Preliminary Results from an Impact Evaluation of Urban Micro-Credit in Hyderabad
by Alexandra Kobishyn

While there is much anecdotal evidence on the impact of microfinance, there have been few rigorous assessments of the extent to which microfinance has improved the lives of beneficiaries. Many microfinance institutions conduct before and after evaluations
of clients, but such snapshots ignore external influences and self-selection biases that could distort the results. Given the scale of the sector, the international attention and the expectations of microfinance’s impact, the importance of a rigorous impact evaluation cannot be overstated. To avoid issues of non-comparability between microfinance clients and nonclients, the Centre for Micro Finance (CMF) in conjunction with Professors Esther Duflo and Abhijit Banerjee (MIT) and the Jameel Poverty Action Lab (J-PAL) conducted a randomised evaluation of the impact of Spandana’s micro-credit program in Hyderabad, Andhra Pradesh.

The evaluation aimed to tease out microcredit’s impact on income, consumption, financial services usage, asset ownership, business scale and profitability, and intra-household decision-making. Spandana is a Non Banking Finance Company (NBFC) that utilises the joint liability model for disbursing micro-credit and has operated out of Hyderabad since 2005. Ms. Padmaja Reddy, Spandana’s founder and executive director, was also instrumental in the conceptualisation of the study. One hundred and four slums in Hyderabad, with an average population of 300, were randomly divided into two groups of 52 slums. Spandana then introduced microcredit into one group of 52 slums, the “treatment group,” and for the duration of the study refrained from
introducing micro-credit into the other group of 52 slums, “the control group.” The research team surveyed both control and treatment groups prior to Spandana’s arrival in 2006, then conducted an endline survey of all the slums, covering about 7,000
households in total, in late 2008.

Thus, the research team was evaluating the impact of Spandana’s microcredit, a Rs. 10,000 ($250) loan, on slum households approximately 15-24 months after loans were first offered. While analysis is ongoing, there are a number of preliminary results available from this groundbreaking study. The data shows a number of distinctions when the treatment slums, those slums
where Spandana introduced microcredit, were compared to the control slums. For example, treatment slums (those opened before the introduction of microcredit) had roughly 30 percent more new businesses than control slums and existing businesses
in treatment slums reported higher profits than comparable businesses in control slums.

The preliminary results also show that formal microcredit impacts different types of borrowers differently, with a strong delineation being whether the borrower has or is likely to start a business. For example, households in treatment areas who already owned
businesses were likely to spend more on durable goods, reduce consumption on “temptation” goods such as alcohol or paan, and see higher business profits than similar households in control areas. Treatment-area households who did not own businesses but
exhibited characteristics highly correlated with starting a new business, such as literacy, tended to borrow and reduce consumption of temptation goods, relative to similar control-area households. However, treatment households who did not own businesses and did not show a propensity to start one tended to use the credit simply to consume more, including more on temptation goods, than similar control households.

Thus, one of the implications is that formal micro-credit might encourage households with an investment goal, such as opening or expanding a business, to shift their consumption patterns toward investment and away from temptation goods. However,
households who do not have an investment opportunity/goal may use microcredit to pay off higherinterest loans, or simply to borrow against the future. Incorporating such empirically-backed nuance to the design of microfinance products could have positive implications for the way practitioners engage with different types of clients.

Another claim often made about microfinance is that it can have dramatic impacts on non-financial outcomes such as women’s
empowerment, girls’ schooling, health or wellbeing. However, access to Spandana’s microcredit product did not cause changes in education and health expenditures or outcomes, nor were there any perceptible improvements in happiness or empowerment. It is worth reiterating the timing of the endline as only 15- 24 months after the introduction of microcredit; further study is needed to gauge whether improvements in social indicators manifest only after several loan cycles, when the income from businesses created or expanded with microloans may translate into higher household consumption.

While the early results of this study do not trumpet microfinance as a quick fix to poverty, they add focus to the oft-blurry approach to understanding the sector as well as show the promise of bringing formal financial services to the poor. Similar and further research in diverse contexts and countries will only sharpen our knowledge as well as our toolkit for eradicating poverty.

For more on the study and its results check out the following link:

The Miracle of Microfinance? Evidence from a Randomized Evaluation by Banerjee, Duflo et al
May 2009, Electronic version