Matching for Microcredit

Market design approaches to financial inclusion

In many of the poorest countries, informal micro-enterprises account for more than 2/3 of economic activities. However, micro-entrepreneurs often find it difficult to grow their businesses because they lack the necessary collateral to access credit and severe information asymmetries between banks and borrowers exist. In response to this market restriction, a simple market design idea has allowed local bankers to revive dormant credit markets and also won one of them the Nobel Peace Prize. Today, the German Kreditanstalt für Wiederaufbau (KfW) is the world’s largest lender to these credit markets. The idea: give loans to groups of borrowers that act as mutual bails, providing incentives to select only creditworthy peers. This idea, however, has little bite in rural settings, where rice farmers, say, group with other rice farmers – all exposed to very similar business risks. During a severe drought, for example, all farmers default simultaneously and the bank loses the security of the bail-out.

The market design process

Market design considers market rules to relax said market restrictions. In the microcredit example, a simple market rule is to limit the grouping of borrowers with similar economic activities – e.g. allow for a maximum of three rice farmers per group. This approach is novel in that it is not concerned with finding the optimal credit contract but instead aims to identify rules for group formation to improve repayment – taking interest and joint liability rates as given. The market design process starts by devising such rules, combining economic theory with a sound understanding of the market context. It proceeds by testing the rules using game theoretic and/or econometric techniques and supporting the implementation, ideally with a built-in, quasi-experimental evaluation. For the analysis of non-experimental survey data, group formation creates an endogeneity problem. The empirical contribution of this project is the development of a structural model – implemented in R package matchingMarkets – that generalises the well-established Heckman correction to allow for the equilibrium analysis of group formation games. The identification assumption is that the characteristics of all agents in a market affect who forms a group with whom, but the repayment of any group is determined only by its own members. This exogenous variation is similar to an instrumental variable. The bottom line for the proposed market rule is that it demonstrably leads to less group defaults, which enables the banks to reduce interest rates, allowing more micro-entrepreneurs to access credit at affordable rates and grow their businesses.

Lessons learned

The main lesson learned from this project is that the composition of groups (or “matches”) does matter and the market design process can help devise better market rules in such matching markets. The methodology is applicable to other matching markets. Examples include the design of quotas for (ethnic) minorities in public school choice and entry labour markets, the state regulation of company mergers and the design of matching algorithms to facilitate decentralised land reform.

Read the full paper in the Cambridge Working Papers in Economics Series.


comments powered by Disqus