Professional Documents
Culture Documents
Access to this document was granted through an Emerald subscription provided by emerald-srm:331053 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
CAER
5,3 Determinants of repayment
performance of group lending
in China
328
Evidence from rural credit cooperatives’
program in Guizhou province
Zhang Qinlan and Yoichi Izumida
Graduate School of Agricultural and Life Sciences,
The University of Tokyo, Tokyo, Japan
Downloaded by Universite Laval At 16:45 09 April 2016 (PT)
Abstract
Purpose – The purpose of this paper aims to explore how borrower and group-level characteristics
affect repayment decisions of group borrowers by highlighting the case of rural credit cooperatives
(RCCs) in Guizhou province in Southwest China.
Design/methodology/approach – The Logit model was applied to test the determinants of
repayment performance of RCCs’ group lending. The authors used the survey data of 245 farm
households in Guizhou province, collected in 2008.
Findings – The empirical results indicate that there is a serious mismatch between joint liability
mechanisms and the social and economic conditions in rural China. Mechanisms such as threatening to
withhold defaulters’ future loans from RCCs failed to work. In addition, higher household incomes also did
not improve repayment performance. However, factors such as a higher degree of acquaintanceship in a
group, migrant income, and employment in government agencies, positively improved the chances of
repayment.
Practical implications – Group lending is more suitable for poorer areas with few opportunities for
migration and limited access to finance. In addition, constructing the trustworthy relationship between
micro-lenders and customers and designing diverse and flexible financial services to meet
heterogeneous demands are equally important.
Originality/value – This paper is an attempt to empirically explore the determinants of repayment
performance in group lending programs in China. The results provide meaningful policy implications
for the government and rural financial institutions.
Keywords China, Rural development, Rural finance
Paper type Research paper
1. Introduction
Group lending, one of the major innovations of microfinance, has been widely
replicated and adopted in developing countries over the past three decades. It is a
contractual innovation to overcome imperfect information in rural financial markets by
addressing four main problems: adverse selection, moral hazard, monitoring, and
enforcement (Besley and Coate, 1995; Ghatak and Guinnane, 1999; Gine and Karlan,
2010). Despite the widespread popularity of joint liability contracts, empirical studies
China Agricultural Economic Review have not shown complete success. One criticism is that the benefits of group lending
Vol. 5 No. 3, 2013
pp. 328-341 have been exaggerated and the method is often too rigid to meet borrowers’ needs
q Emerald Group Publishing Limited (Conning, 2005; Armendariz and Morduch, 2005). In addition, joint liability may lead to
1756-137X
DOI 10.1108/CAER-08-2012-0083 excessive tensions among peers and may even worsen the dropout rate. Today, many
micro-lenders have chosen flexible approaches, and tend to pay more attention to Repayment
individual liability contracts. performance of
China introduced microfinance in the form of joint liability in 1993; subsequently,
group lending schemes have expanded significantly. Currently, there are many types group lending
of microfinance institutions (MFIs) in rural China. Among them, rural credit
cooperatives (RCCs) have been the most popular ones. Statistics from the China
Banking Regulatory Commission (CBRC) show that by the end of 2007, there were 329
over 78 million rural households benefitting from RCCs’ microloans, accounting for
33.2 percent of overall households in China. The outstanding amount of rural group
loans distributed by the RCCs, the Rural Cooperative Banks, and the Rural Commercial
Banks during this period stood at 135.1 billion Yuan (He et al., 2009).
Although RCCs’ microloan programs generated positive profits soon after their
operations, many researchers claimed that they still faced difficulties such as improving
operating efficiency, liberalizing interest rate caps and so on (He et al., 2009). Among these
difficulties, it is crucially important to improve loan recovery performance, especially that
Downloaded by Universite Laval At 16:45 09 April 2016 (PT)
of group lending programs. Yang (2012) pointed out that in some areas, default rates of
group loans reached 10-50 percent and resulted in considerable losses. In addition, the scale
of group lending in many branches began to shrink (Du, 2008). According to the Almanac
of China’s Finance and Banking (2003-2009), from 2002, the outstanding loans of group
lending have been much lower than those of individual lending. Both the academic
researchers and practitioners have questioned the sustainability of group lending in China.
Therefore, it is critical to examine the mismatch between joint liability mechanisms
and the social and economic conditions in China, and explore how borrower and
group-level characteristics affect borrowers’ repayment performance. We adopt Logit
regression to analyze the case of RCCs in Guizhou province in Southwest China.
The sample for this study is the survey data of 245 farm households collected in 2008.
The rest of the paper is organized as follows: Section 2 briefly reviews the rural
financial system and the development of group lending in China. Section 3 presents the
theoretical framework, while Section 4 discusses the data used in this study.
In Section 5, the econometric results are presented and the final section offers
conclusions and suggestions for the improvement of microcredit provisions in China.
programs were those conducted by the RCCs. In July 1999, the PBOC issued the
Provisional Method of Microcredit Loans Management of RCCs for Rural Households,
initiating the practice of individual and group lending activities. According to Gale and
Collender (2006), in 2003, about half of the RCCs had inducted group lending programs
and provided $9.4 billion loans to 9.5 million clients. By June 2005, the outstanding
loans had increased to $11.7 billion. Therefore, the RCCs had gained market leadership
in rural areas.
The third period, beginning in 2005, can be regarded as a normalization or
adjustment phase. New types of MFIs were encouraged to enter the rural financial
market. In 2005, the PBOC launched a pilot program of credit-only microloan
companies. By the end of 2006, the CBRC granted village/township banks, credit-only
companies, and postal saving banks to develop microloan businesses in China.
Although these policies were expected to promote competition and expand credit
opportunities in low-income communities, no significant improvement has been
observed as yet.
3. Theoretical model
A series of theoretical literature on microfinance has proposed numerous models to
explain how the joint liability approach works before and after the loans’ disbursement
(Ahlin and Townsend, 2003; Brehanu and Fufa, 2008; Armendariz and Labie, 2010).
Before the disbursement of loans, most lenders face the problem of adverse selection.
In principle, group lending with joint responsibility can mitigate this inefficiency by
encouraging the applicants to self-select their best partners (Ghatak and Guinnane,
1999; Van Tassel, 1999; Godquin, 2004; Armendariz and Morduch, 2005). It is expected
that borrowers from the same village have sufficient information about incomes,
repayment capacities and creditworthiness of neighboring households, and that they
will use this information to form homogeneous groups. However, after the loan
disbursements, the MFIs may confront moral hazard and enforcement problems.
Available literature shows that threatening not to refinance defaulters or offering
larger loans to borrowers who repay their debts, creates an incentive for peer
monitoring, peer pressure, and intra-group help among the borrowers (Ghatak and
Guinnane, 1999; Armendariz and Morduch, 2005). Under the threat of being excluded
from future loans if one group member defaults, the group has the incentive to monitor Repayment
their peers to use the loans in profitable ways, and to exert pressure to make the performance of
potential defaulter reconsider his decision. Further, each member will support the
others if they face repayment difficulties. group lending
While group members have many incentives to achieve high repayment rates, they
still face a prominent problem of high monitoring costs, even for members living in
close proximity. Armendariz and Morduch (2005) argue that group lending can 331
encourage borrowers to help each other only if peer monitoring is not very costly and
social sanctions are sufficiently strong. In addition, under some conditions, borrowers
under group lending contracts may collude and this creates serious risks for MFIs
(Besley and Coate, 1995; Laffont and Rey, 2003).
The analytical framework here involves the borrowers’ repayment behavior after
loan disbursement. When the project returns are realized, how do group members
make their repayment decisions? Under the individual lending contracts, borrowers
simply compare their own costs and benefits of default. However, under group-lending
Downloaded by Universite Laval At 16:45 09 April 2016 (PT)
contracts, their decisions are also influenced by other members. The repayment
mechanisms such as peer monitoring, mutual help, and peer sanctions help in
heightening motivation for repayment. At the same time, these methods result in high
monitoring costs. Thus, borrowers must evaluate the overall costs as compared to the
benefits of default.
Based on the above analysis, we assume that a borrower has an opportunity to obtain
a loan from one MFI. He could generate p units of net income from the microloan
investment project and w units from other activities. When the loan is due, the amount
including both principal and interest r should be repaid. At the same time, he may
encounter some income shocks during the loan period, such as natural disasters,
sickness, unemployment, and so on. These shocks may affect his repayment behavior.
Here, we denote l as the losses due to negative shocks, thus the household’s total
expected income can be represented as p þ w 2 l. In the case of default, the borrower
may not repay the principal and interest, but his access to loans from this MFI would
cease. Denote the present value of all future loan profits as B, and the function of private
profits as p(· ). Under the individual lending contracts, if the net expected benefits of
default are smaller than B, the borrower should choose to repay. That is:
Under the group lending contracts, the defaulter should also receive social sanctions (S )
from other group members. Therefore, the borrower will repay only if:
Another factor that affects a borrower’s repayment decision is monitoring costs. When a
borrower suspects that monitoring costs are higher than the evaluation of his
investment’s income, he may be prone to default and evade the responsibility of peer
monitoring. Let C be the monitoring costs. Then equation (2) becomes:
4. Data
Downloaded by Universite Laval At 16:45 09 April 2016 (PT)
We collected data in Chishui city, in Guizhou province, from March through April of
2008. The surveys covered 245 households in four densely populated villages. In each
village, we randomly handed out 80 questionnaires to households that had ever
received group loans from RCCs and helped them respond to all the questions. The
questionnaire included items about borrowers’ socioeconomic characteristics, labor
market participation, land usage, household income, negative shocks, financial
activities, repayment performance, and group structure.
Guizhou is a very poor province located in Southwest China. In 2009, more than
20 percent of the population was still below the poverty line. In comparison to other
provinces, poor people in the province are mainly concentrated in remote and
mountainous areas, where the vast majority of households are engaged in agricultural
production. In the early 1990s, microfinance projects started in Guizhou, and
provisions of loans through microfinance become one of the most important tools for
poverty alleviation.
We selected Chishui, a moderately developed city in the northwest of Guizhou as the
survey area. The city is in a large mountainous area, with poor traffic conditions, and a
poorly developed industrial sector.
Table I highlights the major socio-economic characteristics of the sample
households. First, a majority of sample households earn their income from multiple
sources. The agricultural land of both paddy and dry land area was small in the survey
area, but almost all households were engaged in agriculture. About 82.9 percent were
planting bamboo. In addition, over half of the sample households had at least one
member working outside the home, and 12.7 percent were operating off-farm
businesses (OFB). Thus, agricultural income, bamboo income, and non-farming income
made up the majority of households’ overall income. Second, most of the sample
households were lower-income households. In the survey, each respondent was asked
to appraise his household income level. The proportion of households assessed as
low-income, middle-income, and upper-income was 36.7, 59.2, and 4.1 percent,
respectively. In addition, a strikingly large percentage of sample households had low
education levels.
Concerning households’ financing sources (this information is not listed in the table),
the informal sector played an important role. Most (73.9 percent) of the
sample households reported to rely first on relatives when they needed money.
Repayment
Household characteristics Proportion Mean
performance of
Age of the household head group lending
Less than 30 7.3 41.0
30-50 77.6
More than 50 15.1
Education level of the household head 333
Illiterate or primary school 25.3
Junior middle school 62.0
Senior middle school 10.6
College or above 2.1
Income level
Low income 36.7
Middle income 59.2
Upper income 4.1
Whether the family owns OFB
Downloaded by Universite Laval At 16:45 09 April 2016 (PT)
Yes 12.7
No 87.3
Total paddy area
Less than 0.1 hectares 55.1 0.11
0.1-0.2 hectares 27.8
More than 0.2 hectares 17.1
Total dry land area
Less than 0.1 hectares 53.5 0.11
0.1-0.2 hectares 34.7
More than 0.2 hectares 11.8
Total bamboo grove area
Without bamboo grove 7.1 0.17
Less than 0.1 hectares 49.8
More than 0.1 hectares 33.1
NFM working outside the home
0 43.7 0.82
1 or 2 members 53.9 Table I.
More than 3 members 2.4 Socioeconomic
characteristics of the
Source: Field survey in Guizhou sample households
Another 14.3 percent attached the most importance to borrowing from friends. However,
37.6 percent of the respondents still considered RCCs as their second most important
source of borrowing.
Table II shows the relationship between loan usage and repayment rates. We found
that the borrowers used RCCs’ microloans for limited purposes: about 46.9 percent of
the households used their loans for building rural roads, 22.9 percent for agricultural
production purposes, and 14.3 percent for children’s education. Only a few loans were
used for building houses, medical treatment, living expenses, marriage and funeral
expenses, and so on.
The information on other key variables of the group lending contracts is presented
in Table III. Loan amounts ranged from 800 to 15,000 Yuan, and the group size often
exceeded five persons, sometimes going up to ten persons. Moreover, group members
have close relationships with other members in terms of kinship, residential distance,
and cooperation.
CAER Before moving to the econometric analysis, we try to examine the relationship
5,3 between repayment performance and loan and group characteristics. For the analysis,
borrowers are classified into two categories: those who have fully made their required
loan repayments by the due date are classified as “repaid”, and those who have missed
repayments are classified as “default” (Sexton, 1997; Coke, 2002). Field data shows that
serious repayment problems existed in the survey area. The default rate was
334 34.7 percent (65 households), and only 65.3 percent (160 households) repaid both the
principle and interest before maturity.
In order to analyze the correlation between loan size and borrowers’ repayment
performance, we categorized the sample households into three groups by individual
loan size: less than 3,000 Yuan, 3,000-5,000 Yuan, and more than 5,000 Yuan.
No obvious difference has been found in the three groups. However, the repayment
5. Econometric results
Empirical model
To test the hypotheses discussed in Section 3, we adopt the Logit model. The model
Downloaded by Universite Laval At 16:45 09 April 2016 (PT)
aims to identify the relationship between the probability of repaying a loan and a
vector of explanatory variables such as: borrower and group-level characteristics,
borrower’s expectation of future loans from RCCs, intra-group monitoring costs, and
social sanctions. The dependent variable yi is dichotomous with a value of 1 if the
sample borrower i completed repayment before the due date. 1i is an error term.
The model is formulated as the equation below:
1
pi ¼ prob ð yi ¼ 1j zi Þ ¼ ;
1 þ e 2zi
Taking the natural logarithm:
pi X X X X
zi ¼ log ¼aþ b1i x1i þ b2i x2i þ b3i x3i þ b4i x4i þ 1i
1 2 pi
The first category (x1i) indicates the number of family members (NFM), the number of
household members in the labor force (NFL), age of the household head (AGE), sex of
the borrower (SEX), and loan size (AOL). This category of variables is used to test
whether certain individual characteristics or loan size helps the client heighten the
possibility of repayment.
The second series of variables (x2i) indicates a household’s income and income
related factors. We use six variables: a dummy for investment profitability (OCP),
which denotes whether the returns on the microloan investment can cover the principal
and interest required to repay; owned farming area (FA); owned bamboo area (BA); a
dummy for OFB such as managing shops for the sales of candy, beverages, and so on
in the village; the number of household members working outside the home (NWO);
and a dummy for external shocks such as sickness, crop losses, or owing other debts
during the loan period (HNS). Diverse income sources, including both farming and
non-farming income, are expected to improve the repayment performance.
The third category (x3i) is used to test H1 on the borrower’s expectation of future
loans from RCCs. Three variables are taken into consideration: the number of outside
loan (NOL) opportunities besides RCCs, a dummy variable for the need of a loan from
RCCs in the near future (NL), and a dummy variable for whether RCCs are the
borrower’s first choice for a loan (RFC).
CAER Finally, the variable denoted as x4i reflects the homogeneity among group members
5,3 and the intensity of social ties. Since homogeneity and social ties are expected to affect
both intra-group monitoring costs and social sanctions, we test H2 and H3 together.
Five variables are used here: percentage of group members who are relatives (POR),
percentage of group members who are from the same village (POV), percentage of
group members who have cooperative relationships with the respondent (POC),
336 farthest geographic distance from the respondent’s house to a peer’s house (FGD), and
a dummy variable for whether any family member is a village official (VO). Higher
POR, POV, or POC show that the borrower has more relatives, friends, or cooperators
among the group members. These variables enable us to investigate whether kinship
or friendship provides any advantage in the monitoring, mutual help, or sanction
process, or if it increases the likelihood of default due to collusion. Yet, we did not
control village dummy variables in this study because all the sample villages had
roughly the same economic, social, and geographic conditions.
Downloaded by Universite Laval At 16:45 09 April 2016 (PT)
Empirical results
The results of the estimations of repayment behavior are reported in Table IV. From
this Table IV we obtain the following findings.
First, the loan size (AOL) was not correlated with repayment behavior. This result
indicates that a smaller loan was not easier for the borrower to repay. Many borrowers
expressed that they felt a certain reluctance to repay if other members defaulted
households in rural China. This group usually has higher migrant income, better
education, and a stronger sense of following the law. The dynamic incentives of group
lending may be more effective in inducing these people to make rational investments
and to refrain from strategic defaults. Other income sources, such as owning a larger
FA or BA, owning OFB, or having positive profits on a microloan investment (OCP),
were not shown to significantly improve repayment performance. The estimation
result of the variable of OCP was worth noting. Most borrowers reported that they
were grouped randomly with their preferred members. Low-risk borrowers who are
more likely to make positive profits (high value of OCP), were expected to team up with
other low-risk borrowers. However, there is no evidence that so-called assortative
matching increased the possibility of repayment. One explanation that some low risk
clients chose to default was attributed to strategic default. Many sample households
claimed that they preferred individual lending because it relaxed compulsory liability.
In addition, the dummy variable HNS (whether the family experienced negative income
shocks), was found to have a negative impact at the 1 percent level. This means that
almost all households that suffered from external shocks failed to receive any
monetary help from their peers. Therefore, our estimation results have not provided
clear evidence to support the notion that a higher household income leads to better
repayment performance.
Variables such as the number of outside loan opportunities besides RCCs (NOL),
whether the RCCs were the borrower’s first choice for a loan (RFC), and whether the
borrower needed the RCCs’ loan in the near future (NL) were used to test H1.
Surprisingly, RFC and NL were found to be significantly negative. This means that the
threat of stopping future lending failed to work. To understand the reasons, it is
necessary to consider the background of the RCCs’ program at the survey site. In our
samples, large parts of group loans were used to build rural roads or plant bamboo, not
for sustained investments. Hence, when borrowers could not find new investment
opportunities, their capital demands decreased dramatically. In addition, the high
default rate was supposed to be related to the failure of some subsidized microloan
programs in the 1990s. Households regarded those poverty alleviation loans as gifts from
the government. Therefore, when they received microloans from the RCCs, they treated
the loans as another dole from the government. The variable NOL was also significantly
CAER negative at the 5 percent level. Clients who had easier access to borrowing from relatives
5,3 or friends had less incentive to minimize their default. Thus, H2 was not proved.
Finally, group member homogeneity and social ties were tested. We found significant
evidence that when a group had a higher percentage of members from the same village
(POV) or had cooperative relationships with the respondent (POC), repayment
performance improved. Therefore, we could draw one conclusion that the more intensive
338 the intra-group communication was, the less repayment difficulties a group might face.
On one hand, group members were willing to support their friends or relatives to cover
repayment problems. On the other hand, they faced much heavier social sanctions in the
case of voluntary default. This could be especially true in the remote and less-developed
villages, where financing sources are limited, and members are more sensitive to being
taunted or rejected by others. Of course, we should pay attention to the possibility of the
problem of endogeneity as well. Households with high creditworthiness might have less
trouble finding peers in their village, which could explain why they usually had higher
repayment capacities. Hence, a better repayment rate may come from the
Downloaded by Universite Laval At 16:45 09 April 2016 (PT)
6. Conclusions
The RCCs’ group lending program has been developing for over ten years in China.
However, most studies have not supported the view that the RCCs successfully
achieved financial sustainability. This paper aimed to shed light on the key factors
behind the failure by applying econometric methods to the case of Guizhou province in
Southwest China. From the collected field data, it is clear that the joint liability
contracts have not achieved a high rate of success as theories claimed. The delinquency
rate in the survey area was as high as 34.7 percent. RCCs failed to develop sustainable
programs in practice.
A theoretical model has been constructed to analyze the repayment behavior of
borrowers. Group members are expected to repay when the costs of default exceed the
benefits. Empirical results show that higher household income did not improve
repayment performance except migrant income, and mechanisms of threatening to
withhold defaulters’ future credit opportunities also did not work. However, group
homogeneity and social ties significantly improved the repayment rate in the survey
site. When a group contained a higher proportion of acquaintances, the chance of
repayment significantly increased. In addition, having a family member working in
government agencies also effectively prevented default.
However, we cannot simply make a conclusion here that strengthening group
homogeneity or social ties is bound to increase the repayment rate. As shown, it is
possible to lead to serious collusion risk. Thus, MFIs have to be careful in discussing
the function of social ties in practice. Our opinion is that group lending is more
attractive for poorer areas with few opportunities for migration and limited access to
finance. Here, trust between households is easier to build and people are more willing
to provide mutual help. On the contrary, in the richer villages, individual action may be
more effective.
How to overcome the mismatch between group lending and the demand for Repayment
financial services in rural China is still a challenge. Currently, both RCCs and other performance of
MFIs provide limited and rigid menus of microfinance products, but rural individuals
from different areas show diverse demands. They need a broader range of products group lending
with more flexible designs in the form of collateral, credit terms, and price settings, as
well as renegotiation approaches for the defaulting clients. For this reason, we think
the government should encourage a broader participation of all types of MFIs, 339
especially NGOs, which have been successful in serving the unbanked poor in specific
areas and have realized financial sustainability there.
In addition, it is important for MFIs to construct a trustworthy relationship between
themselves and their customers through a better understanding of their clients as well as
by promoting customers’ loyalty. Only in this way, are MFIs likely to maintain
long-term cooperation with their clients and improve their financial profitability.
Moreover, it is critical to choose suitable districts by checking the level of household
assets and their level of financial discipline. Last, but not least, we should recognize that
Downloaded by Universite Laval At 16:45 09 April 2016 (PT)
a good product design cannot guarantee the success of rural financial reform. Increasing
households’ wealth, building a successful microfinance culture, and reducing the
government intervention in rural financial markets are more important in China.
However, the problem of endogeneity in this paper has remained unsolved for
reasons of sample selection. We failed to control a broader range of unobserved
characteristics that may affect individual repayment decisions. In the future,
a well-designed laboratory experiment should be introduced to test the effectiveness of
joint liability contracts.
References
Ahlin, C. and Townsend, R. (2003), “Using repayment data to test across models of joint liability
lending”, working paper, Development of Economics, University of Chicago, Chicago, IL.
Almanac of China’s Finance and Banking (2003-2009), China Finance Press, Beijing.
Armendariz, B. and Labie, M. (2010), “Introduction and overview: an inquiry into the mismatch
in microfinance”, in Armendariz, B. and Labie, M. (Eds), Handbook of Microfinance,
World Scientific Publishing, New York, NY, pp. 3-13.
Armendariz, B. and Morduch, J. (2005), The Economics of Microfinance, The MIT Press,
Cambridge, MA.
Besley, T. and Coate, S. (1995), “Group lending, repayment incentives and social collateral”,
Journal of Development Economics, Vol. 46 No. 1, pp. 1-18.
Brehanu, A. and Fufa, B. (2008), “Repayment rate of loans from semi-formal financial institutions
among small-scale farmers in Ethiopia: two-limit Tobit analysis”, The Journal of
Socio-Economics, Vol. 37 No. 6, pp. 2221-2230.
CGAP (2005), Rural Credit Cooperatives in China, PlaNet Finance, available at:
www.microfinancegateway.org/gm/document-1.9.26797/76.pdf
Coke, R. (2002), “Microfinance borrower default: evidence from the Philippines”, doctoral thesis,
American University, Washington, DC.
Conning, J. (2005), “Monitoring by delegates or by peers? Joint liability loans under moral
hazard”, working paper, Department of Economics, Hunter College and The Graduate
Center, City University of New York, New York, NY, June.
CAER Du, X. (2008), “The current supply of microfinance service in China”, paper presented at International
Symposium Geneva, Switzerland, October 1-2, available at: www.microfinancegateway.
5,3 org/p/site/m/template.rc/1.9.40223/
Gale, F. and Collender, R. (2006), “New direction in China’s agricultural lending”, Outlook Report,
WRS 0601, USDA/ERS, Washington, DC.
Ghatak, M. and Guinnane, T. (1999), “The economics of lending with joint liability: theory and
340 practice”, Journal of Development Economics, Vol. 60 No. 1, pp. 195-229.
Gine, X. and Karlan, D. (2010), “Group versus individual liability: long term evidence from
Philippine microcredit lending groups”, working paper, Economics Development,
Yale University, New Haven, CT.
Godquin, M. (2004), “Microfinance payment performance in Bangladesh: how to improve the
allocation of loans by MFIs”, World Development, Vol. 32 No. 11, pp. 1909-1926.
He, G., Du, X., Bai, C. and Li, Z. (2009), “China microfinance assessment industry report”,
working paper, China Association of Microfinance, Beijing, February 17.
Downloaded by Universite Laval At 16:45 09 April 2016 (PT)
Laffont, J. and Rey, P. (2003), “Moral hazard, collusion and group lending”, working paper, IDEI,
Toulouse and University of Southern California, December 23.
Park, A. and Ren, C. (2001), “Microfinance with Chinese characteristics”, World Development,
Vol. 29 No. 1, pp. 39-62.
PBOC (2010), China Rural Finance Service Report, People’s Bank of China, available at: www.pbc.
gov.cn/image_public/UserFiles/goutongjiaoliu/upload/File/%E4%B8%AD%E5%9B%
BD%E5%86%9C%E6%9D%91%E9%87%91%E8%9E%8D%E6%9C%8D%E5%8A%
A1%E6%8A%A5%E5%91%8A2010.pdf
Sexton, D.E. (1997), “Determining good and bad credit risks among high- and low-income
families”, Journal of Business, Vol. 50 No. 2, pp. 236-239.
Van Tassel, E. (1999), “Group lending under asymmetric information”, Journal of Development
Economics, Vol. 60 No. 1, pp. 3-25.
Yang, F. (2012), “Research on optical rural household group lending model and countermeasures:
based on three typical cases”, Rural Finance Research, Vol. 2, pp. 10-15.
Appendix
Description of variables
Table AI.
341
performance of