ABSTRACT There is evidence that the rural credit markets of developing countries exhibit assortive matching patterns that sort and link borrowing entrepreneurs and lenders according to classes. This dissertation provides an explanation for this empirical regularity. Assortive matching patterns result from cost complementarities between lending technologies and borrower abilities to convey information about their risk type to lenders. Lenders use either standard or idiosyncratic information to support screening, monitoring, and enforcement activities. The value of standard information is not affected by the distance between borrowers and lenders; thus, other lenders can read this information at comparable costs. In contrast, idiosyncratic information is generated at low marginal cost by interactions among local residents, or among agents in the same market. Consequently, the cost of gathering idiosyncratic information is affected by the identity and location of the lender. The data set comes from a survey of rural entrepreneurs and a series of case studies about non-bank and informal lending technologies in three rural areas of Mexico. Rural entrepreneurs exhibit low participation rates in credit markets and are rather monogamous debtors. A benchmark model for a pairwise analysis of entrepreneur and lender interactions is developed to demonstrate the effects on equilibrium interest rates and loan size of: (i) fixed and proportional costs of borrowing and lending; and (ii) the information acquisition behavior of lenders. The benchmark model is extended to derive conditions for assortive matching. It is shown that an assortive matching pattern results in which formal lenders specialize in serving rural entrepreneurs with high endowments of standard information. ii Assortive matching patterns are associated with market segmentation, with degrees of market power, and with interest rate differentials between formal and informal lenders. A multinomial logit model is fitted to test for the matching hypothesis that entrepreneurs with greater endowments of standard information are more likely to be matched with formal lenders who specialize in interpreting standard information. The dissertation found that rural entrepreneurs with formal accounting practices, with higher endowments of collateralizable real estate assets, and with a history of prior access to credit are more likely to obtain loans from formal lenders. iii Is too Risky Did not need Self-selected out Denied Cash-Loan Borrowers -II i i 1 I I i 0 8 16243240 %ltEs Figure 2 Participation and Reasons for not Participating in the Market for Cash Loans. 2.10 Credit Contracts: The Supply Side The main objective or this section is to show the characteristics of credit contracts in Mexico. These characteristics are the result of the interaction between loan applicants and lenders, and of lending technologies. Five elements of the credit contract are considered: loan amount, collateral and guarantees, maturity, interest rates, and borrower transaction costs. 2.10.1 Maturity. Most of the credit transactions are of short duration. There are differences in terms to maturity among contracts provided by formal and informal lenders (see Table 7). Bank loans have the longest repayment terms; the average is 21 months, almost twice the average repayment term that chartered non-bank intermediaries require of REs. Loans from the informal sector (i.e., moneylenders and friends and relatives) have much shorter repayment periods and quite flexible repayment schedules. Loans with a flexible repayment schedule are those whose maturity is contingent on the borrowers ability and willingness to pay, because a schedule is not establised at disbursement. About one-half of the loans by informal lenders have flexible repayment schedules (see Table 7). Flexibility is more common for loans from friends and relatives-59 percent of transactions. Flexibility does not mean long repayment 29 R Figure 5 Loan Demand and Fixed Fac,ors or Production The properties of the REs iso-expected profit curve are given by the following proposition. Proposition 2: Characteristics of the Iso-expected Profit Curves of Rural Entrepreneurs. (i) Lower isoprofit curves correspond to higher levels of profit. (ii) Iso-expected profit curves are concave around B d. The isoprofit curves are positively sloped for B Bd the converse is true. Figure 6 shows the relationship between an entrepreneurs loan demand and his isoprofit curves. 61