WPS5808 Policy Research Working Paper 5808 After the Microfinance Crisis Assessing the Role of Government-Led Microcredit Alternatives Jordi de la Torre Xavier Giné Tara Vishwanath The World Bank Development Research Group Finance and Private Sector Development Team September 2011 Policy Research Working Paper 5808 Abstract In light of the recent microfinance crisis in South India, cooperatives tend to be used as political instruments government-run institutions in general, and primary and, as a result, borrowers prioritize all debt obligations agricultural credit cooperatives in particular, may end (microfinance institutions, informal lenders, etc.) before up playing a larger role in the provision of financial repaying their primary agricultural credit cooperative services for the poor. Using survey data collected in loans. The authors suggest that if the performance 2007 from three districts in Andhra Pradesh, this paper of primary agricultural credit cooperatives does not assesses the performance of 72 primary agricultural improve, a larger government role in the supply of credit credit cooperatives and finds lack of training among the may undermine the culture of repayment. management. In addition, primary agricultural credit This paper is a product of the Finance and Private Sector Development Team, Development Research Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at xgine@worldbank.org and tvishwanath@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team After the Microfinance Crisis: Assessing the role of government-led microcredit alternatives Jordi de la Torre World Bank Xavier Giné World Bank and BREAD Tara Vishwanath World Bank JEL Codes: C81, G21, O12, O16 De la Torre: j.delatorreanglada@gmail.com, Giné:xgine@worldbank.org, Vishwanath:tvishwanath@worldbank.org. We would like to thank the valuable comments of Giovanna Pernushi and Niraj Verma. The views expressed in this paper are those of the authors and should not be attributed to the World Bank, its executive directors, or the countries they represent. 1. Introduction The microfinance revolution that started 30 years ago has resulted, as of last count in more than 3,500 microfinance institutions (MFIs) that collectively serve almost 150 million clients worldwide (Daley-Harris, 2009). Proponents of microfinance claim that by using innovative contract features, such as joint liability, frequent repayments and escalating loan sizes, MFIs have managed to achieve high repayment rates while serving a fraction of the world’s poor. Without proper access to financial services, the poor have to rely on their own scare resources and uncertain incomes to invest in education, start new businesses or enhance the productivity of the land they cultivate. Underinvestment in these crucial assets tends to keep their overall productivity low, making it very hard for them to escape the vicious cycle of poverty. The recent crisis in AP, however, has called into question the role that MFIs play in providing access to financial services to the poor. The crisis was triggered by an ordinance promulgated by the government of Andhra Pradesh1 in response to allegations of over-indebtedness of clients due to careless lending practices and rumors of debt-related suicides among MFI borrowers. While there may be some truth to these allegations, partly a result of the rapid growth that MFIs have experienced in recent years, especially in AP, MFIs do serve an important segment of society, particularly those individuals without assets that can be used as collateral. Repayment rates remain low today, and if proper measures are not successfully implemented, the microfinance sector may wither or disappear altogether.2 If this were to happen, what then are the feasible credit alternatives? And will they contribute to the efficiency of the credit markets for the poor? 1 The ordinance required that repayment collections occur at panchayat offices, waived retroactively loans if twice the principal had already been repaid, and required registration of MFIs with district authorities who could, at any time, cancel it. 2 Most of the measures proposed come from the report written by the Malegam committee, set up in October 2010 following the ordinance promulgated by the government of AP. The Committee was tasked with proposing regulations for the microfinance sector to curb issues related to over-borrowing, multiple-lending and forceful recovery methods. The proposals include, among others, the creation of a credit information bureau, the adoption of a customer protection code, the disclosure of interest rate calculation and perhaps more controversial, a cap on lending rates and the limits on microloan sizes and incomes of borrowers (see, for example, http://blogs.cgdev.org/open_book/2011/01/retort-to-the-malegam-report.php for further details). 2 Perhaps one consequence of a diminished role of MFIs as credit providers for the poor may be an enhanced role for the government. In order to assess whether this would be desirable, and given its relevance, a closer look at the evidence regarding the effectiveness of institutions with significant government involvement in providing access to credit for the poor is clearly warranted. In this article we report findings from a survey conducted in three districts in AP in 2007 designed to assess the performance of rural credit cooperatives and in particular Primary Agricultural Credit Cooperatives (PACS).3 The experience of PACS may prove useful to shed light into the debate about the role of government intervention in access to credit for the poor. We find evidence that PACS were used as political instruments and that borrowers responded by prioritizing all debt obligations (MFIs, informal lenders, etc) before repayment of PACS loans. This indicates that the culture of repaying institutions with significant government involvement may have been low and that a private sector solution that is shielded from the government’s interference may be desirable. The article proceeds as follows. We first describe the data and the sources of credit used by the sample of households interviewed, we then discuss the PACS management practices and the impact of politics in the culture of repayment. Section 4 concludes. 2. Data and Sources of Credit Our data includes 72 PACS from the districts of Anantapur, Mahabubnagar and Vizianagaram in Andhra Pradesh with audited recovery rates as of June 2004 between 25 and 75 percent. We chose a sample of PACS in each of the 3 districts with repayment rates in 2004 just above and below the thresholds of 30 and 50%. We then chose a representative sample of borrowing households and another sample of non-borrowers from the villages where the PACS operates. Since we have the proportion of PACS members in each village, we can weight responses from 3 Primary Agricultural Cooperative Societies (PACS) are the lowest tier of the rural cooperative banking system. They focus primarily on providing short term (seasonal) credit for agricultural purposes and are regulated by NABARD. According to the Reserve Bank of India, there were 94,942 PACS in 2008, each one serving 7 villages on average, and a total membership of 131 million individuals, 79 million of which borrowers (Johnson and Meka 2010). 3 the members and non-members and report summary statistics that are representative of the villages where PACS operate. In 2007, we interviewed 1,060 farmers from 106 different villages, of which 847 were PACS members and 213 were non members. The population in these villages is mostly engaged in agriculture (82.6 percent) and only about half of household heads are literate.4 Half of the sample own less than 5.5 acres of land. 6.3 percent of our farmers are marginal farmers (owning less than 1 acre), 40.2 percent are small farmers and the remaining 54.3 percent own more than 5 acres of land. Among PACS borrowers, 8.1 percent are marginal farmers and almost 42 percent are large farmers. Thus, we find little evidence that PACS prioritize marginal and small farmers. This lack of targeting is consistent with cross-country evidence by Beck, Demirguc-Kunt and Martinez-Peria (2008) which find no significant association between greater government ownership of banks and financial access across countries. Table 1 reports the preferred source of credit and actual use by source at the time of the survey in 2007. It is noteworthy to observe the relatively limited role of self-help groups (SHGs)4 both as the preferred as well as actual choice. The data show that only 6 percent of households had a loan with a SHG, and perhaps as a consequence, they were almost never the preferred source for credit (less than 2 percent of households rank SHGs above other sources). This is not too surprising, as SHGs target the poorest households in the community. Marginal farmers exhibited a slightly higher preference for SHG loans than small and large farmers but they ended up borrowing the least from this source (less than 10 percent of households). Informal loans, quite surprisingly, were the preferred and used source by a large margin among all groups of farmers. This is probably due to the higher flexibility they offer. Loans from PACS were the preferred source for marginal farmers, and indeed the dominant source of credit for marginal and the second source for small farmers, below informal credit. In sum, the larger penchant of marginal farmers for SHGs did not translate into greater access relative to small and large farmers who also borrowed more heavily from formal and informal sources. 4 Another government initiative that is often seen as a viable alternative to microfinance, especially in AP, is the SHG model based on teaching community members how to provide and manage financial services among themselves. Poor women come together in a group that meets periodically, each contributes savings to a group account that is then on lent to members. SHG are an example of a “promoter� approach to use the term coined by Stuart Rutherford in his 2000 book The Poor and Their Money. 4 We use stratification weights to report population averages. 4 These differences in lending methodology are echoed in Table 2 which reports the characteristics of loans in the sample by credit source. The main differences in the characteristics lie in collateral requirements, interest rate, maturity and average loan size. Both PACS and formal sources of credit require collateral. In terms of changes in the interest rate, informal loans charge nearly twice as much as loans from PACS and they also have a shorter maturity. The problem with PACS, however, is that the average loan size was smaller (a median of 10,000 Rs. versus 20,000 Rs. for informal sources) and collateral is required. Indeed, half of the PACS borrowers said they could not cover expenses in agricultural inputs with a PACS loan plus cash at home. Three in four said they would have liked to borrow more and most of them also borrowed from other sources besides PACS, mainly from informal sources. This was especially true for larger farmers. 5 3. PACS Management We also inquired into the management practices of PACS and the evidence pointed to a lack of the proper incentives to improve repayment. Secretaries of PACs, for instance, were remunerated by and large using a fixed compensation, unrelated to repayment or other performance measures. Secretaries had on average only 1.5 years of formal education. They had been trained in accounting for less than two months and they only had an average of three years of experience as accountants. Presidents, on the other hand, had higher qualifications, but secretaries play a more important role in the day to day operations. Perhaps not surprisingly, the quality of record keeping for more than 50 percent of PACS was poor, and only 5 percent of PACS provisioned loan losses. Questions in the survey about governance presented clear evidence of misalignment between the perceptions of members and those of PACS committee members and presidents. The perception of outside influence is a telling example: between 20 and 30 percent of members perceived that admission and termination of members and credit decisions were influenced by outsiders. In contrast, no president or committee member thought so. In addition, nearly 40 percent of members thought that loan decisions of a PACS tend to favor certain members, while less than ten percent of presidents and committee members thought so. Respondents were usually skeptical of the motives of PACS management. They believed, with a probability of close to 50 percent, that PACS’s presidents do not look after members’ interests, and they assigned a probability of 40 percent to the proposition that the president would abuse his power on his own interest. 4. The role of politics A first hint of the potential role of politics comes from looking at the political connections that PACS members and management had relative to non members. The data showed that close to 60 percent of PACS members had the same political affiliations as the head or ward member of Panchayat. In the case of presidents or community members, this figure climbed to nearly 75 percent and most committee members had the same political affiliation as the president. Consistent with this finding, there is evidence that links the behavior of government-run credit providers to the electoral cycle. Cole (2009) for example finds that agricultural credit increases 6 by 5–10 percentage points in an election year, particularly in districts with high level of electoral contestation. This pattern is not found in nonelection years or in lending by private banks. Cole also shows that this capture is costly as elections affect negatively loan repayment, and election- year credit booms do not increase agricultural output either. An explanation for why government-run lending institutions in India are more generous during elections comes from Cole, Healy and Werker (2009) who using rainfall, public relief, and election data examine the reaction of voters to the response of governments to adverse shocks. They find that voters only respond to government relief efforts during the year immediately preceding an election.5 Unfortunately, the culture of subsidies and frequent government reliefs undermines the culture of repayment as borrowers take advantage of the lax enforcement of credit contracts. The data clearly show that farmers anticipated that, with probability close to 50 percent, a relief package that benefits PACS (but not MFIs or informal lenders) would be announced in case of a drought. Since farmers do not benefit from a relief package if they repay the loan before the announcement is made, all farmers have an incentive to wait (even after the loan due date) until the announcement (if any) is made. We also asked farmers about their perceived probability of different events related to loan repayment, and their answers showed similar patterns. For example, farmers perceived that their loan with a PACS would be rescheduled in the event of a drought with a 70 percent chance, while this probability dropped to 55 percent for Commercial Banks and only 25 percent for MFIs. The same pattern occurs with the perceived probability of loan forgiveness. Besides, PACSs appeared more likely to issue loans in amounts lower than requested (the difference between PACS and MFIs exceeds 30 percentage points) and less likely to issue new loans even if the previous ones were fully repaid. PACS were less trustworthy of savings and more likely to have a share of the loan deducted as a bribe. Not surprisingly, the loans most likely to suffer from this problem were those issued by PACS. MFI loans, on the contrary, were perceived as less likely to do so. There was a nearly 20 percentage point gap between the two loan sources. 5 The international evidence shows the same patterns. Dinç (2005), for example, shows that increased lending by government-owned banks right before elections is not specific to India but can be observed in data from 22 developing countries. Khwaja and Mian (2005) also find that in Pakistan, politically connected firms are able to secure larger and cheaper loans from state-owned banks and default on these loans much more than other non- connected firms. 7 Besides, loans from PACS were perceived as less likely to be repaid even if a good harvest followed a drought than were either loans from commercials banks or MFIs. Finally, respondents answered a hypothetical question that asked to provide a ranking of lenders they would pay first if they had an outstanding loan from different lenders but did not have enough money to meet all repayment obligations. As it turns out, PACS loans always appeared last in the repayment priority list, behind all other loan types. The first in the list was always a loan from an MFI. Thus, the frequent announcement of relief measures appears to undermine the culture of repayment. 5. Conclusion If an expanded role for government has to be the solution, care should be taken in ensuring the effects of government interventions have the intended consequences. There is strong evidence pointing to the crucial importance of government institutions regarding the functioning of the financial sector in general and access to credit in particular. However, instead of replacing private agents in the market, governments can make a more valuable contribution in securing an adequate environment for private credit institutions to flourish and to improve institutional mechanisms in ways that further the availability of credit to poor individuals. The range of policies that would contribute to this goal goes beyond the extent of this article: building these institutions involves decision-making with long time-horizons, and the guidance provided by existing research is only partial and incomplete. We have, however, some good indications on where the right path might lie. Providing good institutions, such as efficient, speedy and fair courts is an area where only the government can play an active role. Also, governments can provide much needed registries of credit information, liens and property ownership; and make sure legislation is reliable and property rights and responsibilities of companies and financial institutions are properly defined. Reforms in these areas may be difficult to accomplish in the short term, but there are many other dimensions where progress comes at lower costs: Djankov, McLiesh and Shleifer (2007) use a panel of aggregate data for 129 countries over 25 years to confirm that creditor rights and the existence of credit registries (public or private) exhibit a strong correlation with higher ratios of private credit to GDP. Interestingly, the information infrastructure (for instance, credit registries) matter relatively more (compared to creditor rights) in poorer countries than in rich countries. 8 The experience of India in the 1990 also gives insights into ways the government can work to ensure that access to credit is available to those who need it the most: Visaria (2006) shows that the new expedited mechanism introduced for loan contract enforcement resulted in a sizeable increase in loan recoveries, as market participants were able to bypass the inefficient court procedures then in place. Also, consistent with the evidence we presented in this paper, Caprio and Honohan (2004) argue about the importance of ensuring the independence of bank supervisors from the political sphere, and of the supervised entities themselves, in order for banks to promote social welfare and not their own or that of the officials. In sum, in this paper we have provided some evidence showing that government controlled PACS may suffer from weak governance and poor management practices and may be prone to be captured by politicians. This evidence is consistent with cross-country evidence that emphasizes governance and institutional issues as critical in explaining differences in the performance of public sector interventions. Some of the recommendations of the Malegam Committee may restore confidence in the microfinance sector but it is critical that the details are worked out and they are properly implemented. If not, we risk that the ordinance promulgated in AP will inflict lasting damage to the repayment prospects of all loans, worsening the credit prospects for everyone, especially the most vulnerable. References The Andhra Pradesh Micro Finance Institutions (Regulation of Money Lending) Rules 2010. G.O. MS. No. 356, Panchayat Raj & Rural Development (RD-I). Government of Andhra Pradesh. Extraordinary Rules Supplement to Part-VII of The Andhra Pradesh Gazette, October 19, 2010. Beck, Demirguc-Kunt and Martinez-Peria (2008). Bank Financing for SMEs around the world. Drivers, obstacles, Business Models, and Lending Practices. World Bank Policy Research Working Paper 4785 Caprio, Gerard, and Patrick Honohan. 2004. “Can the Unsophisticated Market Provide Discipline.� In Market Discipline across Countries and Industries, ed. William C. Hunter, George G. Kaufman, Claudio Borio, and Kostas Tsatsaronis, 349-62. Cambridge, MA: MIT Press) 9 Cole, Shawn. 2009. "Fixing Market Failures or Fixing Elections? Agricultural Credit in India." American Economic Journal: Applied Economics, 1(1): 219–50.Cole, S., Healy, A. and Werker, E. (2009). Do Voters Demand Responsive Governments? Evidence from Indian Disaster Relief. Daley-Harris, S. (2009) State of the Microcredit Summit Campaign Report. Dinç, Serdar. 2005. “Politicians and Banks: Political Influences on Government-Owned Banks in Emerging Countries�. Journal of Financial Economics 77 (2): 453-79 Johnson, Doug and Meka, Sushmita (2010). Access to Finance in Andhra Pradesh. Institute for Financial Management and Research. Centre for Micro Finance. Khwaja, Asim Ijaz and Atif Mian. 2005. “Do Lenders Favor Politically connected Firms? Rent Provision in an Emerging Financial Market.� Quarterly Journal of Economics 120 (4): 1371-411 La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert W. Vishny. 1997. “Law and Finance� Journal of Political Economy 106 (6): 1113-55. Murray, Jessica and Rosenberg, Richard (2006). Community-Managed Loan Funds: Which Ones Work? CGAP Focus notes. Rutherfurd, Stuart (2000). The Poor and their Money. Practical Action Publishing, UK Yezdi H Malegam (2011). Report of the Sub-Committee of the Central Board of Directors of Reserve Bank of India to Study Issues and Concerns in the MFI Sector. Reserve Bank of India. 10 Table 1: Access to Credit Households' preferred sources of credit 60 50 40 30 % 20 10 0 All Marginal Farmers Small Farmer Large Farmer PACS SHG Formal Institutions Informal Sources Percentage of households that have at least one loan with… 90 80 70 60 50 % 40 30 20 10 0 All Marginal Farmers Small Farmers Large Farmers PACS SHG Formal Institutions Informal Source Source: Household survey conducted in three districts in AP, India in 2007. 11 Table 2: Loan Characteristics PACS SHG Formal Sources Informal Sources Number of Loans since April 06 861 175 454 1,051 Median Borrowed Amount 10,000.0 7,000.0 20,000.0 20,000.0 Mean Borrowed Amount 15,734.3 8,725.1 44,565.0 26,522.4 Std Borrowed Amount 29,993.2 6,871.5 141,475.3 40,678.1 Maturity (Months) 24.32 14.80 15.55 17.11 Annual Interest Rate (%) 16.85 19.07 20.10 31.71 Pct. of loans that equire Physical Collateral (%) 98.14 4.00 91.41 8.09 Source: Household survey conducted in three districts in AP, India in 2007. Formal Sources includes Basix / Bartiya Samruddi / KBS Bank or NGO, commercial banks, rural bank and government bank. Informal Sources refers to friend/relative, local money lender, landlord/employer, local grocery store, pawn broker, SHG/ Raitu Mitra and others. Physical Collateral includes land, house, gold, crops or livestock (the rest of loans required guarantor, fund-savings or other collaterals). Table 3: Perceptions of Access to credit and repayment Comm. PACS - PACS - Comm. PACS MFI Comm.Ba Bank - Bank MFI nk MFI Perceived probability of … (0-10) Repayment rescheduled if drought 7.078 5.448 2.446 1.629*** 4.631*** 3.002*** Repayment forgiven if drought 6.564 4.621 2.015 1.943*** 4.549*** 2.606*** Loan amount lower than requested 6.441 5.109 3.604 1.332*** 2.838*** 1.506*** New loan after full repayment 4.961 6.276 6.818 -1.315*** -1.857*** -0.542** New loan even if not full repayment 3.696 3.650 4.253 0.046 -0.557** -0.603** Share deducted from loan as bribe 4.722 2.903 2.985 1.819*** 1.737*** -0.082 Share cap./savings misused and lost 4.801 2.374 3.576 2.426*** 1.225*** -1.201*** Repayment of loan in case of droughts but good harvest* 4.962 6.186 6.613 -1.224*** -1.651*** -0.427** Rank of repayment of loan if not enough funds to repay all (1 to 3) * 2.486 1.907 1.607 0.579*** 0.879*** 0.300*** Source: Household survey conducted in three districts in AP, India in 2007. An asterisk (*) indicates that the sample is restricted to those with more than one source of credit. 12