Principal investigator: Omar Barroso Khodr
Authors: Vivian Hoffmann, Vijayendra Rao, Vaishnavi Surendra, Upamanyu Datta
Original title: Relief from usury: Impact of a self-help group lending program in rural India
Location of the Intervention: Credit Policies and Economic Development.
Sample Size: 180 panchayats (local administrative units)
Primary Variable of Interest: The result of the interest for families in the village in the panchayat in the year 't'.
Type of Intervention: Credit subsidy policies, poverty reduction
Methodology: ANCOVA specification; random walk; Romano-Wolf stepwise method; and use of randomization inference.
Summary
The provision of low-cost microcredit to low-income populations, generally through self-help groups (SHGs), has been widely adopted as a key strategy for poverty reduction in developing countries. However, evidence on the impact of this approach is scarce. Using a randomized implementation program in over 180 panchayats (Village councils), the authors assess the impact of a government-led GAAs initiative in the Indian state of Bihar. Two years after the program began, the authors observed a significant increase in GAAs membership and loans granted by these groups, resulting in a corresponding decrease in the use of informal credit. Furthermore, the authors note that there are fewer informal lenders operating in the beneficiary villages, and those who remain charge lower interest rates. While these impacts on the credit market may lead to substantial improvements in long-term economic well-being, the program's impact on these outcomes is modest in the short term.
- Policy Problem
The authors discuss critical policy implications related to large-scale Self-Help Group (SHG) programs conducted by the Indian government. The central issue is that, although billions of dollars are invested in these programs, there is a notable lack of rigorous evidence on their real impact, especially in the informal credit market. This evidence gap undermines large poverty reduction initiatives, in which policymakers allocate vast resources based primarily on observational data that can be misleading. The article directly addresses this problem by presenting a randomized evaluation—which they consider the gold standard in impact analysis—to offer a more reliable measure of program effects and answer the fundamental question.Are these substantial investments being well applied?
A central policy dimension explored is the impact of GAA programs on informal credit markets. Prior to this study, the effect of large-scale initiatives on the informal sector, on which a large portion of the rural poor population depends, was poorly understood. The authors demonstrate that such programs not only introduce a new source of credit but also actively destabilize the informal market. The main findings reveal a substitution effect, in which households exchange high-cost informal debt for low-cost loans offered by GAA, as well as competitive pressure that forces informal lenders to reduce their interest rates, leading some to exit the market. According to the results, this results in significant welfare gains, especially for marginalized groups such as Registered Castes and Tribes, who experienced a 20% reduction in the cost of their loans.
In this context, the article also critically examines the broader socioeconomic impact. GAA programs are often justified by promises of poverty reduction, increased productive investment, and women's empowerment. However, the results raise concerns for policymakers: the effects were modest and limited, such as a slight increase in consumer assets, without growth in productive assets or consumption itself. The authors explain that, consequently, there was no perceptible impact on women's empowerment. According to them, this factor requires a reconsideration of policies: if the main proven benefit is the reduction of the cost of debt, and not the promotion of transformative change, the program's objectives and public communication need to be recalibrated.
Furthermore, the article presents robust evidence on targeting and equity. Given that informal credit markets are highly stratified and disadvantage the poorest, the study demonstrates that well-targeted GAA programs can successfully reach marginalized groups and offer them disproportionate financial relief. This confirms that such initiatives can be an effective tool in promoting financial equity.
In summary, the implications for policymakers are clear. The article argues that investing in GAA programs is an effective strategy for dismantling exploitative informal lenders, representing its most relevant effect. However, caution is advised regarding expectations of transformative changes in empowerment or income, and the need to base future financing decisions on rigorous evidence is emphasized. The main recommendation is to focus on the primary and proven benefit: offering low-cost credit capable of stimulating competition and reducing the financial burden on the poor—which, in itself, already constitutes a significant achievement.
- Policy Implementation Context
The authors emphasize that public policies granting credit through GAAs (Group Asset Access) have been implemented in a context where their overall impact, especially on the informal credit market, remains theoretically ambiguous and empirically uncertain. According to the authors, the available evidence on similar large-scale government-led GAA programs was based on observational methods, such as propensity score matching. These studies presented heterogeneous results regarding economic outcomes, such as asset ownership, but frequently reported positive effects on women's autonomy. However, the reliance on non-experimental data kept open the possibility of bias, hindering the formulation of definitive causal conclusions about the real effects of the policy.
Thus, a crucial aspect of the implementation context is the policy's intended interaction with the informal credit sector, in which a significant portion of the rural poor—especially in regions like Bihar—still resorts to loans with high interest rates. Theoretically, the introduction of a low-cost competitor, such as GAA programs, could destabilize this market through competitive pressure, reducing rates for everyone. However, economic theory also suggests alternative scenarios: the new credit could induce moral hazard among borrowers or provoke market segmentation, potentially raising interest rates for those not served by GAA (Hoff, Karla and Stiglitz, 1997; Kahn and Mookherjee, 1998; McIntosh and Wydick, 2005). This theoretical ambiguity highlights a relevant gap in the policy's evidence base, since understanding this interaction is essential to assessing its overall implications for welfare.
Existing empirical studies on this specific interaction have also yielded divergent results: some have not identified a substitution of informal lenders, while others have pointed to more complex relationships. A fundamental limitation of these studies—including recent Randomized Controlled Trials (RCTs) on microcredit—has been the difficulty in definitively measuring the impact on informal markets. This is largely due to low program participation rates or contexts where informal credit does not constitute the primary source of financing. Thus, the policy was implemented and expanded with billions of dollars in investment without robust causal evidence regarding one of its main mechanisms of change: the ability to overcome and discipline high-cost informal lenders who perpetuate poverty. Finally, this study seeks to fill this critical gap by evaluating a program with high participation in a region heavily dominated by informal credit, and providing the first consistent causal test of the impact of this policy on this market.
- Evaluation Details
The assessment was conducted in Bihar, an Indian state with an dire need for effective poverty reduction strategies, where 32 million people lived below the poverty line. A striking feature of this context was the severe lack of access to formal financial services among the poor population. At the time, credit through GAAs or Microfinance Institutions (MFIs) was quite limited, representing only 3,2% of outstanding loans. In contrast, credit offered by loan sharks at high interest rates was widespread, creating a clear market gap and an urgent demand for affordable alternatives. This situation directly justified the implementation of the Jeevika project by the government in partnership with the World Bank.
According to the authors, the evaluated intervention corresponds to a large-scale, government-led, comprehensive banking and asset management (GAA) program. Its main components include the mobilization of women in small groups that meet weekly, later organized into larger structures at the village and urban center levels, to facilitate access to formal banking. The program's activities go beyond credit granting, incorporating a training curriculum focused on women's empowerment, advocacy, and the development of basic reading, writing, and math skills. Although its target audience is poor and marginalized communities, the program is widely accessible to any adult woman residing in the areas of operation.
According to the authors, from a financial standpoint, the central axis of the intervention is access to subsidized credit. After three months of small, regular savings, the groups become eligible for a significant amount of capital earmarked for loans. These loans are offered at a monthly interest rate of 2%, strategically set at less than half the rate practiced in the informal market, making it a highly attractive alternative. The evaluation focuses specifically on this credit mechanism, since other planned components—such as income-generating training—had not yet been implemented in the region studied during the analysis period. This allows for a clear and isolated test of the impact of offering low-cost credit, with collective responsibility, in a scenario dominated by loan sharks who charge high rates.
- Method
According to the authors, the project evaluation Jeevika A rigorous experimental design was employed. The intervention was implemented in a randomized manner in 180 panchayats (local administrative units), which constituted the primary unit of analysis. To increase the precision of the experiment, the units were initially paired with their nearest neighbors within the same block, based on similar baseline levels of high-cost debt. Then, one panchayat from each pair was randomly assigned to the early implementation group (treatment), while the other was assigned to late implementation (control), establishing a clear counterfactual for comparison.
The authors reveal that the household sampling strategy was carefully planned to reach the program's target beneficiaries. The study aimed to prioritize socioeconomically disadvantaged groups without sacrificing variability; the sample was stratified to include 70% households from registered castes and tribes (SC/ST) and 30% from other castes. In segregated villages (tolas), households were selected using a random walk method with a skip pattern, ensuring randomness within each stratum. For the analysis, inverse probability sampling weights were constructed so that the results reflected the actual caste composition in each village, assigning equal weight to all castes.
Therefore, data collection was comprehensive, conducted at both the household and village levels. Baseline and follow-up surveys gathered detailed information from families on debt, assets, consumption, livelihoods, and women's empowerment. Additionally, village-level surveys were conducted with community members familiar with the local context to reach consensus on key attributes such as demographics, sources of credit, interest rates, and wages.
Empirically, the primary method used to estimate the program's impact was an ANCOVA specification, which controls for baseline values of outcome variables, increasing statistical power. This procedure provides an intention-to-treat (ITT) estimate of the intervention's effect.
Finally, the study also pre-specified a heterogeneity analysis of treatment effects, testing whether the impact differed for registered caste and tribe families (SC/ST), the central target group, using interaction terms in the regression. Statistical robustness was ensured by different strategies: pooled standard errors at the panchayat level; creation of indices for families of related variables in order to handle multiple hypothesis tests, with p-values adjusted by the Romano-Wolf stepwise method; and use of randomization inference to further validate the significance of treatment effects.
- Main results
The study results reveal significant impacts of the Jeevika program, particularly in reshaping household credit patterns and influencing local informal credit markets. One of the main direct effects was a marked increase in participation in Self-Help Groups (SHGs) and in obtaining loans through them. According to the authors, in the intervention areas, 55% of households had at least one member in a SHG and 30% had obtained loans through these groups, rates much higher than in the control areas. This participation was especially pronounced among registered caste/tribe (SC/ST) households, which were the focus of the program.
Consequently, there was a reduction in the use of informal lenders by families in the program areas, with a 5 to 6 percentage point drop in obtaining new informal loans. This substitution led to a robust and significant decrease in the cost of credit for families: the average monthly interest rate on recent loans fell by 0,7 percentage points (a 13% reduction). The authors emphasize that this effect was stronger for SC/ST families, who faced higher rates at the beginning of the study.
However, despite this shift in credit sources, the program did not significantly increase the total amount of household debt and may even have reduced it, suggesting that Jeevika provided cheaper credit rather than simply increasing indebtedness. In market terms, the program had measurable indirect effects on the informal credit sector.
The analysis, which corrected for potential alterations in the borrower profile, found that Jeevika led to a modest, yet statistically significant, reduction in informal interest rates. This effect was concentrated among SC/ST households, which showed a 4,7% decrease. Corroborating this finding, village-level data—less susceptible to compositional bias—showed a similar decline.
Furthermore, the program contributed to a contraction in the informal lending sector itself, with the number of active informal lenders in the villages decreasing by approximately 10% in the intervention areas. The combined impact of this set of market-level outcomes was highly significant. In this way, the study also established important contextual conclusions. At baseline, SC/ST households demonstrated greater financial vulnerability, with a higher prevalence of debt, smaller loans, and significantly higher interest rates compared to other households.
Finally, although the randomized design was generally successful, the authors observed some minor imbalances between the treatment and control groups at baseline. However, they employed robust statistical methods, including ANCOVA, to control for these pre-existing differences and confirmed that the main outcomes regarding access to credit and interest rates remained strong and significant even after adjusting for multiple hypothesis testing.
- Lessons in Public Policy
The Jeevika program demonstrates several important public policy lessons related to microcredit intervention and its impact on local markets in informal economies. GAAs-based microcredit policy is an effective tool for reshaping household credit patterns, offering a robust alternative to informal lenders. The key lesson is that government support for GAAs can achieve a high level of uptake and use of institutional credit, especially among the most vulnerable groups, such as Registered Caste and Tribe (SC/ST) households. By increasing access to formal loans, public policy can promote a direct substitution of informal credit, as evidenced by the 5-6 percentage point drop in new loans from loan sharks.
A second lesson is that such credit programs can lead to a significant and robust reduction in the cost of credit for households. The study shows that the introduction of Jeevika forced a 13% drop in the average monthly interest rate for recent loans. This cost-reduction effect was more pronounced for SC/ST households, who originally paid the highest rates, demonstrating that public policy not only improves access but also acts as a mechanism for financial protection and equity for the most vulnerable groups. It is important to note that this policy managed to provide cheaper credit without increasing the total indebtedness of households.
Finally, the study highlights the power of credit interventions in indirectly influencing local informal markets. By providing a viable alternative, the program not only reduced the interest rate charged by remaining informal lenders (a statistically significant effect) — concentrated in SC/ST households — but also caused a physical contraction in the informal sector, with a 10% decrease in the number of active lenders in the villages. This suggests that microcredit policy has a market leverage effect, improving credit conditions for all and diminishing the power of loan sharks, which is a highly significant public policy outcome.
References
Hoff, Karla and Joseph E. Stiglitz. 1997. “Moneylenders and bankers: price-increasing
subsidies in a monopolistically competitive market.” Journal of Development Economics, 52(2),
pp. 429-462.
Kahn, Charles M. and Dilip Mookherjee. 1998. “Competition and incentives with
nonexclusive contracts.” The RAND Journal of Economics, pp.443-465.
McIntosh, Craig, & Bruce Wydick. 2005. “Competition and Microfinance.” Journal of
Development Economics, 78: 271-298
