Fungibility and the Design and Evaluation of Agricultural Credit Projects

Abstract

Agricultural credit is an important element in development efforts in most low income countries. Some countries such as India, Brazil, and Thailand assign credit a leading role in rural development. The World Bank, the Inter-American Development Bank, and the Agency for International Development have aggressively promoted agricultural credit, committing in excess of $5 billion through hundreds of projects. The popularity of credit is due in part to the notions that loans are necessary to accelerate technological change in farming and that formal credit is required to release peasants from dependence on moneylenders. In certain situations the relative ease with which credit projects can be initiated adds to their appeal. Most credit projects are aimed at stimulating the production of commodities such as rice or dairy products, augmenting the use of an input like fertilizer or improved breeding stock, encouraging investment in machinery and irrigation, or providing more financial services to target groups such as the rural poor, cooperative members, or corn producers. Agricultural banks, cooperative banks, credit unions, and supervised credit agencies have been created under some of these projects. Other projects have augmented loanable funds flowing through existing parts of rural financial markets (RFMs). A number of these projects have been evaluated formally.' Major measures of performance emphasized by donor agencies are disbursement of project funds and recovery rates on loans to farmers. Most evaluations also attempt to measure the impact of loans on farm activities. Impact is usually expressed in terms of increases in crop area or yields financed by the project and by the quantity of animals, fertilizer, or tractors bought with loans. Numbers, amounts, and kinds of loans made, and farm income and net worth are also used as performance measures. These evaluations typically include little analysis in depth of the credit institutions handling project funds. While project evaluations may show slow loan disbursement or loan repayment problems, they often indicate that production, input use, investment, and target group participation goals were generally met, and that projects achieve many of their objectives. Despite this, a number of observers are increasingly concerned about the quality and quantity of services provided in low income countries by rural credit institutions and by the RFMs of which they are a part. Critics charge that although donor funding for agricultural credit has increased substantially, the real value of total agricultural loans has decreased in many countries, that concessionary loans often end up in the hands of the well-to-do, that loans for agricultural purposes are diverted to nonagricultural uses, that policies in many RFMs encourage consumption and discourage savings, that the term structure of agricultural loans often contracts or fails to expand, and that RFMs are adopting few cost-decreasing technologies in the provision of financial services. It is disconcerting that rural financial markets could perform poorly while projects within these markets are judged to be doing well. An attempt is made in the following discussion to resolve this paradox by showing how design and evaluation procedures which ignore fungibility lead to faulty conclusions about agricultural credit project results.2