the SOEs were located were sparsely populated, the SOEs were also obliged
to supply social services, such as housing and schooling, health and child care, and
pension benefits, to employed families.
In the 1980s, enterprise reforms, accompanied by various price and market reforms,
introduced material incentives for SOE management. Simultaneously, SOE ownership
was decentralized from the central to sub-national governments at all levels. Managerial
contracts now commit the managers to specific performance targets, such as profitabil-
ity. After transferring taxes and an agreed-upon share of profits to the local government
(which is prescribed for the duration of the contract), managers enjoy extensive auton-
omy in using the retained profits for firm investments and capacity expansion (Nolan
and Wang, 1999). They may also transfer more than the contractually required amount
to the local government in order to signal their ability. We will, later on, assume that
the managers can freely choose the amount to transfer. The contracts effectively link
managers’ rewards and careers to firm performance. The contracts generally have a
duration of three or five years (Groves et al., 1995).
SOEs have enormous importance to local governments: they provide public goods
such as schools and hospitals, and they are also a considerable source of revenue.
Nevertheless, SOEs often are very pollution-intensive and depleting natural resources,
because they mostly belong to the industrial sectors of iron and steel, machinery con-
struction, mining, automotive, and manufacturing, as well as the processing of wood,
leather and chemical products.
5
2.3 Environmental Protection: Authority and Enforcement
This section reviews the system of environmental authority and enforcement in which
SOEs operate. First of all, the enviornmetnal legal framework is well-developed in
China, including 22 laws and over 44 regulations, 500 standards and more than 600
other legal documents addressing pollution control, natural resource protection and
environmental regulation of consumer products.3 As mentioned in the introduction, the
authority regarding the environment lies with the central government and is delegated
to several distinct ministries. The Ministry of Environmental Protection (MEP) is
in charge of environmental regulation, while the Ministry of Land Resources (MLR),
Ministry of Water Resources (MWR) and State Forestry Administration (SFA) are
in charge of natural resources, which are mostly owned by the central government.
Their sub-level bureaus are designed to work independently and undertake monitoring
and inspection activities to make sure that SOEs and other firms comply with the
environmental regulations.
In the National Pollutant Discharge Reporting and Registration Program, as an
example, each firm reports its input uses of raw materials, its pollutant discharges,
and its environmental management. Using the self-reported data, local environmental
protection bureaus aggregate the data and submit it to the MEP to compile a national
pollution register. In reality, monitoring is very incomplete, and few environmental
agencies have accurate data on firms’ actual emissions and use of natural resources
or raw materials. Likewise, in the Pollutant Discharge Standard, the pollution charges
on 65 kinds of water pollutants and 44 kinds of air pollutants (MEP, 2003) are set
by the central government (i.e., the MEP), but, in practice, the actual payments by
enterprises are limited, not only by difficulties in measurement and monitoring, but
also by the “protective umbrella” of local governments (OECD, 2009).4
2.4 Implications for the State Forest Sector
This section introduces the empirical context of our analysis: the state-owned forest
sector and its SOEs. State forests make up 42 percent of China’s total forest area and
68 percent of the total national forest stock.5 Since the 1950s, 135 SFEs have been
3 See, for example, the website of the Department of Policies, Laws and Regulations, Ministry of Environ-
mental Protection, P.R. China: http://zfs.mep.gov.cn/.
4 According to anecdotal evidences, except for a small number of severe violations against criminal law, the
sub-bureaus seldom disclose enforcement decisions. It also seems that they often go with the local governments’
preferences regarding the trade-off between development and environment (Lo et al., 2006).
5 Chinese forests, which are the fifth largest in area world-wide (208 million hectares), account for merely 22
percent of the total land area, which is below the 30.8 percent global average (FAO, 2016). They are unevenly
distributed, with the majority located in the South (about 68 million ha) and 43 million hectares located in
the Northeast, 83 percent of which are natural forests (Xu, 2013), in which no sowing or planting is allowed.
6
established to meet the growing demand for timber in China’s development process.
The SFEs manage the forest-rich regions in the Northeast, Northwest and Southwest,
and their main business is timber extraction.
Before the reforms of the 1980s, the SFEs, like other SOEs, were under the control
of the central government, specifically, the Ministry of Forestry. With the economic
reforms in the 1990s, the ownership of the SFEs was devolved to sub-national govern-
ments, which were given great autonomy in appointing managers and making financial
and production decisions.6 In 1998, the Ministry of Forestry was downgraded to be-
come the SFA, which still maintains authority in forest governance on behalf of the
central government. The SFA thereby represents the central government as the for-
est owner. The monitoring bodies of the SFA are State Forest Resource Monitoring
Offices (SFRMOs) for timber production and forest protection; they also implement
the national forest policies. For example, the Forest Law from the mid-1980s required
reforestation after commercial harvest, and a logging ban associated with the Natural
Forest Protection Program (NFPP) was introduced in 1998 (Xu, 2013).
The managers of SFEs have commitments to the SFA in terms of sustainable use of
forests and commitments to their jurisdictional governments based on contracts that
specify revenue-sharing and social obligations. They are, of course, free to over-fulfill
these targets to please the local government. The goals written in the contracts –
profitability, job creation, payment of pension benefits and provision of other social
benefits, and profit-sharing – are relatively easy to monitor. By contrast, the ecological
targets on forest protection are more difficult to measure and do not affect human
welfare immediately; the impact of over-harvesting on forest degradation is cumulative
and takes a longer time. SFE managers are thus reasonable in prioritizing economic
performance over natural resource protection, leading to over-harvesting and forest
degradation (Xu et al., 2004).
3 Theoretical Framework
As described in the previous section, the management and monitoring of Chinese state-
owned forests involves a complex interplay of three actors: the SFA, the local govern-
ments, and the SFEs, represented by their managers. The managers have career con-
cerns and care about being promoted, while they also have to make sure that ecological
targets imposed by the SFA are not violated (at least not by too much) - else, they will
be fined. We interpret this complex relationship as a two-principal, one-agent relation-
ship, with the principals being the SFA and the local government, and the agent being
6 The SFEs were mainly devolved to provincial governments, and also to prefecture- and county-level
governments.
7
the manager of an SFE. We will argue that one of the principals (the SFA) can only
imperfectly observe the agent’s effort with respect to meeting the ecological target,
while the other one (the local government) decides upon his promotion.
In what follows, we characterize the relationship between a representative SFE
(through its manager), the SFA and a representative local government. In particular,
we describe the SFA’s and local government’s objectives and interests, and then turn
to the description of the manager’s payoffs and maximization problem. We take the
agent’s contract as given and will not characterize the optimal contract here. This
would be beyond the scope of this paper.
3.1 SFA and Local Governments
As the owner of all state forests, the central government through the SFA wants to
ensure that the forests are able to fulfill their ecological functions. Therefore, it sets an
upper limit to the total allowable harvested volume, x¯. Let X be each SFE’s overall
forest volume. Then X − x¯ is the protected volume, which may not be logged. We
assume that SFEs always log a volume x which is at least as large as x¯, i.e., the “cap”
on legally logged forest is binding. To enforce this cap, the SFA or its sub-bureaus
make monitoring visits in each SFE. However, due to the large area of the forests and
the huge administrative cost that a complete inspection would require, the monitoring
technology is imperfect. We assume that the SFA only receives the signal xs about the
total logged volume (where the superscript “s” stands for signal):
xs = µ(x) +  ,  ∼ N (0, σ2) . (1)
µ is assumed to be twice continuously differentiable and strictly concave, and  is
normally distributed with mean zero and variance σ2.
The massive floods of the Yangtze River, the Nen River, the Songhua River and
the Pearl River in the summer of 1998 convinced the SFA to implement the Natural
Forest Protection Program (NFPP). The main policy instrument of this program is a
logging ban on most commercial logging in order to increase the capacity of the forests
to absorb rainwater. However, logging is still allowed in plantation forests and for forest
tending. To compensate the SFEs for the logging ban and the associated reduced profit,
the SFA supports them with an annual transfer of T .
The local government gets twofold benefits from the SFE’s gross profit piG. First, it
receives corporate income tax payments, which we model as a profit tax. These amount
to τpiG, where 0 < τ < 1 is the tax rate. Second, it receives an additional share of
the profit as a transfer from the SFE, amounting to γpiG, where 0 ≤ γ ≤ 1. Whereas
8
the tax rate is exogenously given for the SFE, the transferred amount can be chosen
by the SFE through its manager.7
3.2 SFEs and Managers
Each SFE cuts a volume x of timber and sells at the exogenously given price P . Cutting
costs are C(x), with C′(x), C′′(x) > 0. The profits, net of taxes and net of the transfer
T received from the SFA for offsetting the financial implications of the logging ban after
1998, are given by:
pi = (1− τ )piG + T , (2)
where gross profit is piG = Px − C(x). From these profits, transfers to the local
government can be made, and social welfare benefits G > 0 to residents in the local
communities of an SFE must be provided. The latter consist of unemployment benefits,
health insurance, etc., and are exogenously given in our model. After transfers, taxes
and the provision of social welfare benefits, the remaining profits have to satisfy:
(1− γ)pi −G ≥ 0 . (3)
In other words, the transfer to the local government cannot be set so high that this
inequality is violated, i.e., γ is always strictly smaller than unity.
The manager earns a fixed wage w¯. Beyond that, he may be promoted into the
ranks of local government officials based on economic achievements, and may also be
fined for failing to meet ecological targets. If promoted, the agent receives economic
benefits from the time of being promoted until retirement (and possibly beyond). The
net present value of a promotion (compared to the earnings profile of continuing to work
as a manager) equals b. We model the likelihood of being promoted as a continuous
variable which depends on the amount of transfers that are voluntarily handed over to
the local government. Either a higher share γ or a higher profit pi (or both) will thus
increase the agent’s likelihood of receiving a promotion. The likelihood reads:8
Pr(b) = 1− e−γpi ∈ [0, 1) . (4)
If fined, the agent will suffer an income loss equal to f . The likelihood of being
caught cheating upon the ecological target positively depends on the deviation of the
7 Before the manager signs a contract for a certain number of years, individual targets for the transfers
are negotiated. We conjecture that managers have an incentive to over-fulfill their targets and strategically
negotiate lower targets in order to be able to demonstrate their good performance.
8 This specification of the probability function is chosen to analytically separate the variance component
of the signal in the agent’s payoff function.
9
signal from the ecological target, xs − x¯, and is equal to
Pr(f) = 1− e−(xs−x¯) ∈ [0, 1) . (5)
The manager chooses the share of pi transferred to the local government, i.e., τ , and
the logged volume, x (where we assume that x ≥ x¯ always holds), so as to maximize
his payoff
EU = E [w¯ + b(1− e−γpi)− f(1− ex¯−xs)] . (6)
Plugging in (1) and considering that E [ek] = ek2σ22 for any constant k when  ∼
N (0, σ2), expected utility can be written as:
EU = w¯ + b(1− e−γpi)− f(1− ex¯−µ(x)−σ22 ) . (7)
The first-order conditions with respect to γ and x read as follows:
bpie−γpi > 0 , (8)
bγ(1− τ )[P − C′(x)]e−γpi − fµ′(x)ex¯−µ(x)−σ22 = 0 , (9)
which implies P − C′(x) > 0 at the optimum.
The first condition is never fulfilled with equality. Instead, a corner solution is
chosen, i.e., γ is chosen as high as possible so that equation (3) is still satisfied. The
agent thus does not gain from transferring less to the local government than pi − G,
which increases his chances of getting a promotion. The second condition trades off the
marginal benefits of cutting more timber (first term in equation (9)) with the marginal
costs of doing so (second term). The marginal benefits are the increased likelihood of
receiving a promotion because profits have increased. At the same time, the agent runs
a higher risk of being fined because the likelihood of being detected cheating increases
with cutting (marginal costs).
In general, it cannot be shown that the maximization problem is concave for all pa-
rameter constellations. We assume that it is concave for empirically relevant parameter
constellations.
Applying the implicit function theorem to equation (9), we derive the following
10
comparative statics results:
∂x
∂x¯ =
fµ′(x)ex¯−µ(x)−σ22
Γ > 0 , (10)
∂x
∂T =
bγ2(1− τ )[P − C′(x)]e−γpi
Γ < 0 , (11)
∂x
∂b =
−γ(1− τ )[P − C′(x)]e−γpi
Γ > 0 , (12)
∂x
∂f =
µ′(x)ex¯−µ(x)−σ22
Γ < 0 , (13)
∂x
∂σx =
−fσµ′(x)ex¯−µ(x)−σ22
Γ > 0 , (14)
∂x
∂γ = −
b(1− τ )[P − C′(x)]e−γpi [1− γpi]
Γ R 0 , (15)
where Γ ≡ ∂2EU/∂x2 < 0 by the assumed concavity of the maximization problem.
The interpretation of the comparative statics goes as follows: A marginal increase
in the ecological target x¯ increases the amount of logged timber due to the less binding
constraint. If the transfer T from the SFA to the SFE marginally increases, the manager
cuts less timber because he can maintain the same or a similar level of profits with
less logging. Therefore, ceteris paribus, less logging does not decrease the manager’s
likelihood of getting a promotion. Marginally increasing b or marginally decreasing f
has qualitatively the same effect: more timber will be cut.
If the signal with respect to x becomes more blurred, i.e., σ marginally increases,
the agent can cut more without having to fear that he will be caught with a higher
probability. Interestingly, if the transfer rate γ goes up, the reaction of x is ambiguous
in sign. On the one hand, a higher γ directly increases the marginal benefits of cutting
more forest because a higher transfer of funds to the local government, given a certain
profit, increases the likelihood of being promoted. For this reason, the agent would like
to cut more. On the other hand, profits can be decreased by cutting less trees while
still having the same probability of being promoted due to a higher γ. Therefore, the
agent would like to cut less.
4 Empirical Strategy
In this section, we state our hypotheses based on the predictions from the theoretical
model and lay out our empirical strategy to test them. We will test the hypotheses
with a focus on the northeastern state-owned forests in China.
11
4.1 Hypotheses
We base our empirical analysis on the following hypotheses, with two of the predictions
coming from the comparative statics results above.
Hypothesis 1: Logging and asymmetric information .
A larger degree of asymmetric information between the SFA and the SFE, as measured by the
forest area and volume as well as the area and volume share of the natural forest, increases
the amount of logging and thus leads to more deforestation.
This hypothesis is in line with equation (14). The idea is that the larger the SFE’s
forest area and volume, the harder and more costly it is for the SFA and its sub-level
bureaus to monitor that SFE’s logging activities. This claim is also backed by the
literature on public harvesting and concession contracts, as well as anecdotal evidence
on Chinese SFEs (Poore, 1993; Gray, 2002; Palmer, 2000; Johnson, 2002; Xu et al.,
2004). SFEs with larger areas of forests, especially natural forests, can more easily hide
their operations from SFA inspectors, including logging beyond the specified quotas.
Furthermore, in natural forests, in which no sowing or planting is allowed, it is less
obvious how much cutting was actually done. The reason for this is that natural forests
are usually less regular in their structure than plantation forests, and inspection of
selective logging in mostly natural forests is useless, because largely intact canopy will
not look very different from selectively logged canopy when viewed by the human eye
from some distance.
Hypothesis 2.1: Logging and voluntary transfers: positive correlation .
If a higher percentage of profit is transferred to the local government, the manager of an SFE
increases the logged volume x and thus induces more deforestation.
Hypothesis 2.2: Logging and voluntary transfers: negative correlation .
If a higher percentage of profit is transferred to the local government, the manager of an SFE
decreases the logged volume x and thus induces less deforestation.
The above two hypotheses are obviously related. While the first one corresponds to
equation (15) being positive, the second one corresponds to the same equation being
negative. Because the theory does not give a clear prediction regarding the sign of this
effect, we split the hypothesis into two sub-hypotheses.
Hypothesis 3.1: Likelihood of being promoted and share of profits transferred
to local government .
The likelihood of an SFE manager getting promoted increases with the share of profits trans-
ferred to the local government, i.e., ∂Pr(b)∂γ > 0.
Hypothesis 3.2: Likelihood of being promoted and absolute amount of profits
transferred to local government .
The likelihood of an SFE manager getting promoted increases with the absolute amount of
12
transferred profits, i.e., ∂Pr(b)∂pi > 0.
The last two hypotheses come from our assumption about the probability function
of getting promoted. We want to test whether the share and the absolute amount
of transferred profits matter for the likelihood of promotion. In contrast to the other
hypotheses, this will shed light on the other side of the coin: how managers’ economic
and environmental performance is actually evaluated by the SFEs’ owners, the local
governments.
4.2 The Role of Asymmetric Information in Deforestation
The first hypothesis is tested by estimating a simple reduced form relationship between
information asymmetry and deforestation:
H = f(θ,V) + u, (16)
where H is deforestation or the rate of change in forest vegetation from one year to
the next, respectively. θ represents the difficulty of measuring the manager’s effort and
thus the degree of asymmetric information between the SFA and the SFE (σ in the
theory part). As discussed previously, as proxy variables for this asymmetry, we take
the total forest area as well as the total forest stock volume, and area and volume
shares of natural forest.
V denotes control variables that affect H, and u is a disturbance term. The vector
V contains, first, the characteristics of natural forests, including their area and volume
shares, because they constitute an important part of the total forests managed by an
SFE. Given their more diversified ecosystem services, natural forests face more restric-
tions on cutting; in fact, higher demands are placed on the protection of natural forests
because of their ecological importance. Second, the vector contains the afforested area,
including areas newly afforested and reforested in the current year and their shares of
the total forest land area. We include the latter variable because future vegetation will
differ by the various levels of afforestation and post-harvest reforestation. Third, we
include SFE characteristics such as population density and employee structure, that is,
the ratio of employees to supported laid-off and retired people. Fourth, we control for
manager characteristics because a manager takes actions that affect the enterprise’s
productivity and thus the forest stock. This includes manager age, education, and
managerial experience. In addition, regional demographics and economic factors are
included in V , such as population and GDP, ecological pressure from agricultural and
urban populations, influences from the non-farming sector, and the timber price. We
also control for whether the previous manager was promoted, and whether he defor-
13
ested, as well as the interaction term of these two.
We employ a linear panel fixed effects model to estimate equation (16), controlling
for year trends, fixed and year effects related to upper bureaus. The panel fixed effects
model allows us to capture the influence of time-invariant factors. One such important
factor is historical importance, measured by the contributions to social and economic
development, which might give some SFEs greater bargaining power, so that its logging
activities have low probabilities of punishment. The year dummy variables control for
weather changes and, in particular, for the influence of, for example, a drought year
on the vegetation values. The upper bureau fixed and year effects enable us to capture
time-varying factors, including changes in technology and equipment efficiency at the
upper level bureaus. The upper level bureaus of the SFEs belonged to the central
government prior to the economic reform and served as a government body which
allocated tasks and provided equipment to the SFEs. With the economic reform, many
SFEs have been delegated to the local government while their upper level bureaus
still serve as their parent enterprises. These upper level bureaus are not the same
sub-bureaus that are responsible for monitoring and inspection of the forests, as we
introduced in Section 2.
4.3 Political Career Interest and Environmental Protection Effort
We test the remaining two sub-hypotheses by exploring the determinants of a manager’s
promotion into the ranks of local or upper level governments. We focus on whether or
not sub-national governments in charge of selecting firm managers were driven by
both environmental protection effort and the voluntary transfer of profits. We estimate the
relationship between various characteristics of the incumbent manager of an SFE and
his fate when his tenure as the manager was at an end (i.e., at the end of his current
managerial contract):
Pr[Y = 1|z, V ] = Φ(α+ β′z + δ′V). (17)
The dependent variable is the likelihood of an SFE manager being promoted to the
local (or upper level) government. Φ is a standard normal cumulative distribution func-
tion. z denotes the decision variables of each SFE manager – the percentage of profits
transferred, natural logs of the transferred amount, yearly average and aggregated
transfers, and environmental protection efforts, i.e., the total afforestation area and
the area in proportion to the total forest land area. V is a vector of manager personal
and SFE characteristics, as discussed for equation (16). α, β and δ are parameters to
be estimated.
14
In the V -vector, a manager’s age, education, and expertise in management seem to
be important evaluation criteria, in addition to political loyalty (Li and Zhou, 2005).
Expertise is correlated with the manager’s economic performance and political loyalty
is reflected by the share of profits transferred to the local government. We also control
for the manager’s number of years to retirement and for his education in the estimation
of equation (17).
Table 1 provides the definitions of all variables used in the estimation. The variables
in the vector V are characteristics that may also influence H and Pr[Y = 1] when we
estimate the equations (16) and (17), respectively. Data sources, summary statistics,
and estimation methods are described in the next section.
< Table 1 here >
5 Data
The data used for analysis comes from the economic survey data measuring various
characteristics of 24 randomly surveyed SFEs and their respective managers, which are
combined with satellite data measuring vegetation as a measure of deforestation.
5.1 Satellite Data Set
Given the difficulty of obtaining truthful reporting of deforestation by SFEs (Alford and
Shen, 1998; Brandt and Zhu, 2000), we develop an independent measure of deforestation
by using satellite imagery from two NASA satellite sensors – the AVHRR sensor carried
on the Polar Orbiting Environmental Satellites (POES) and the MODIS sensor on
board the Terra and Aqua satellites (Tucker et al., 2010; NASA LP DAAC, 2015). In
particular, we combine the Normalized Differentiated Vegetation Index (NDVI) (1981-
2002) and the Enhanced Vegetation Index (EVI) (2003-2009) to construct an annual
measure of forest quality and quantity for each year. Using this measure, we are able
to encompass both legal and illegal logging undertaken in the SFEs.9
The NDVI is the ratio of the difference in reflectance of sunlight, ρ, between the
near-infrared (NIR) and red bands (visible radiation), divided by the sum of near-
infrared and red band radiation (Tucker 1980; Sellers 1985; Sellers et al., 1994):
NDV I = ρNIR − ρREDρNIR + ρRED . (18)
9 Illegal logging includes both over-harvesting by the SFE itself and unlawful logging of forests, by the
rural population, for example. An SFE is responsible for both within its management area.
15
Calculations of NDVI for a given pixel result in a number ranging from −1 to +1.
No green leaves results in a value close to zero, while a value close to +1 indicates the
highest possible density of green leaves. In other words, unhealthy or sparse vegetation
implies low values of NDVI, and healthy or abundant vegetation implies high values of
NDVI.
We obtain a generalized annual measure of the state of the forest in the area of an
SFE by taking the mean value of the vegetation index over all pixels of size 250 by
250 meters from 1981 to 2009. Deforestation is then measured as a negative annual
rate of change in vegetation values. The latter measure of deforestation differs from
Burgess et al. (2012) in that we assess the forest change observed at an SFE average,
while they treat a pixel inside a village as being deforested if 90% of the area is cleared.
Thus, their method captures only large-scale changes such as observed clear-cutting.
Selective logging, by contrast, changes the forest canopy moderately and is not captured
in their estimates. Given that selective logging is a reasonable logging method and is
usually adopted by forest managers over medium and large scales of forests, the average
measure of forest change in our study will provide an objective estimate of an SFE
manager’s environmental performance.
Due to data availability, we combine the NDVI data with EVI data after 2002.
EVI is an optimized vegetation index. It accounts for and reduces background and
atmospheric noise. It is also more responsive to canopy structural variations, canopy
type, plant physiognomy, and canopy architecture (Huete et al., 2002). In contrast to
the NDVI, the EVI is computed according to the following equation:
EV I = 2.5× ρNIR − ρREDρNIR + 6ρRED − 7.5ρBLUE + L. (19)
where ρΛ, Λ ∈ (NIR,RED,BLUE), is atmospherically-corrected or partially atmosphere-
corrected surface reflectance, and L is the canopy background adjustment, which ad-
dresses non-linear, differential NIR and red radiant transfers through the canopy.
The final outputs of forest change data are values of vegetation index rescaled to
be bounded between -1 and 1 for each pixel and averaged over all pixels for each
SFE. We match the pixels with the maps of areas of the 24 SFEs that were surveyed;
the description is in the next section. As for the years available in the survey data,
vegetation change refers to the year pairs 1980-81, 1985-86, 1990-91, 1995-96, 2000-01,
2004-05 and 2008-09. Thus, the switch from NDVI to EVI in 2003 does not cause any
data inconsistencies in our forest change measures because we do not compare NDVI
data with EVI data – only NDVI with NDVI data and EVI with EVI data. Figure
1 illustrates the study area – maps of the 24 SFEs. Figure 2 codes the pixels with
vegetation indices in Northeast China in the study period.
16
< Figures 1 and 2 here >
5.2 Economic Survey Data Set for State Forest Enterprises
Our data on 24 SFEs in Northeast China comes from a survey conducted by the Envi-
ronmental Economics Program in China (EEPC) at Peking University. The northeast-
ern state forests are located in three provinces (Heilongjiang, Inner Mongolia and Jilin)
and are managed by 84 SFEs in total. These SFEs played an important role in the
old industrial base in this area and contributed to China’s growth by providing more
than half of the national timber production in the past five decades. However, this
enormous contribution to social and economic development has come at the expense
of environmental degradation. By the beginning of the 2000s, 60 out of the 84 SFEs
had almost depleted their mature forests (Xu, 2013).
To study the institutional system of state forest management and provide a roadmap
for feasible reform, EEPC researchers conducted surveys in 24 randomly selected SFEs.
The first wave survey was conducted from June to October 2005, and in summer 2009
the SFEs were revisited, and the earlier process of data collection was repeated. For each
SFE, data was collected for the period 1980-2008 on logging and reforestation activities,
employment structure, income and expenditure, and investments in other assets. The
data on forest resources and the SFEs’ social services, tax payments and transfers
to local jurisdictional governments, as well as on the timber price, was provided by
provincial forestry authorities (the sub-level bureaus of the SFA).
In addition to the average values of our constructed vegetation indices – the annual
rate of change between two consecutive years and the deforestation dummy – Table 2
reports the mean values of socio-economic variables for the SFEs and the counties in
which the SFEs are located, as well as the SFE managers’ characteristics, during the
period 1980-2008. The availability of data on the transferred amount of profits allows
us to have nine observations in 1980, 13 in 1985, 15 in 1990, and around 20 for the
remaining years. Overall, the observed SFEs reported a stable total land area but forest
stock volumes fluctuated by more than 10 percent in the 1990s and the early 2000s. An
annual deforestation rate of three to five percent was observed in 1990, 2000 and 2004
(which means that the rates of change in vegetation index values were negative in the
following years). In each year, at least one to four SFEs over-harvested their forests,
as shown by negative rates of change in vegetation value, and this figure rises to ten in
1990 and even more in 2000 and 2004.
In our sample, natural forests account for more than half of the forests in each
SFE. They decrease in volume on average over time, while plantation forests increase
in both area and volume. Following Xu et al. (2004), we use areas of forest land and of
17
natural forest, and area and volume shares of natural forest, as proxies for the degree
of asymmetric information between the SFA and the SFEs. Natural forests also differ
from plantation forests in their growth rates. Plantation forests are used for timber
production, and therefore fast-growing tree species are planted. Depending on climatic
conditions, natural forests grow at an annual rate of about two to three percent in the
SFE-managed areas. Natural and plantation forests also provide different ecological
functions. Estimating equations (16) and (17), we control for the growth of natural
forests in addition to their area and volume, also as a percentage of total forest.
As discussed earlier, the transfer of profits to jurisdictional local governments is
unique to China. The percentage of profits transferred, the absolute amount, and the
amount of timber production may signal the SFE manager’s ability. They are also sup-
posed to be among the evaluation criteria for his economic performance. We do not
focus on tax payments due to the fact that these are exogenously set. The transferred
amount of profits is subject to the manager’s decision, and we are interested in investi-
gating how the transfer decision determines a manager’s political career at the end of
his tenure as manager. The transferred amount ranges from 10 to 800 million CNY; as
a share of the SFEs’ disposable profit, the transfer ranges from less than ten percent
to as high as 90 percent. Timber production was very high in the beginning and de-
creased over time, with an average of 357,463 cubic meters in 1980, and dropped by 40
percent in 1995 to 215,827 cubic meters; after the implementation of the NFPP, there
was soon another 38 percent decrease, to 133,502 cubic meters in 2000, suggesting that
the logging ban was effective. Nevertheless, during the period with the logging ban,
the average timber production of the SFEs was still over 110,000 cubic meters in the
2000s. Because timber production is self-reported data and may be subject to mea-
surement error and reporting fraud, our econometric estimations do not include this
variable. Similar reasoning applies to the total harvested volume. Apart from transfers,
the yearly afforestation area, both in absolute amount and as a percentage of the total
land area, may serve as a signal for an SFE manager’s effort in forest conservation.
These two variables are observed to have an average decrease over time.
In addition to the above variables, a manager’s personal characteristics are impor-
tant. The average age of the managers is less than 50 years, varying from 34 to 60
years. The retirement age in China is 65 for males, and we conjecture that the proba-
bility of being promoted and hence the logging and transfer decisions may differ when
approaching this age. Therefore, we control for the number of years for an incumbent
manager until retirement age. We observe that the number of managers with higher
education significantly increased over the sample period, and all the observed managers
have a college education in the 2000s. Managerial experience, i.e., the number of years
as the manager of an SFE, is about 3.5 years on average, varying from one to 13 years.
18
For each manager, the promotion status is valued 1 if he was promoted to an upper
political position in a local or higher government after the tenure as manager was at an
end. For promoted managers observed in two or more periods, we recode their earlier
promotion status into 0 and leave only their last period of promotion equal to 1. In
each observed year except 2008, one to two thirds of the managers of the northeastern
state-owned forests were promoted at a later time.
Other variables that are important and may drive deforestation come from social
and economic pressures to harvest forests (Hyde et al., 1996; Amacher et al., 1998;
Rozelle et al., 1998; Rozelle and Huang, 2000; Foster and Rosenzweig, 2003). The
sources of these pressures include SFE population and social welfare payments (be-
cause laid-off and retired employees are required to be supported by the SFE)10, the
regional endowment with land and the share of agricultural population, urban pop-
ulation, population growth, income growth, and changes in the timber price. Among
these factors, SFE and local county population experienced first an overall increasing
and then a decreasing trend (since the late 1990s), while population density at SFEs
has kept rising. The social welfare payments by SFEs become less over time, as indi-
cated by the increasing ratio of employees to supported laid-off and retired persons.
The timber price during the period experienced an increasingly rising trend, due to
the booming timber markets and forest resource scarcity. At the regional level, pop-
ulation and GDP in counties increased, and the urban populations exhibited only a
slight decline. Agricultural population density was high in the 1990s and the local non-
farming sector experienced small fluctuations. The growth of agricultural and urban
populations may drive deforestation due to an increased need for agricultural land and
higher demand for forest products. The development of the non-farming sector, e.g.,
the manufacturing industries, may influence forest change directly through the growth
in processing related to forest and non-forest products, and indirectly by contributing
to regional economic growth. Therefore, we control for the variations in these factors
in the econometric estimations.
In total, we have a sample size of 24 (SFEs) for the seven year pairs for which we have
vegetation data, which yields 168. From this sample, we have to drop 54 observations
because of missing data on the transfers to the local government. We proceed with a
sample of 114 SFEs.
< Table 2 here >
10 The economic reform in China in the 1990s brought about a number of laid-off employees in state-owned
enterprises, some of whom receive a monthly allowance for subsistence, paid by the SOEs. The retired workers
are also supported by the SOEs.
19
6 Econometric Results
In this section, we present the econometric results of the examination of the previously
stated hypotheses regarding the impact of the information asymmetry on deforestation
and the role of personal interests in pursuing a future political career.
6.1 Deforestation and Information Asymmetry
Table 3 presents the estimation results of fixed-effects linear models on equation (16).
The dependent variable in Columns 1 and 2 is the rate of change of the vegetation index;
in Column 3, the dependent variable is the deforestation dummy. Column 1 focuses on
how the rate of change is affected by various measures of asymmetric information with
respect to monitoring by the central government, which is associated with different sizes
of the forests managed by an SFE. The models control for an SFE’s (natural) resource
structure, for managerial and regional characteristics, for last year’s vegetation value
and for the previous manager’s performance, as well as for year trends and upper
bureau fixed and year effects. The parameter estimates do not have statistical power
and are not reported.11
In support of Hypothesis 1, we find that the coefficients of the total forest land
area, the forest volume and the area share of natural forests are statistically significant
in Columns 1 and 2, and the total forest land area is also significant in Column 3. For
Columns 1 and 2, an increase in the information asymmetry as measured by our three
proxies leads to a smaller rate of change in the vegetation. This smaller rate implies
that the vegetation grows less on average, and the rate might even become negative
or more negative, indicating more logging activities and, as a result, deforestation.
Taking the marginal effect (-0.301) of the total forest land area on the rate of vegetation
change in Column 2 as an example, an SFE with a one percent larger area of forest land
compared to the average (256,200 ha) will have a 30.1 percent decrease from the average
rate of vegetation change in the following year (0.0183). This effect suggests that an
average-sized SFE being given a 10 percent larger area will experience a decrease of 5.5
percentage points in the average rate of change in vegetation, or an annual deforestation
rate of 3.7 percent in this area next year. This is a sizable effect in destroying forests.
Similar correlations hold for the other two asymmetric information proxies. If we only
look at SFEs that experience deforestation (Column 3), then deforestation is more
likely to happen in an SFE with larger forest land area, ceteris paribus. This lends
strong support to our Hypothesis 1, and is in line with findings by Xu et al. (2004).
In Columns 2 and 3, we add the variable of the percentage of profits transferred
11 The complete results are available upon request.
20
by the SFEs to the jurisdictional sub-national governments. The parameter estimate
of the variable does not have statistical power in explaining the variation in the rate
of change of the vegetation index in Column 2. However, it is statistically significant
at the 5 percent level when we look at the binary variable deforestation as the de-
pendent variable in Column 3. Column 3 is estimated by a linear probability model
with SFE fixed effects. The parameter estimate (0.963) implies that, when evaluated
at the avearge, an SFE manager whose transfer rate to the local government increases
by one standard deviation (0.288) leads to a stronger likelihood of having an overall
negative rate of change of the vegetation in his area by 0.277. This effect is equivalent
as 58 percent of the mean deforesatation rate (0.474). This finding provides support
for Hypothesis 2.1 but not for Hypothesis 2.2. While a higher transfer rate does, on
average, increase logging, it makes it more likely that forests are being degraded.
< Table 3 here >
6.2 Political Career Interest and Environmental Protection
The above analysis assumes that the jurisdictional sub-national governments are hap-
pier with more transfers and thus are more likely to promote the responsible manager
into the ranks of a government official. We now examine the determinants of an SFE
manager’s probability of getting a promotion to sub-national governments as a result
of these transfers, which we interpret as a better political relationship. Table 4 reports
the maximum likelihood estimation results of a probit model on equation (17) on the
promotion status of the incumbent SFE managers in the observed years.12 To allow
for heterogeneity across observations and control for SFE fixed effects across years, we
estimate the probit model with the SFE dummy variables and with the robust standard
errors option – adjusted at SFE level.
While no detailed evidence is available on how the local governments evaluate
and select SFE managers, we conjecture that a manager’s evaluation may depend
on multiple-year averages of transfers or cumulative performance rather than simply
on annual performance. In Table 4, the first three columns report the results of regres-
sions with the percentage of SFE profits transferred to local government, controlling in
Columns 1-3 for different variables that signal the manager’s performance and ability:
the natural log of the total transferred amount in the current observational year, the
multiple-year average transfer and the aggregated amount of the transfer. By doing
this, the effect of past performance is incorporated into the analysis of the determi-
12 A manager who had a tenure for five years or longer appears at least twice in our observations. This is
why we control for the number of years in office as a proxy for managerial experience.
21
nants of personnel promotion. Column 4 computes the marginal effects of a selection
of coefficients in Column 1 at sample means.
Consistent with Hypothesis 3.1, the percentage of transferred profits has a positive
impact on the probability of being promoted. The sign of the coefficient is positive and
significant at the 5 percent level in all columns. Moreover, the marginal effect of the
percentage of profits shared with the local government on a manager’s promotion prob-
ability is reasonably large. When evaluated at the mean of the independent variables,
it is 0.941, suggesting that, when an average-level manager increases the transfer rate
by one standard deviation (0.29) from the mean (0.288), his probability of promotion
will increase by 0.273 or 66 percent of the average promotion probability (0.412).
We do not find evidence supporting a positive relationship between the amount of
transferred profits and the probability of promotion and thus we reject Hypothesis 3.2.
For this, we consider current year and past average and cumulative transfers of the same
incumbent manager. In Columns 1 and 2, the coefficients of the transferred amount
are insignificant and have ambiguous signs; in Column 3, the cumulative amount has
a negative impact on the promotion probability at the 10 percent level of statistical
significance. One plausible interpretation for this result is that a bigger cumulative
amount, aggregated from past performance, would signify larger quantities of logging
and potentially cause trouble for the manager from environmental authorities. Overall,
this suggests that it is not the absolute amount of transfer that matters for the promo-
tion decision but rather the transfer rate. No matter the size of the cake, a manager
who hands over a larger share of the cake is more likely to get promoted. It seems that
the transfer rate is a better signal to the local government of the manager’s ability
than the absolute amount of the transfer.
Interestingly, an SFE manager’s efforts in environmental protection – measured
by the yearly afforestation area (log) and its share of the total forest land area – do
not have a positive effect on promotion as one might expect. First, a larger afforested
area has no significant effect in increasing an SFE manager’s promotion probability,
given the statistically insignificant coefficient in all columns. Second, however, in the
cases where the afforested area accounts for more than three percent of the total SFE
forest land area (sample mean), a standard deviation (seven percent) increase in the
total afforested and reforested area as a share of the total forest land area decreases
the probability of promotion for the manager by 0.296 or 71 percent of the average
promotion probability. It seems that afforestation and reforestation are not appreciated
by the manager’s superiors – on the contrary, more afforestation/reforestation may be
a bad signal to the local government because it shows that there must have been a lot
of deforestation previously.
Additionally, the point estimates are not influenced by adding manager character-
22
istics and the importance of the SFE, which is why we only report the table with all
controls. For the importance of an SFE, we consider the management scale in terms
of land area and population and the volume share of its natural forests. With these
variables controlled for, the regressions take into account various scales, where some
managers just need to do a little and can be easily promoted because the scale de-
termines the importance of the position of the SFE in the local economy and society.
Total land area and natural forest volume have the positive and significant impact on
promotion that we expected. Other economic and environmental efforts of the manager
equal, a manager with a one percent larger area than the average is more likely to be
promoted, by 3.3 percentage points (8 percent of the mean promotion probability of
0.412). Similarly, if the natural forest volume as a share of the total forest volume is
one percentage point higher than the average, the manager’s promotion is increased by
6 percentage points (15 percent of the mean probability of 0.412).
Among the manager characteristics of age, education and managerial experience,
we find that the younger a manager is, the more likely he is to get a promotion. This
effect is significant at the one percent level. The marginal effect of this is, however,
small in absolute value. Given that the mean age of the managers is 48, the mean num-
ber of years to the retirement age of 65 is 17 years. One year less from this mean (17)
increases his promotion likelihood by 0.072, or 17 percent of the average probability of
promotion. This result is consistent with other studies on political promotion of person-
nel from the local level of governments in China, e.g., Li and Zhou (2005). Equally or
more important than the incumbent manager’s own characteristics is the performance
of his predecessor, which has a significant impact on his promotion probability. In
particular, the promotion of one’s predecessor significantly decreases one’s promotion
probability, at the one percent level in all three columns of Table 4. The marginal effect
of the predecessor’s promotion on the incumbent manager’s promotion is -0.857. This
number means that, if the predecessor was promoted, this may raise the standard for
the incumbent manager to be promoted. It might become harder for the subsequent
manager to outperform his predecessor if the latter got promoted. By contrast, the
predecessor’s environmental performance in terms of deforestation, and the condition
that the predecessor both deforested and was promoted, do not always influence the
incumbent manager’s promotion probability in a statistically significant way, except in
Column 1. One plausible interpretation for this result is that earlier deforestation may
impede the incumbent’s performance due to the lack of resources available for logging;
however, a manager’s efficiency, especially in difficult situations like this, could also be
valued highly and considered in sub-national governments’ evaluation criteria.
In summary, our empirical findings lend support to the notion that: 1) an SFE
manager (the agent) utilizes the information asymmetry with the central government
23
(principal 1) to over-use natural resources; and 2), he satisfies the jurisdictional sub-
national government (principal 2) with a higher percentage of profits, which indeed
benefits his personal political career.
< Table 4 here >
7 Conclusion
In this paper, we identify a new channel through which natural resources such as
forest stocks get degraded. We establish a theoretical model to show that, in this two-
principal, one-agent relationship, managers of state forest enterprises take advantage
of the asymmetric information between the SFE and the SFA to increase logging and
thereby maximize profits and further their own careers. Their jurisdictional local gov-
ernments are driven by their own economic and career concerns. The principal-agent
relationship between the local government and the SFE manager drives the latter to
share a larger proportion of his profit cake with the former. This might even add to
the difficulty for the other principal – the SFA – to monitor its target if the local
government is able to protect an SFE from the SFA’s influence, which is suggested by
anecdotal evidence, as reported in Xu et al. (2004). For both the local government and
the SFE manager, environmental degradation plays a minor role because authority over
the environment does not lie with the local governments and the SFA has difficulties
in monitoring the exact amount of logged forest.
Our empirical analysis focuses on the state-owned forests in northeastern China,
using economic survey data and satellite imagery. We find that the rate of forest change
decreases and the likelihood of deforestation increases with the scale of forests, due to
the difficulty of measuring and monitoring. We also find that a larger share of the cake
to local government positively impacts the likelihood of deforestation. Futhermore, we
find that it is the share of the cake, not the size of the cake or the environmental protec-
tion effort, that determines an SFE manager’s political promotion. We suggest that the
degree of information asymmetry can be reduced through better monitoring technolo-
gies such as real-time satellite imagery, by transferring authority over forests to local
governments, which will then also be in charge of conservation, or simply by designing
incentive-compatible contracts for SFE managers with both of their principals.
Our theoretical model not only applies to the two-principal, one-agent relationship
in the state forestry sector but also to other sectors, e.g., the industrial SOEs which
have emissions as a by-product of output. In that case, x needs to be interpreted as
output, which is normalized in such a way that one unit of output causes one unit of
emissions, and thus x¯ is the maximum permissible level of emissions. However, it is
24
more difficult to develop an objective measure of emissions because satellites cannot
capture a factory’s emissions.
References
[1] Acemoglu, Daron, and Simon Johnson. 2005. Unbundling institutions. Journal of Political
Economy 113 (5): 949-95.
[2] Alford, W. P. and Y. Shen. 1997. Limits of the law in addressing China’s environmental
dilemma. Stanford Environmental Law Journal 16: 125-150.
[3] Amacher, G., W. Cruz, D. Grebner, and W. Hyde. 1998. Environmental motivation for
migration: population pressure, poverty, and deforestation in the Philippines. Land Eco-
nomics 74 (1): 92-101.
[4] Besley, T. and S. Coate. 1997. An economic model of representative democracy. The
Quarterly Journal of Economics 112 (1): 85-114.
[5] Besley, T. and S. Coate. 2003. Centralized versus decentralized provision of local public
goods: a political economy approach. Journal of Public Economics 112 (1): 85-114.
[6] Brandt, L., and X. Zhu. 2000. Redistribution in a decentralized economy: growth and
inflation in China under reform. Journal of Political Economy 108 (2): 422-439.
[7] Burgess, R., M. Hansen, B. A. Olken, P. Potapov, and S. Sieber. 2012. The political
economy of deforestation in the tropics. The Quarterly Journal of Economics 127 (4):
1707-1754.
[8] Delacote, P., A. Dragicevic,, and S. Garcia. 2014. Principal-agent relationship in re-
source management, multiple principals and spatial dynamics (No. 2014-10). Laboratoire
d’Economie Forestiere, AgroParisTech-INRA.
[9] Fehr, E. and K. M. Schmidt. 2004. Fairness and Incentives in a Multi-task Principal-Agent
Model. Scandinavian Journal of Economics 106 (3): 453-474.
[10] Food and Agriculture Organization of the United Nations (FAO). 2016. Global Forest
Resources Assessment 2015: How Are the World’s Forests Changing? Second edition.
FAO, Rome.
[11] Foster, A.D., and M.R. Rosenzweig. 2003. Economic growth and the rise of forests. The
Quarterly Journal of Economics 118 (2): 601-637.
[12] Gray, J. A., 2002. Forest Concession Policies and Revenue Systems: Country Experience
and Policy Changes for Sustainable Tropical Forestry (Vol. 522). World Bank Publica-
tions.
[13] Groves, T., Y. Hong, J. McMillan, and B. Naughton. 1995. China’s evolving managerial
labor market. Journal of Political Economy, 103 (4): 873-892.
25
[14] Holmstrom, B. and P. Milgrom. 1991. Multitask principal-agent analyses: incentive con-
tracts, asset ownership, and job design. Journal of Law, Economics, and Organization 7:
24-52.
[15] Holmstrom, B. and P. Milgrom. 1994. The firm as an incentive system. The American
Economic Review 84 (4): 972-991.
[16] Huete, A., K. Didan, T. Miura, E. P. Rodriguez, X. Gao, L. G. Ferreira. 2002. Overview
of the radiometric and biophysical performance of the MODIS vegetation indices. Remote
Sensing of Environment 83 (2002): 195-213.
[17] Hyde, W. F., G. S. Amacher, and W. Magrath. 1996. Deforestation and forest land use:
theory, evidence, and policy implications. The World Bank Research Observer 11 (2):
223-248.
[18] Jin, Hehui, Yingyi Qian, and Barry Weingast. 2005. Regional decentralization and fiscal
incentives: federalism, Chinese style. Journal of Public Economics 89 (9-10): 1719-1742.
[19] Johnson, S. 2002. Documenting the undocumented. Tropical Forest Update 12 (1): 6-14.
[20] Li X., and C.G.W. Chan. 2016. Who pollutes? Ownership type and environmental per-
formance of Chinese firms. Journal of Contemporary China 25 (98): 248-263.
[21] Li, H. and L.-A. Zhou. 2005. Political turnover and economic performance: the incentive
role of personnel control in China. Journal of Public Economics 89 (9-10): 1743-1762.
[22] Lin, Justin Yifu, and Zhiqiang Liu. 2000. Fiscal decentralization and economic growth
in China. Economic Development and Cultural Change 49 (1): 1-21.
[23] Liu, J. and J. Diamond. 2008. Revolutionizing China’s environmental protection. Science
319 (5859): 37-8.
[24] Lo, C.W.-H., G.E. Fryxell, and W.W.-H. Wong. 2006. Effective regulations with lit-
tle effect? The antecedents of the perceptions of environmental officials on enforcement
effectiveness in China. Environmental Management 38: 388-410.
[25] Maskin, Eric, Yingyi Qian, and Chenggang Xu. 2000. Incentives, information, and orga-
nizational form. The Review of Economic Studies 67 (2): 359-378.
[26] Ministry of Environmental Protection (MEP), P.R. China. 2003. Measures for the Ad-
ministration of the Charging Rates for Pollutant Discharge Fees [in Chinese]. (Ministry
of Environmental Protection, P.R. China, 2003-07-01). http://zfs.mep.gov.cn/
[27] Montinola, Gabriella, Yingyi Qian, and Barry R. Weingast. 1995. Federalism, Chinese
style: the political basis for economic success in China. World Politics 48 (01): 50-81.
[28] Murphy, K., and J. Zimmerman. 1993. Financial performance surrounding CEO
turnover. Journal of Accounting and Economics 16: 273-315.
[29] NASA LP DAAC. 2015. Vegetation Indices 16-Day L3 Global 250m: MOD13Q1. NASA
EOSDIS Land Processes DAAC, USGS Earth Resources Observation and Science (EROS)
Center, Sioux Falls, South Dakota https://lpdaac.usgs.gov.
26
[30] National Bureau of Statistics of China (NBSC). 1999, 2011, 2015. China Statistical Year-
book, 2015. China Statistics Press, Beijing, China (2015)
[31] Nolan, P. and X. Wang. 1999. Beyond privatization: institutional innovation and growth
in China’s large state-owned enterprises. World Development 27 (1): 169-200.
[32] OECD. 2009. Ensuring Environmental Compliance: Trends and Good Practices. Organ-
isation for Economic Co-operation and Development. ISBN: 9789264059580. Released:
15/06/2009.
[33] Palmer, C. 2001. The extent and causes of illegal logging: an analysis of a major cause
of tropical deforestation in Indonesia. CSERGE working paper, Economics Department,
University College London.
[34] Poore, E. 1993. The sustainable management of tropical forests: the issues. In: Rietber-
gen, E. (Ed.), The Earthscan Reader in Tropical Forestry. Earthscan Publications Ltd,
London.
[35] Qian, Y. and B. R. Weingast. 1997. Federalism as a commitment to preserving market
incentives. The Journal of Economic Perspectives 11 (4): 83-92.
[36] Qian, Y. and C. Xu. 1993. Why China’s economic reforms differ: the M-form hierarchy
and entry/expansion of the non-state sector. Economics of Transition 1 (2): 135-170.
[37] Rozelle, S.,H. Albers, L. Guo, V. Benziger. 1998. Forest resources under economic reform
in China. China Information 13: 106-125.
[38] Rozelle, S. and J. Huang. 2000. China: From Afforestation to Poverty Alleviation and
Natural Forest Management, World Bank Publications.
[39] Sellers, P.J. 1985. Canopy reflectance, photosynthesis, and transpiration. International
Journal of Remote Sensing 6 (8): 1335-1372.
[40] Sellers, P. J., S. O. Los, C. J. Tucker, C. O. Justice, G. J. Collatz, and D. A. Randall.
1994. A global 1 degree by 1 degree NDVI data set for climate studies. Part 2: the adjust-
ment of the NDVI and generation of global fields of terrestrial biophysical parameters.
International Journal of Remote Sensing 15: 3519-3546.
[41] Thomas, J. and T. Worrall. 1990. Income fluctuation and asymmetric information: an
example of a repeated principal-agent problem. Journal of Economic Theory 51 (2): 367-
390.
[42] Tirole, J. 1986. Hierarchies and bureaucracies: on the role of collusion in organizations.
Journal of Law, Economics, and Organization 2 (2): 181-214.
[43] Tucker, C.J. 1980. Remote sensing of leaf water content in the near infrared. Remote
Sensing of Environment 10: 23-32.
[44] Tucker, C.J., J. Pinzon, M. Brown, and GIMMS/GSFC/NASA. 2010. ISLSCP II GIMMS
Monthly NDVI, 1981-2002. In Hall, Forrest G., G. Collatz, B. Meeson, S. Los, E. Brown de
27
Colstoun, and D. Landis (eds.). ISLSCP Initiative II Collection. Data set. Available on-
line [http://daac.ornl.gov/] from Oak Ridge National Laboratory Distributed Active
Archive Center, Oak Ridge, Tennessee, U.S.A.
[45] Ward D. and I. Filatotchev. 2010. Principal-principal-agency relationships and the role
of external governance. Managerial and Decision Economics 31 (4): 249-261.
[46] Weisbach, M. 1988. Outside directors and CEO turnover. Journal of Financial Eco-
nomics 20 (1-2): 431-460.
[47] World Bank. 1993. China: Budgetary Policy and Intergovernmental Fiscal Relations.
Report No. 11094-CHA. Washington, D.C.
[48] Xu, Chenggang. 2011. The fundamental institutions of China’s reforms and development.
Journal of Economic Literature 49 (4): 1076-1151.
[49] Xu, Chenggang, and Juzhong Zhuang. 1998. Why China grew: the role of decentraliza-
tion. In Emerging from Communism: Lessons from Russia, China, and Eastern Europe,
ed. Peter Boone, Stanislaw Gomulka, and Richard Layard, 183-212. Cambridge and Lon-
don: MIT Press.
[50] Xu, Jintao. 2013. State Forest Reform in Northeast China: Issues and Options. PROFOR
Working Paper. World Bank. Washington DC: Program on Forests (PROFOR).
[51] Xu, Jintao, Ran Tao, and Gregory S. Amacher. 2004. An empirical analysis of China’s
state-owned forests. Forest Policy and Economics 6 (3) (2004): 379-390.
[52] Zhang, K. Z., J. Wang, and X. Y. Cui. 2011. Fiscal decentralization and environmental
pollution: from the perspective of carbon emission [in Chinese]. China Industrial Eco-
nomics 10: 008.
28
Figure 1: The study area: maps of 24 SFEs
29
Figure 2: Forest change in Northeast China, 1981-2009
(a) NDVI 1981 (b) NDVI 1986
(c) NDVI 1991 (d) NDVI 1996
(e) NDVI 2001 (f) EVI 2001
(g) EVI 2005 (h) EVI 2009
Data sources: Tucker et al., 2010; NASA LP DAAC, 2015
30
Table1:Definitionsofvariables
Variable:Definitionandunitofmeasurement
Measureofdeforestation:
AnnualrateofchangeofvegetationindexAnnualrateofchangeofNDVI/EVIvaluesforanSFE-managedforestlandarea
DeforestationBinary,=1ifnegativeannualrateofchangeofvegetationindex,=0otherwise
DifficultyinevaluatingtheSFA’starget:
TotalforestlandareaTotalforestlandareamanagedbyanSFE(ha)
TotalforestvolumeTotalforestlivingstockvolumemanagedbyanSFE(1000m3)
AreashareofnaturalforestNaturalforestareaasashareofthetotallandareainanSFE
VolumeshareofnaturalforestNaturalforestvolumeasashareofthetotalforestlivingstockvolumeinanSFE
AreashareofplantationforestPlantationforestareaasashareofthetotallandareainanSFE
VolumeshareofplantationforestPlantationforestvolumeasashareofthetotalforestlivingstockvolumeinanSFE
ProfitstransferredtolocalgovernmentAmountofanSFE’sprofitstransferredtothejurisdictionalgovernment(10000CNY)
ProportionofprofitstransferredPercentageofanSFE’sprofitstransferredtothejurisdictionalgovernment
Controlvariables:
AnnualgrowthrateofnaturalforestvolumeAverageannualgrowthrateoftimbervolumeinnaturalforestsofanSFE
ShareofyearlyafforestationareaTotalafforestedandreforestedareaattimetasashareoftotalforestlandareainanSFE
LogofyearlyafforestationareaNaturallogofyearlyafforestationarea(ha)
PopulationPopulationsizeofanSFE
PopulationdensityPersonsperhectareinanSFE
Ratioofemployeestothesupportedlaid-offandretiredRatioofthenumberofemployeestothenumberofsupportedlaid-offandretiredpersons
CountypopulationTotalpopulationsizeinthecountywhereanSFEislocated
CountyGDPpercapitaAnnualGDPinthecountywhereanSFEislocated(CNY)
CountyurbanpopulationshareUrbanpopulationasashareoftotalpopulationinthecountywhereanSFEislocated
CountyagriculturalpopulationdensityAgriculturalpopulationdensityinthecountywhereanSFEislocated(personsperha)
Countynon-agriculturalGDPshareNon-agriculturalGDPasashareoftotalGDPinthecountywhereanSFEislocated
TimberpriceTimberprice(1000CNYperm3)
Managercharacteristics:
AgeCurrentageoftheincumbentmanagerofanSFE(years)
EducationEducationoftheSFEmanager(1ifcollege,0iflower)
ManagerialexperienceNumberofyearsbeinganSFEmanager
PromotionattheendofcurrentmanagerialcontractManagerbeingpromotedtogovernmentposition(1=yes,0=no)
PromotionofpreviousmanagerWhetherthepreviousmanagerwaspromotedtolocalgovernment(1=yes,0=no)
Notes:1.SFA:StateForestryAdministrationofP.R.China.
2.SFEs:stateforestenterprises.
3.Laid-offreferstothelaid-offworkersinanSFEwhoreceiveamonthlyallowanceforsubsistencepaidbytheSFE.
31
Table2:SummarystatisticsofSFEsandcounties(means)
FullByyear
sample1980198519901995200020042008
Measureofdeforestation:
NDVI(EVIfrom2003)0.662n/a0.7250.7530.7730.7390.5120.516
NDVIinthenextyear(EVIfrom2003)0.6700.7250.7320.7650.7640.7610.5200.500
Annualrateofchange,vegetationindex0.01830.09120.0483-0.04050.0251-0.0448-0.05090.149
Deforestation0.47400.1110.3080.6670.2110.9470.8090
DifficultyinevaluatingtheSFA’starget:
Totalforestlandarea256,200248,853252,037242,041274,746235,894259,685272,471
Totalforeststock20,02622,66920,42818,42220,20317,36819,17323,365
Areashareofnaturalforest0.6710.6650.6420.6320.6600.6710.6820.725
Volumeshareofnaturalforest0.8550.9000.8970.8910.8790.8300.8020.832
Areashareofplantationforest0.1340.06840.07380.1030.1410.1750.1750.141
Volumeshareofplantationforest0.09370.01910.04570.05520.06390.1290.1500.126
Profitstransferredtolocalgovernment2,190516.7925.61,6863,6112,3422,9651,796
Proportionofprofitstransferred0.2880.3010.2950.3340.4910.2280.2910.0807
Controlvariables:
Annualgrowthrateofnaturalforestvolume0.02760.02310.02200.02670.02850.03300.03160.0236
Shareofyearlyafforestationarea0.03000.05150.04540.04520.04400.02490.01300.00593
Yearlyafforestationarea5,46611,6587,7238,0757,8124,1612,1301,359
Population36,78135,94137,56938,01240,10536,58033,47736,163
Populationdensity0.1940.1730.2030.1970.1920.1940.1810.212
Ratioofemployeestothelaid-offandretired0.6870.1010.2080.3240.3930.7751.0571.411
Countypopulation345,142233,954244,995442,822522,122397,232271,976235,229
CountyGDPpercapita7,8841,0741,7581,9454,5818,33915,66314,591
Countyurbanpopulationshare0.6540.6580.6770.6050.6450.6710.6920.624
Countyagriculturalpopulationdensity0.6870.5010.4401.1840.9130.8800.4020.436
Countynon-agriculturalGDPshare0.3590.4340.3930.4820.3860.2920.3020.301
Timberprice0.4540.01770.03690.08450.1900.4690.8421.092
Managercharacteristics:
Age48.0349.2245.5447.0750.0546.7447.1050.33
Education0.8250.2220.6920.8000.684111
Managerialexperience3.5091.2221.7694.8003.6323.0532.9525.833
Promotionattheendofcurrentmanagerialcontract0.4120.3330.6150.6670.4740.5790.2860.111
Promotionofpreviousmanager0.5710n/a0.2310.6670.8420.5790.6670.333
Numberofobservations:1149131519192118
Notes:1.SFA:StateForestryAdministrationofP.R.China.
2.SFE:stateforestenterprise.
3.Laid-offreferstothelaid-offworkersinanSFEwhoreceiveamonthlyallowanceforsubsistencepaidbytheSFE.
Sources:Author’scomputationfromSFEsurveydata,andsatelliteimagedata(Carrolletal.,2003;Tuckeretal.,2010;NASALPDAAC,2015).
32
Table 3: SFE fixed-effects (FE) estimates: information asymmetry and deforestation
1 2 3
Dep. Var.: Rate of vegetation change, t+ 1 Deforestation
Proportion of profits transferred (γt) -0.023 0.963**
(0.053) (0.386)
Log of total forest land area -0.292* -0.301* 1.881*
(0.144) (0.148) (1.070)
Log of forest stock volume -0.106** -0.104** 0.102
(0.048) (0.049) (0.358)
Area share of natural forest -0.577*** -0.551*** 1.696
(0.156) (0.169) (1.223)
Volume share of natural forest 0.139 0.144 -0.466
(0.155) (0.158) (1.145)
Log of vegetation index value at t− 1 YES YES YES
SFE characteristics YES YES YES
County characteristics YES YES YES
Manager characteristics YES YES YES
Promotion of previous manager YES YES YES
Promotion of previous manager × He deforested YES YES YES
Upper bureau fixed and year effects YES YES YES
Year trends YES YES YES
Observations 105 105 105
R-squared 0.953 0.953 0.916
Notes: The dependent variable in Models 1 and 2 is the rate of vegetation index change at t+1. The dependent
variable in Model 3 is binary, with value 1 meaning deforestation given a negative rate of vegetation index
change at t + 1. SFE characteristics include annual growth rate of natural forest volume, log area of newly
afforested forest and as a percentage of total forest area, population density at SFE, and ratio of employees to
the supported laid-off and retired people. County characteristics include population, GDP per capita, urban
population share, agricultural population density, non-agricultural GDP share, and timber price. Manager
characteristics include age, education level, and managerial experience. Standard errors are in parentheses.
Significance is denoted as: * significant at 10%; ** significant at 5%; *** significant at 1%.
We lose 9 observations (from 114 in total) because we included NDVI values that are lagged by one year
and these are not available prior to 1980.
33
Table 4: Probit estimates: determinants of an SFE manager’s political promotion
Dep. Var.: Promotion (1=yes, 0=no) 1 2 3 4
Proportion of profits transferred (γt) 2.388** 2.861** 2.904** 0.941**
(1.084) (1.245) (1.324) (0.423)
Log amount of transferred profits 0.146
(0.409)
Log amount of yearly average transfer -0.430
(0.358)
Log amount of aggregated tranfer -0.505*
(0.297)
Log of yearly afforestation area 0.232 0.240 0.245
(0.232) (0.225) (0.224)
Share of yearly afforestation area over the total -10.724* -15.930** -17.143** -4.228*
(5.920) (7.089) (7.167) (2.366)
Log of total forest land area 8.337* 10.134** 10.378** 3.287*
(4.652) (4.487) (4.713) (1.809)
Share of natural forest volume over the total 15.431** 11.148* 10.286 6.084**
(7.195) (6.529) (6.726) (2.746)
Log of population 0.276 0.408 0.386
(0.432) (0.434) (0.416)
Population share of the county total 0.098 -0.291 -0.366
(1.790) (1.958) (1.955)
Manager education -0.312 -0.400 -0.434
(0.924) (0.909) (1.046)
Managerial experience 0.032 -0.039 0.096
(0.112) (0.100) (0.166)
Number of years to retirement (i.e., 65 for males in China) 0.182*** 0.203** 0.203** 0.072***
(0.069) (0.081) (0.087) (0.027)
Promotion of previous manager (1 if yes, 0 otherwise) -2.174*** -2.072*** -2.124*** -0.857***
(0.629) (0.757) (0.799) (0.243)
Deforestation by previous manager (1 if yes, 0 otherwise) -1.942** -1.555 -1.526 -0.766**
(0.819) (0.962) (0.983) (0.321)
Promotion of previous manager × He deforested 1.736* 1.538 1.556 0.684*
(0.993) (1.072) (1.148) (0.388)
Observations 105 105 105
SFE Fixed-effects YES YES YES
Pseudo R2 0.486 0.500 0.508
Notes: The dependent variable is whether the incumbent manager is promoted to a government position at the end
of the current managerial contract.
Robust standard errors in parentheses are clustered at the SFE level. The significance levels of 1%, 5% and 10% are
noted by ***, ** and *.
Column 4 reports the marginal effects computed at sample means for Column 1.
34

Previous doctoral theses in the Department of Economics, Gothenburg 
 
Avhandlingar publicerade innan serien Ekonomiska Studier startades  
(Theses published before the series Ekonomiska Studier was started): 
 
Östman, Hugo (1911), Norrlands ekonomiska utveckling 
Moritz, Marcus (1911), Den svenska tobaksindustrien 
Sundbom, I. (1933), Prisbildning och ändamålsenlighet 
Gerhard, I. (1948), Problem rörande Sveriges utrikeshandel 1936/38 
Hegeland, Hugo (1951), The Quantity Theory of Money 
Mattsson, Bengt (1970), Cost-Benefit analys 
Rosengren, Björn (1975), Valutareglering och nationell ekonomisk politik 
Hjalmarsson, Lennart (1975), Studies in a Dynamic Theory of Production and its 
Applications 
Örtendahl, Per-Anders (1975), Substitutionsaspekter på produktionsprocessen vid 
massaframställning 
Anderson, Arne M. (1976), Produktion, kapacitet och kostnader vid ett helautomatiskt 
emballageglasbruk 
Ohlsson, Olle (1976), Substitution och odelbarheter i produktionsprocessen vid 
massaframställning 
Gunnarsson, Jan (1976), Produktionssystem och tätortshierarki – om sambandet mellan 
rumslig och ekonomisk struktur 
Köstner, Evert (1976), Optimal allokering av tid mellan utbildning och arbete 
Wigren, Rune (1976), Analys av regionala effektivitetsskillnader inom industribranscher 
Wästlund, Jan (1976), Skattning och analys av regionala effektivitetsskillnader inom 
industribranscher 
Flöjstad, Gunnar (1976), Studies in Distortions, Trade and Allocation Problems 
Sandelin, Bo (1977), Prisutveckling och kapitalvinster på bostadsfastigheter 
Dahlberg, Lars (1977), Empirical Studies in Public Planning 
Lönnroth, Johan (1977), Marxism som matematisk ekonomi 
Johansson, Börje (1978), Contributions to Sequential Analysis of Oligopolistic 
Competition 
 
Ekonomiska Studier, utgivna av Nationalekonomiska institutionen vid Göteborgs  
Universitet. Nr 1 och 4 var inte doktorsavhandlingar. (The contributions to the department 
series ’Ekonomiska Studier’ where no. 1 and 4 were no doctoral theses): 
 
2. Ambjörn, Erik (1959), Svenskt importberoende 1926-1956: en ekonomisk-
statistisk kartläggning med kommentarer 
3. Landgren, K-G. (1960), Den ”Nya ekonomien” i Sverige: J.M. Keynes, E. 
Wigfors och utecklingen 1927-39 
5. Bigsten, Arne (1979), Regional Inequality and Development: A Case Study of 
Kenya 
6. Andersson, Lars (1979), Statens styrning av de kommunala budgetarnas struktur 
(Central Government Influence on the Structure of the Municipal Budget) 
7. Gustafsson, Björn (1979), Inkomst- och uppväxtförhållanden (Income and 
Family Background) 
8. Granholm, Arne (1981), Interregional Planning Models for the Allocation of 
Private and Public Investments 
9. Lundborg, Per (1982), Trade Policy and Development: Income Distributional 
Effects in the Less Developed Countries of the US and EEC Policies for 
Agricultural Commodities 
10. Juås, Birgitta (1982), Värdering av risken för personskador. En jämförande studie 
av implicita och explicita värden. (Valuation of Personal Injuries. A comparison of 
Explicit and Implicit Values) 
11. Bergendahl, Per-Anders (1982), Energi och ekonomi - tillämpningar av input-
output analys (Energy and the Economy - Applications of Input-Output Analysis) 
12. Blomström, Magnus (1983), Foreign Investment, Technical Efficiency and 
Structural Change - Evidence from the Mexican Manufacturing Industry 
13. Larsson, Lars-Göran (1983), Comparative Statics on the Basis of Optimization 
Methods 
14. Persson, Håkan (1983), Theory and Applications of Multisectoral Growth 
Models 
15. Sterner, Thomas (1986), Energy Use in Mexican Industry. 
16. Flood, Lennart (1986), On the Application of Time Use and Expenditure 
Allocation Models. 
17. Schuller, Bernd-Joachim (1986), Ekonomi och kriminalitet - en empirisk 
undersökning av brottsligheten i Sverige (Economics of crime - an empirical 
analysis of crime in Sweden) 
18. Walfridson, Bo (1987), Dynamic Models of Factor Demand. An Application to 
Swedish Industry.  
19. Stålhammar, Nils-Olov (1987), Strukturomvandling, företagsbeteende och 
förväntningsbildning inom den svenska tillverkningsindustrin (Structural Change, 
Firm Behaviour and Expectation Formation in Swedish Manufactury) 
20. Anxo, Dominique (1988), Sysselsättningseffekter av en allmän arbetstidsför-
kortning (Employment effects of a general shortage of the working time) 
21. Mbelle, Ammon (1988), Foreign Exchange and Industrial Development: A Study 
of Tanzania. 
22. Ongaro, Wilfred (1988), Adoption of New Farming Technology: A Case Study 
of Maize Production in Western Kenya. 
23. Zejan, Mario (1988), Studies in the Behavior of Swedish Multinationals. 
24. Görling, Anders (1988), Ekonomisk tillväxt och miljö. Förorenings-struktur och 
ekonomiska effekter av olika miljövårdsprogram. (Economic Growth and 
Environment. Pollution Structure and Economic Effects of Some Environmental 
Programs). 
25. Aguilar, Renato (1988), Efficiency in Production: Theory and an Application on 
Kenyan Smallholders. 
26. Kayizzi-Mugerwa, Steve (1988), External Shocks and Adjustment in Zambia. 
27. Bornmalm-Jardelöw, Gunilla (1988), Högre utbildning och arbetsmarknad 
(Higher Education and the Labour Market) 
28. Tansini, Ruben (1989), Technology Transfer: Dairy Industries in Sweden and 
Uruguay. 
29. Andersson, Irene (1989), Familjebeskattning, konsumtion och arbetsutbud - En 
ekonometrisk analys av löne- och inkomstelasticiteter samt policysimuleringar för 
svenska hushåll (Family Taxation, Consumption and Labour Supply - An 
Econometric Analysis of Wage and Income Elasticities and Policy Simulations for 
Swedish Households) 
30. Henrekson, Magnus (1990), An Economic Analysis of Swedish Government 
Expenditure 
31. Sjöö, Boo (1990), Monetary Policy in a Continuous Time Dynamic Model for 
Sweden 
32. Rosén, Åsa (1991), Contributions to the Theory of Labour Contracts. 
33. Loureiro, Joao M. de Matos (1992), Foreign Exchange Intervention, Sterilization 
and Credibility in the EMS: An Empirical Study 
34. Irandoust, Manuchehr (1993), Essays on the Behavior and Performance of 
 the Car Industry 
35. Tasiran, Ali Cevat (1993), Wage and Income Effects on the Timing and  
 Spacing of Births in Sweden and the United States  
36. Milopoulos, Christos (1993), Investment Behaviour under Uncertainty: An 
Econometric Analysis of Swedish Panel Data 
37. Andersson, Per-Åke (1993), Labour Market Structure in a Controlled Economy: 
The Case of Zambia 
38. Storrie, Donald W. (1993), The Anatomy of a Large Swedish Plant Closure 
39. Semboja, Haji Hatibu Haji (1993), Energy and Development in Kenya 
40. Makonnen, Negatu (1993), Labor Supply and the Distribution of Economic 
 Well-Being: A Case Study of Lesotho 
41. Julin, Eva (1993), Structural Change in Rural Kenya 
42. Durevall, Dick (1993), Essays on Chronic Inflation: The Brazilian Experience 
43. Veiderpass, Ann (1993), Swedish Retail Electricity Distribution: A Non-
Parametric Approach to Efficiency and Productivity Change 
44. Odeck, James (1993), Measuring Productivity Growth and Efficiency with 
 Data Envelopment Analysis: An Application on the Norwegian Road Sector 
45. Mwenda, Abraham (1993), Credit Rationing and Investment Behaviour under 
 Market Imperfections: Evidence from Commercial Agriculture in Zambia 
46. Mlambo, Kupukile (1993), Total Factor Productivity Growth: An Empirical 
Analysis of Zimbabwe's Manufacturing Sector Based on Factor Demand  
 Modelling 
47. Ndung'u, Njuguna (1993), Dynamics of the Inflationary Process in Kenya 
48. Modén, Karl-Markus (1993), Tax Incentives of Corporate Mergers and          
Foreign Direct Investments 
49. Franzén, Mikael (1994), Gasoline Demand - A Comparison of Models 
50. Heshmati, Almas (1994), Estimating Technical Efficiency, Productivity Growth 
And Selectivity Bias Using Rotating Panel Data: An Application to Swedish 
Agriculture 
51. Salas, Osvaldo (1994), Efficiency and Productivity Change: A Micro Data Case 
Study of the Colombian Cement Industry 
52. Bjurek, Hans (1994), Essays on Efficiency and Productivity Change with 
Applications to Public Service Production 
53. Cabezas Vega, Luis (1994), Factor Substitution, Capacity Utilization and Total 
Factor Productivity Growth in the Peruvian Manufacturing Industry  
54. Katz, Katarina (1994), Gender Differentiation and Discrimination. A Study of 
Soviet Wages 
55. Asal, Maher (1995), Real Exchange Rate Determination and the Adjustment 
 Process: An Empirical Study in the Cases of Sweden and Egypt 
56. Kjulin, Urban (1995), Economic Perspectives on Child Care 
57. Andersson, Göran (1995), Volatility Forecasting and Efficiency of the Swedish 
Call Options Market 
58. Forteza, Alvaro (1996), Credibility, Inflation and Incentive Distortions in the 
Welfare State 
59. Locking, Håkan (1996), Essays on Swedish Wage Formation 
60. Välilä, Timo (1996), Essays on the Credibility of Central Bank Independence 
61. Yilma, Mulugeta (1996), Measuring Smallholder Efficiency: Ugandan Coffee 
and Food-Crop Production 
62. Mabugu, Ramos E. (1996), Tax Policy Analysis in Zimbabwe Applying General 
Equilibrium Models 
63. Johansson, Olof (1996), Welfare, Externalities, and Taxation; Theory and Some 
Road Transport Applications. 
64. Chitiga, Margaret (1996), Computable General Equilibrium Analysis of Income 
Distribution Policies in Zimbabwe 
65. Leander, Per (1996), Foreign Exchange Market Behavior Expectations and 
Chaos 
66. Hansen, Jörgen (1997), Essays on Earnings and Labor Supply 
67. Cotfas, Mihai (1997), Essays on Productivity and Efficiency in the Romanian 
Cement Industry 
68. Horgby, Per-Johan (1997), Essays on Sharing, Management and Evaluation of 
Health Risks 
69. Nafar, Nosratollah (1997), Efficiency and Productivity in Iranian Manufacturing 
Industries 
70. Zheng, Jinghai (1997), Essays on Industrial Structure, Technical Change, 
Employment Adjustment, and Technical Efficiency 
71. Isaksson, Anders (1997), Essays on Financial Liberalisation in Developing 
Countries: Capital mobility, price stability, and savings 
72. Gerdin, Anders (1997), On Productivity and Growth in Kenya, 1964-94 
73. Sharifi, Alimorad (1998), The Electricity Supply Industry in Iran: Organization, 
performance and future development 
74. Zamanian, Max (1997), Methods for Mutual Fund Portfolio Evaluation: An 
application to the Swedish market 
75. Manda, Damiano Kulundu (1997), Labour Supply, Returns to Education, and 
the Effect of Firm Size on Wages: The case of Kenya 
76. Holmén, Martin (1998), Essays on Corporate Acquisitions and Stock Market 
Introductions 
77. Pan, Kelvin (1998), Essays on Enforcement in Money and Banking 
78. Rogat, Jorge (1998), The Value of Improved Air Quality in Santiago de Chile 
79. Peterson, Stefan (1998), Essays on Large Shareholders and Corporate Control 
80. Belhaj, Mohammed (1998), Energy, Transportation and Urban Environment in 
Africa: The Case of Rabat-Salé, Morocco 
81. Mekonnen, Alemu (1998), Rural Energy and Afforestation: Case Studies from 
Ethiopia 
82. Johansson, Anders (1998), Empirical Essays on Financial and Real Investment 
Behavior 
83. Köhlin, Gunnar (1998), The Value of Social Forestry in Orissa, India 
84. Levin, Jörgen (1998), Structural Adjustment and Poverty: The Case of Kenya 
85. Ncube, Mkhululi (1998), Analysis of Employment Behaviour in Zimbabwe 
86. Mwansa, Ladslous (1998), Determinants of Inflation in Zambia 
87. Agnarsson, Sveinn (1998), Of Men and Machines: Essays in Applied Labour and 
Production Economics 
88. Kadenge, Phineas (1998), Essays on Macroeconomic Adjustment in Zimbabwe: 
Inflation, Money Demand, and the Real Exchange Rate 
89. Nyman, Håkan (1998), An Economic Analysis of Lone Motherhood in Sweden 
90. Carlsson, Fredrik (1999), Essays on Externalities and Transport 
91. Johansson, Mats (1999), Empirical Studies of Income Distribution 
92. Alemu, Tekie (1999), Land Tenure and Soil Conservation: Evidence from 
Ethiopia 
93. Lundvall, Karl (1999), Essays on Manufacturing Production in a Developing 
Economy: Kenya 1992-94 
94. Zhang, Jianhua (1999), Essays on Emerging Market Finance 
95. Mlima, Aziz Ponary (1999), Four Essays on Efficiency and Productivity in 
Swedish Banking 
96. Davidsen, Björn-Ivar (2000), Bidrag til den økonomisk-metodologiske 
tenkningen (Contributions to the Economic Methodological Thinking) 
97. Ericson, Peter (2000), Essays on Labor Supply 
98. Söderbom, Måns (2000), Investment in African Manufacturing: A 
Microeconomic Analysis 
99. Höglund, Lena (2000), Essays on Environmental Regulation with Applications 
 to Sweden 
100. Olsson, Ola (2000), Perspectives on Knowledge and Growth 
101. Meuller, Lars (2000), Essays on Money and Credit 
102. Österberg, Torun (2000), Economic Perspectives on Immigrants and 
Intergenerational Transmissions 
103.   Kalinda Mkenda, Beatrice (2001), Essays on Purchasing Power Parity, 
RealExchange Rate, and Optimum Currency Areas 
104. Nerhagen, Lena (2001), Travel Demand and Value of Time - Towards an 
Understanding of Individuals Choice Behavior 
105. Mkenda, Adolf (2001), Fishery Resources and Welfare in Rural  
               Zanzibar 
106. Eggert, Håkan (2001), Essays on Fisheries Economics 
107. Andrén, Daniela (2001), Work, Sickness, Earnings, and Early Exits from the 
Labor Market. An Empirical Analysis Using Swedish Longitudinal Data 
108. Nivorozhkin, Eugene (2001), Essays on Capital Structure 
109. Hammar, Henrik (2001), Essays on Policy Instruments: Applications to Smoking 
and the Environment 
110. Nannyonjo, Justine (2002), Financial Sector Reforms in Uganda (1990-2000): 
Interest Rate Spreads, Market Structure, Bank Performance and Monetary Policy 
111. Wu, Hong (2002), Essays on Insurance Economics 
112. Linde-Rahr, Martin (2002), Household Economics of Agriculture and Forestry 
in Rural Vienam 
113. Maneschiöld, Per-Ola (2002), Essays on Exchange Rates and Central Bank 
Credibility 
114. Andrén, Thomas (2002), Essays on Training, Welfare and Labor Supply 
115. Granér, Mats (2002), Essays on Trade and Productivity: Case Studies of  
 Manufacturing in Chile and Kenya 
116. Jaldell, Henrik (2002), Essays on the Performance of Fire and Rescue Services 
117. Alpizar, Francisco, R. (2002), Essays on Environmental Policy-Making in 
Developing Countries: Applications to Costa Rica 
118. Wahlberg, Roger (2002), Essays on Discrimination, Welfare and Labor Supply 
119. Piculescu, Violeta (2002), Studies on the Post-Communist Transition 
120. Pylkkänen, Elina (2003), Studies on Household Labor Supply and Home 
Production 
121. Löfgren, Åsa (2003), Environmental Taxation – Empirical and Theoretical 
Applications 
122. Ivaschenko, Oleksiy (2003), Essays on Poverty, Income Inequality and Health in 
Transition Economies 
123. Lundström, Susanna (2003), On Institutions, Economic Growth and the 
Environment 
124. Wambugu, Anthony (2003), Essays on Earnings and Human Capital in Kenya 
125. Adler, Johan (2003), Aspects of Macroeconomic Saving 
126. Erlandsson, Mattias (2003), On Monetary Integration and Macroeconomic 
Policy 
127. Brink, Anna (2003), On the Political Economy of Municipality Break-Ups 
128. Ljungwall, Christer (2003), Essays on China’s Economic Performance During 
the Reform Period 
129. Chifamba, Ronald (2003), Analysis of Mining Investments in Zimbabwe 
130. Muchapondwa, Edwin (2003), The Economics of Community-Based Wildlife 
Conservation in Zimbabwe 
131. Hammes, Klaus (2003), Essays on Capital Structure and Trade Financing 
132. Abou-Ali, Hala (2003), Water and Health in Egypt: An Empirical Analysis 
133. Simatele, Munacinga (2004), Financial Sector Reforms and Monetary Policy in 
Zambia 
134. Tezic, Kerem (2004), Essays on Immigrants’ Economic Integration 
135. INSTÄLLD 
136. Gjirja, Matilda (2004), Efficiency and Productivity in Swedish Banking 
137. Andersson, Jessica (2004), Welfare Environment and Tourism in Developing 
Countries 
138. Chen, Yinghong (2004), Essays on Voting Power, Corporate Governance and 
Capital Structure 
139. Yesuf, Mahmud (2004), Risk, Time and Land Management under Market 
Imperfections: Applications to Ethiopia 
140. Kateregga, Eseza (2005), Essays on the Infestation of Lake Victoria by the Water 
Hyacinth 
141. Edvardsen, Dag Fjeld (2004), Four Essays on the Measurement of Productive 
Efficiency 
142. Lidén, Erik (2005), Essays on Information and Conflicts of Interest in Stock 
Recommendations 
143. Dieden, Sten (2005), Income Generation in the African and Coloured Population 
– Three Essays on the Origins of Household Incomes in South Africa 
144. Eliasson, Marcus (2005), Individual and Family Consequences of Involuntary 
Job Loss 
145. Mahmud, Minhaj (2005), Measuring Trust and the Value of Statistical Lives: 
Evidence from Bangladesh 
146. Lokina, Razack Bakari (2005), Efficiency, Risk and Regulation Compliance: 
Applications to Lake Victoria Fisheries in Tanzania 
147. Jussila Hammes, Johanna (2005), Essays on the Political Economy of Land Use 
Change 
148. Nyangena, Wilfred (2006), Essays on Soil Conservation, Social Capital and 
Technology Adoption 
149. Nivorozhkin, Anton (2006), Essays on Unemployment Duration and Programme 
Evaluation 
150. Sandén, Klas (2006), Essays on the Skill Premium 
151. Deng, Daniel (2006), Three Essays on Electricity Spot and Financial Derivative 
Prices at the Nordic Power Exchange 
152. Gebreeyesus, Mulu (2006), Essays on Firm Turnover, Growth, and Investment 
Behavior in Ethiopian Manufacturing 
153. Islam, Nizamul Md. (2006), Essays on Labor Supply and Poverty: A 
Microeconometric Application 
154. Kjaer, Mats (2006), Pricing of Some Path-Dependent Options on Equities and 
Commodities 
155. Shimeles, Abebe (2006), Essays on Poverty, Risk and Consumption Dynamics in 
Ethiopia 
156. Larsson, Jan (2006), Four Essays on Technology, Productivity and Environment 
157. Congdon Fors, Heather (2006), Essays in Institutional and Development 
Economics 
158. Akpalu, Wisdom (2006), Essays on Economics of Natural Resource Management 
and Experiments 
159. Daruvala, Dinky (2006), Experimental Studies on Risk, Inequality and Relative 
Standing 
160. García, Jorge (2007), Essays on Asymmetric Information and Environmental 
Regulation through Disclosure 
161. Bezabih, Mintewab (2007), Essays on Land Lease Markets, Productivity, 
Biodiversity, and Environmental Variability 
162. Visser, Martine (2007), Fairness, Reciprocity and Inequality: Experimental 
Evidence from South Africa 
163. Holm, Louise (2007), A Non-Stationary Perspective on the European and 
Swedish Business Cycle 
164. Herbertsson, Alexander (2007), Pricing Portfolio Credit Derivatives 
165. Johansson, Anders C. (2007), Essays in Empirical Finance: Volatility, 
Interdependencies, and Risk in Emerging Markets 
166. Ibáñez Díaz, Marcela (2007), Social Dilemmas: The Role of Incentives, Norms 
and Institutions 
167. Ekbom, Anders (2007), Economic Analysis of Soil Capital, Land Use and 
Agricultural Production in Kenya 
168. Sjöberg, Pål (2007), Essays on Performance and Growth in Swedish Banking 
169. Palma Aguirre, Grisha Alexis (2008), Explaining Earnings and Income 
Inequality in Chile 
170. Akay, Alpaslan (2008), Essays on Microeconometrics and Immigrant 
Assimilation 
171. Carlsson, Evert (2008), After Work – Investing for Retirement 
172. Munshi, Farzana (2008), Essays on Globalization and Occupational Wages 
173. Tsakas, Elias (2008), Essays on Epistemology and Evolutionary Game Theory 
174. Erlandzon, Karl (2008), Retirement Planning: Portfolio Choice for Long-Term 
Investors 
175. Lampi, Elina (2008), Individual Preferences, Choices, and Risk Perceptions – 
Survey Based Evidence 
176. Mitrut, Andreea (2008), Four Essays on Interhousehold Transfers and 
Institutions in Post-Communist Romania 
177. Hansson, Gustav (2008), Essays on Social Distance, Institutions, and Economic 
Growth 
178. Zikhali, Precious (2008), Land Reform, Trust and Natural Resource Management 
in Africa 
179. Tengstam, Sven (2008), Essays on Smallholder Diversification, Industry 
 Location, Debt Relief, and Disability and Utility 
180. Boman, Anders (2009), Geographic Labour Mobility – Causes and Consequences 
181. Qin, Ping (2009), Risk, Relative Standing and Property Rights: Rural Household 
 Decision-Making in China 
182. Wei, Jiegen (2009), Essays in Climate Change and Forest Management 
183. Belu, Constantin (2009), Essays on Efficiency Measurement and Corporate 
Social Responsibility 
184. Ahlerup, Pelle (2009), Essays on Conflict, Institutions, and Ethnic Diversity 
185. Quiroga, Miguel (2009), Microeconomic Policy for Development: Essays on 
Trade and Environment, Poverty and Education 
186. Zerfu, Daniel (2010), Essays on Institutions and Economic Outcomes 
187. Wollbrant, Conny (2010), Self-Control and Altruism 
188. Villegas Palacio, Clara (2010), Formal and Informal Regulations: Enforcement 
and Compliance 
189. Maican, Florin (2010), Essays in Industry Dynamics on Imperfectly Competitive 
Markets 
190. Jakobsson, Niklas (2010), Laws, Attitudes and Public Policy 
191. Manescu, Cristiana (2010), Economic Implications of Corporate Social 
Responsibility and Responsible Investments 
192. He, Haoran (2010), Environmental and Behavioral Economics – Applications to 
China 
193. Andersson, Fredrik W. (2011), Essays on Social Comparison 
194. Isaksson, Ann-Sofie (2011), Essays on Institutions, Inequality and Development 
195. Pham, Khanh Nam (2011), Prosocial Behavior, Social Interaction and 
Development: Experimental Evidence from Vietnam 
196. Lindskog, Annika (2011), Essays on Economic Behaviour: HIV/AIDS, 
Schooling, and Inequality 
197. Kotsadam, Andreas (2011), Gender, Work, and Attitudes 
198. Alem, Yonas (2011), Essays on Shocks, Welfare, and Poverty Dynamics: 
Microeconometric Evidence from Ethiopia 
199. Köksal-Ayhan, Miyase Yesim (2011), Parallel Trade, Reference Pricing and 
Competition in the Pharmaceutical Market: Theory and Evidence 
200. Vondolia, Godwin Kofi (2011), Essays on Natural Resource Economics 
201. Widerberg, Anna (2011), Essays on Energy and Climate Policy – Green 
Certificates, Emissions Trading and Electricity Prices 
202. Siba, Eyerusalem (2011), Essays on Industrial Development and Political 
Economy of Africa 
203. Orth, Matilda (2012), Entry, Competition and Productivity in Retail 
204. Nerman, Måns (2012), Essays on Development: Household Income, Education, 
and Female Participation and Representation 
205. Wicks, Rick (2012), The Place of Conventional Economics in a World with 
Communities and Social Goods 
206. Sato, Yoshihiro (2012), Dynamic Investment Models, Employment Generation 
and Productivity – Evidence from Swedish Data 
207. Valsecchi, Michele (2012), Essays in Political Economy of Development 
208. Teklewold Belayneh, Hailemariam (2012), Essays on the Economics of 
Sustainable Agricultural Technologies in Ethiopia 
209. Wagura Ndiritu, Simon (2013), Essays on Gender Issues, Food Security, and 
Technology Adoption in East Africa 
210. Ruist, Joakim (2013), Immigration, Work, and Welfare 
211. Nordén, Anna (2013), Essays on Behavioral Economics and Policies for 
Provision of Ecosystem Services 
212. Yang, Xiaojun (2013), Household Decision Making, Time Preferences, and 
Positional Concern: Experimental Evidence from Rural China 
213. Bonilla Londoño, Jorge Alexander (2013), Essays on the Economics of Air 
Quality Control 
214. Mohlin, Kristina (2013), Essays on Environmental Taxation and Climate Policy 
215. Medhin, Haileselassie (2013), The Poor and Their Neighbors: Essays on 
Behavioral and Experimental Economics 
216. Andersson, Lisa (2013), Essays on Development and Experimental Economics: 
Migration, Discrimination and Positional Concerns 
217. Weng, Qian (2014), Essays on Team Cooperation and Firm Performance 
218. Zhang, Xiao-Bing (2015), Cooperation and Paradoxes in Climate Economics 
219. Jaime Torres, Monica Marcela (2015) Essays on Behavioral Economics and 
Policy Design 
220. Bejenariu, Simona (2015) Determinants of Health Capital at Birth: Evidence 
from Policy Interventions 
221. Nguyen, Van Diem (2015) Essays on Takeovers and Executive Compensation 
222. Tolonen, Anja (2015) Mining Booms in Africa and Local Welfare Effects: Labor 
Markets, Women’s Empowerment and Criminality  
223.  Hassen, Sied (2015) On the Adoption and Dis-adoption of Household Energy and 
            Farm Technologies 
224.  Moursli, Mohamed-Reda (2015) Corporate Governance and the Design of Board 
            of Directors 
225. Borcan, Oana (2015) Economic Determinants and Consequences of Political 
Institutions 
226. Ruhinduka, Remidius Denis (2015) Essays on Field Experiments and Impact 
Evaluation 
227.  Persson, Emil (2016) Essays on Behavioral and Experimental Economics, 
Cooperation, Emotions and Health. 
228.  Martinangeli, Andrea (2017) Bitter divisions: inequality, identity and 
cooperation 
229.     Björk, Lisa (2017) Essays on Behavioral Economics and Fisheries: Coordination 
and Cooperation 
230.     Chegere, Martin Julius (2017) Post-Harvest Losses, Intimate Partner Violence    
             and Food Security in Tanzania 
231.     Villalobos-Fiatt, Laura (2017) Essays on forest conservation policies, weather  
            and school attendance 
232.  Yi, Yuanyuan (2017) Incentives and Forest Reform: Evidence from China