The Benefits and Costs of Microfinance: Evidence from Bangladesh

Publication Details

Journal of Development Studies, April 2008, vol.44, iss.4, pp.463-484. Available From:

Link to Source
Matthieu Chemin
Institutional affiliations
None specified
Grant-holding institution
None specified
South Asia
Equity Focus
Evaluation design
Propensity Score Matching (PSM)
Journal Article


This study examines the socioeconomic impact of several microfinance programmes implemented in rural Bangladesh. Advocates of microfinance services—notably microcredit—argue that providing low-income households with better access to financial markets represents a powerful way to foster entrepreneurial initiatives and generate new economic activities. The present work investigates this claim by studying three different institutions (the Grameen Bank, the Bangladesh Rural Development Board, and BRAC) created in the 1970s to provide affordable alternatives to rural households excluded from the traditional banking sector. The institutions offer microloans to small groups of entrepreneurs who obtain their individual credits by providing mutual and morally binding guarantees—a commitment to pay back each other’s loans—rather than conventional collaterals.

The author uses a propensity score matching procedure to evaluate programme impact. Propensity scores are estimated using a model that includes variables such as the individual's sex, age, education, savings, and household size. Beneficiary individuals are then matched with control individuals on the basis of their scores—using both stratification and kernel matching techniques. The analysis includes data collected by the Bangladesh Institute for Development Studies for the years 1991 and 1992. The sample includes 5,043 observations from 72 villages covered by one or several institutions (treatment observations being composed of 705 beneficiaries and 4,338 non-beneficiaries) and 1,094 observations from 15 control villages covered by none of these three institutions. The surveys gathered information on agricultural income, sales of dairy products, revenue from non-farming enterprises, household expenditure, household land, transport assets, school enrollment, and household demographics.

Main findings

The analysis finds contradictory results depending on which control group is used (that is, nonparticipants from control villages or nonparticipants from treatment villages). When using nonparticipants from control villages, the author finds that participation in a microfinance programme led to a positive change in per-capita expenditure (coefficient of 0.028). In contrast, using nonparticipants from treatment villages as the control group reveals that per-capita expenditure actually decreased as a result of programme participation, with coefficients ranging from –0.046 (kernel matching) to –0.035 (stratification matching). The researcher suggests that this negative finding might be due to positive externalities entailed by the presence of the programme—nonparticipants benefiting indirectly from newly created economic activities.

Comparing participants with nonparticipants from control villages, the author also finds that programme participation significantly improved school enrolment among boys (coefficient of 0.035) and girls (coefficient of 0.051) and increased the amount of hours worked by men (coefficient of 17). No significant effects were found on women's nonland assets or the amount of hours worked by women. The researcher hypothesises that men's better response to the programme might be explained by the fact that men tend to borrow higher amounts than women. For instance, the average loan among female clients of the Grameen Bank approximated 11.5 Takas whereas the average loan size for men reached almost 14 Takas.

The researcher compares his results to those obtained in other works assessing the impact of microcredit. Notably, he points out that the estimates measured for per-capita expenditure (using nonparticipants from control villages as the control group) were lower than those found by Pitt and Khandker (1998) but greater than those calculated by Morduch (1999). The study concludes by calling for further research to study the positive externalities of microfinance within beneficiary villages.

Scroll to Top