Heterogeneous effects of microcredit: Evidence from large-scale programs in Bangladesh; Research by Asad Islam

Microcredit has become one of the central elements of pro-poor policy over the past three decades, most notably through its widespread adoption in Bangladesh. However, the effectiveness of microcredit in raising household welfare — and the distribution of those effects — remains a subject of academic debate. This study rigorously evaluates the causal impact of microcredit program participation on household food consumption using nationally representative data and advance econometric methodologies. The primary objective is to identify which groups within the poor population benefit most, with particular attention to gender and poverty status.

Methodology The analysis leverages a unique, large-scale survey covering 3,026 households from 91 villages, drawn across both program and control groups. The dataset — commissioned by PKSF and implemented by BIDS — enables granular assessment by subgroup. To address issues of self-selection and non-random program placement, the authors use village fixed effects and an instrumental variable approach based on land-eligibility criteria, thereby ensuring robust causal inference regarding microcredit’s impact on consumption. The methodology also distinguishes between intention-to-treat (ITT) and treatment-on-the-treated effects.

Results The study’s central finding is that microcredit’s effects on food consumption are highly heterogeneous across household types:

• Poorest Households: The most substantial benefits are observed among the “poorest of the poor,” who register statistically and substantively significant increases in food expenditures after program participation.
• Marginal Participants: Households on the margins of program eligibility — those not in abject poverty, but still poor — exhibit limited or negligible welfare gains from microcredit.
• Gender Disparities: The effects are robustly stronger for female borrowers compared with male participants, reinforcing the gender-sensitive nature of successful microcredit schemes.
• Robustness: These outcomes remain consistent across multiple identification strategies, and persist when adjusting for possible spillover, selection, and non-compliance.

Contextualization with Literature The paper corroborates earlier evidence (e.g., Pitt & Khandker, 1998; Khandker, 2005) regarding microcredit’s positive impact on consumption and female empowerment. However, it adds necessary nuance by illuminating significant distributional effects: not all poor benefit equally, and program targeting is critical. Earlier experimental and quasi-experimental work indicated similar outcomes in terms of business expansion and risk mitigation; this study gives new depth to the consumption channel, especially regarding structural poverty.

Policy Implications

• Targeting: Program expansion should prioritize the poorest households and female borrowers to maximize poverty alleviation and welfare impact.
• Design: Eligibility criteria and community engagement mechanisms can be optimized based on empirical evidence to enhance outreach and effectiveness.
• Assessment: Continuous data collection and advanced analytic strategies are essential for dynamic program evaluation.

Access the complete study:- https://www.sciencedirect.com/science/article/abs/pii/S1049007815000044