Research Methodology and Data

The nature of the Khushhali Bank’s operations lent itself to an impact assessment using prospective clients who have not yet accessed loans as a comparison, or control group. The bank is rapidly expanding into new villages and the number of active clients is increasing at a rate of approximately 20,000 clients every 3 months. Bank management and staff were willing to cooperate with surveyors in identifying new villages that had just received the service and within those villages identifying new clients, allowing them to be surveyed in the interim between their application and the approval to get a microloan and the actual disbursement of the money.

Using the approach of surveying prospective clients who have not yet accessed loans as a control group, impact can be estimated with a single equation:

Y_ij = β₁X_ij + β₂V_j + β₃M_ij + β₄P_ij + β₅T_ij + β₉P_ijT_ij + &epsilon_ij

where Y_ij, is a vector of outcome variables (see Appendix 9.1 [ PDF 72.9KB | 1 page ] for a detailed list of variables and summary statistics for each) X_ij is a vector of household characteristics (see Appendix 9.2 [ PDF 72.6KB | 1 page ]), V_j represents village fixed effects, which control for observable and unobservable variables that may influence program placement, M_ij is a membership dummy variable equal to 1 for any household that participates in the program and T_ij is a measure of treatment: participation in the microfinance program.

The treatment variable is based on three alternative measures of participation in the program:
- ‘Months Since First Borrowed’: the number of months elapsed since the household first borrowed
- ‘Total Amount of Loans’: the total amount ever borrowed by the household
- ‘Number of Loans’: a count of the number of loan cycles the household has borrowed.
The first two measures of treatment, which only measure the impact of access to microfinance, present the most unbiased results.

The hypothesis tested is whether participation in the microfinance program of the Khushhali Bank has a positive effect on various outcome measures. Support for the hypothesis requires that the estimated coefficient β₃ on one of the treatment variables in (1) is statistically significantly positive. A statistically significantly positive coefficient estimate on one of the treatment variables indicates that the degree of participation in the program – either the length of time the client has participated, or how many loans he or she has taken out or the total value of those loans – has an impact.

In addition to the overall impacts of participation in the microfinance program, we examine whether there are any special impact for poorer borrowers. Defining P_ij=1 if a household is in the bottom quintile of the population in terms of monthly per capita food consumption, we first control for the fact that these borrowers are likely to have lower overall measures of welfare by including the dummy in all regressions, and then look for differential impact by interacting that dummy variable with the treatment variables to see whether participation in the program has more impact for those borrowers.¹²

The hypothesis tested is whether participation in the microfinance program for very poor borrowers has a more positive effect on various outcome measures than it does for average borrowers. Support for the hypothesis requires that the estimated coefficient â6 in equation (1), the interaction of the treatment variables with a dummy variable indicating extremely poor borrowers, is statistically significantly positive. A finding of no special impact for these extremely poor borrowers does not mean that the program has no impact on their welfare, but rather that their impact does not differ from the impacts of the program overall.

Estimation of equation (1) above was carried out using primary data from 2,881 rural and urban households in Pakistan. A stratified random sample of 1,454 Khushhali Bank clients and future clients was drawn from 139 rural villages and 3 urban cities where Khushhali operates. A roughly equal number (1,427) of randomly selected non-clients from the same villages or settlements were also surveyed (see appendix 9.1 for details of the survey).

The Khushhali Bank’s mandate is to serve the poor, defined as persons who have a meager means of subsistence and whose total income during a year is less than the minimum taxable limit. Accordingly, Khushhali serves clients who are ‘poor’ and ‘very poor’ but not those who are ‘destitute’ (receiving zakat as discussed in chapter 8) or the ‘non-poor’, who receive enough income to pay income tax. Clients are screened by bank staff and classified into one of the above categories when they apply for the loan. The program also has an element of self-targeting in that participation in the program is voluntary and the loan product – uncollateralized micro-loans of between Rs 3,000- 30,000 – are designed to be attractive to poor clients. These are loans of approximately $50 - $500. Indeed, in the sample drawn for this study, more than 70% of the clients were below the official poverty line of the Government of Pakistan¹³. 20% of the sample, defined here as the ‘core poor’ or ‘poorest of the poor’, were subsisting on less than half of the caloric consumption defined by the Government of Pakistan as poor. Rough calculations of total consumption-expenditure indicate that the 70% of the sample defined as poor are living on approximately 87 cents per day and the bottom 20% of ‘core poor’ are living on less than 50 cents per day (at current exchange rates).

For most of the empirical analysis, ordinary least squares analysis (OLS) was applied in estimating equation (1). For regressions in which the outcome variable of interest was a yes/no dummy variable on qualitative information, logit estimation techniques were used.

Comment(s)

There are [1] comment(s) for this entry. Post a comment.

1. Khan Hidayat Ullah
   (posted 23 November 2006 / 10:14:28 PM)
   
   Econometric Model Selection:
   
   I went through your paper and really liked the approach adopted by you while found your results interesting as well.
   
   I have one comment to make. When it comes the components of microfinance pertaining to the term of loan, e.g., amount, duration and repayment mode, in practice these factors are determined by the both the borrower and lender through negotiation between line staff of MFI and Managers of CBOs. This results into adoption of resolution by members in joint meeting held by field staff MIF and CBO.
   
   Keeping these factors in view, the claim that an explanatory variable relating to terms of micro-credit is exogenous becomes doubtful. This creates endogenity problems in the model and the results which are estimated by this model can not be claimed as unbiased.
   
   In you model you have used variables like amounts of loan and times of loans borrowed in past, which brings in endogenity problem in the model and leads to biased estimates, if I am correct? I would like to know how you addressed this problem in your econometric model. If not what kind of instrumental variable you intend to use in order to deal with situation will.