Micro-Lending's Broken Promise And What Actually Builds Household Wealth

The rise and partial fall of microcredit is one of the most instructive episodes in the history of development economics — not because it involved fraud or bad faith, but because it illustrates how a plausible mechanism story can generate enormous institutional momentum before the evidence is in, and how the evidence, when it arrives, struggles to overcome the institutional investments already made.

The Theoretical Case and Why It Was Compelling

The theoretical case for microcredit rested on a simple model: poor people have entrepreneurial ideas but lack capital; capital markets fail poor people because they lack collateral; if you solve the capital access problem through unsecured small loans, entrepreneurship will flourish, income will rise, and poverty will decline. Grameen Bank's group lending model added an elegant mechanism: borrowers in solidarity groups were jointly responsible for each other's repayment, which substituted social collateral for physical collateral and aligned incentives without requiring traditional loan security.

This model was theoretically coherent. The Grameen Bank's early repayment rates were impressively high — 97 percent or above was frequently cited. The case studies of Bangladeshi women using small loans to buy sewing machines or mobile phones for community airtime rental, and transitioning from poverty to modest stability, were real. The theory seemed to match the evidence.

What the early evidence did not show — because the studies were not designed to detect it — was whether microloan recipients were better off because of the loans or simply because they were selected into a program that provided social support, training, regular savings discipline, and peer network effects along with the credit. Grameen's group lending model bundled many interventions simultaneously. Attributing the observed benefits to the credit specifically was an error of interpretation rather than evidence.

The Randomized Trial Evidence

Beginning in the mid-2000s, development economists began designing randomized controlled trials to isolate the effect of credit access from confounding factors. The results, published across multiple countries and research teams, converged on a finding that Dean Karlan of Yale University, one of the leading researchers in this area, has summarized as: microcredit has modest but positive effects at the margin for existing entrepreneurs, and near-zero effects on average household consumption or poverty escape rates for typical program participants.

The Hyderabad study by Banerjee, Duflo, Glennerster, and Kinnan (2015), considered the most rigorous evaluation of a standard microcredit program, found that after 15–18 months, households with microcredit access showed no significant improvement in average consumption, women's empowerment measures, or education outcomes compared to control households. Existing business owners showed modest improvements in business profitability. But the typical microcredit borrower — not a pre-existing entrepreneur but a household managing poverty through diverse coping strategies — did not benefit measurably from credit access.

Similar null or near-null findings on average household welfare effects were produced by studies in Morocco (Crépon et al., 2015), Mongolia (Attanasio et al., 2015), Bosnia and Herzegovina (Augsburg et al., 2015), Mexico (Angelucci et al., 2015), and Ethiopia (Tarozzi et al., 2015). The convergence of negative results across such diverse geographic and institutional contexts was unusual in development economics, where context-dependence typically limits generalization. It suggested that the null result was robust.

The Structural Barriers That Credit Cannot Solve

The microcredit story assumed that lack of capital was the binding constraint preventing poor entrepreneurs from growing their enterprises. The research found that for most people in poverty, this assumption was wrong. The binding constraints are typically several layers deeper.

Market size and purchasing power constraints: In an economy where the majority of potential customers are also poor, there is a fundamental ceiling on the scale achievable by a small service or goods business. If you borrow money to expand your tailoring business in a low-income neighborhood, you will still be constrained by the number of potential customers who can afford tailored clothes. Adding more sewing machines does not create more customers. This is not a problem that credit can solve.

Regulatory and infrastructure constraints: The World Bank's "Doing Business" indicators, whatever their methodological limitations, consistently document that in the countries with the highest rates of extreme poverty, the regulatory burden on small formal businesses is severe — multiple registrations, licenses, and inspections that impose significant compliance costs. Combined with inadequate power supply, poor roads, and limited telecommunications, these infrastructure and regulatory constraints impose real costs on small business operations that credit cannot reduce.

Skills and knowledge constraints: Most extremely poor households lack the business management skills, market knowledge, and production expertise required to successfully operate and grow a microenterprise. Credit provides capital. It does not provide competence. The Grameen model was aware of this and built training and peer learning into its program design. Later, lower-cost microfinance institutions that stripped out these support elements to reduce costs further degraded outcomes.

Competition from imports: Many of the goods that microentrepreneurs might produce — textiles, processed foods, consumer goods — face direct competition from subsidized imports from wealthy country manufacturers. A Bangladeshi tailor with a microloan cannot compete on price with a Chinese factory. A Kenyan processed food producer cannot compete on price with European agricultural surpluses dumped below cost in African markets. These competitive dynamics are structural and political, and they constrain small enterprise viability in ways that more credit cannot address.

The Debt Trap and the Andhra Pradesh Crisis

While the average effects of microcredit are near zero, the tail risks are significantly negative. In contexts where aggressive microfinance lending outpaced borrowers' capacity to repay, cycles of multiple indebtedness emerged — borrowers taking new loans to repay old ones, entering debt spirals that eliminated household financial stability entirely.

The 2010 Andhra Pradesh microfinance crisis in India is the most documented case. Multiple microfinance institutions competed aggressively for borrowers in the same villages, resulting in many households carrying loans from four or more different lenders simultaneously. When repayment collection practices became coercive, a wave of borrower suicides — estimated at over 50 in a two-month period — triggered a state government moratorium on collections and a near-collapse of the microfinance sector in the state. Over $4 billion in loans became uncollectable. The crisis was not an anomaly — similar episodes occurred in Bolivia, Nicaragua, and Morocco — but it was the most severe.

The Andhra Pradesh crisis illustrated that the institutional incentives within the microfinance sector had diverged significantly from borrower welfare. Microfinance institutions competed on loan volume rather than borrower outcomes. Investors, including international impact investment funds, rewarded growth metrics rather than repayment quality or borrower welfare indicators. The result was a system optimized for loan disbursement rather than for the household wealth outcomes that the sector claimed to produce.

What Actually Builds Household Wealth

The positive evidence base for household wealth building in low-income contexts is robust and largely points toward different mechanisms than credit:

Direct cash transfers: The GiveDirectly program, which transfers unconditional cash to extremely poor households in Kenya and Uganda, has been more rigorously evaluated than almost any other development intervention. Studies including a twelve-year randomized controlled trial found large, lasting effects on household assets, consumption, and psychological wellbeing. Cash transfers outperform in-kind aid on virtually every household welfare metric and cost substantially less per dollar of benefit delivered. The success of cash transfers challenges the paternalistic assumption that poor people need guidance on how to spend resources — the evidence shows they use cash effectively and according to their actual priorities.

Universal basic services: The most dramatic poverty reductions in the historical record were produced by state investment in universal services: public education, universal healthcare, social insurance against illness and old age. South Korea, Taiwan, China, and the East Asian Tigers all achieved rapid poverty reduction through state-led industrialization combined with universal investment in human capital. The microfinance model emerged partly as a market substitute for these public investments in contexts where states were too weak or too captured by elite interests to provide them. The substitute has not performed as well as the original.

Land rights and secure tenure: For the majority of the world's extremely poor people, who live in rural areas and whose livelihoods depend on agriculture, secure land tenure is the primary asset that determines household economic trajectory. The inability to borrow against land, pass it to children, or invest in long-term improvements without risk of dispossession constrains everything downstream. Programs that secure land rights — at lower cost per household than microfinance programs — produce economic effects that outlast the programs themselves.

Savings rather than credit: Multiple evaluations have found that savings products — particularly commitment savings accounts that make it harder to withdraw funds — produce welfare outcomes comparable to credit access without the risk of debt burden. BRAC's village savings groups, VSLAs (Village Savings and Loan Associations) run by CARE, and other savings-first approaches consistently outperform credit-first approaches on household stability and investment outcomes. The constraint that savings addresses — households' inability to accumulate capital in the face of social pressures to redistribute to relatives — is more commonly binding than the credit constraint that microcredit addresses.

The lesson from two decades of microcredit evaluation is not that financial tools are irrelevant to poverty reduction. It is that credit is a tool suited to a specific situation — someone with a viable enterprise who needs capital to grow it — and has been systematically oversold as a tool suitable for general poverty reduction. The people who are poorest are poor not primarily because they lack credit. They are poor because they lack land, security, health, education, and political power. Those deficits require structural interventions. They will not be solved by a loan.