Microfinance—extending tiny, collateral-free loans to poor entrepreneurs—was once heralded as a near-miraculous tool for ending poverty, on the premise that the poor lack only access to capital. Early evidence came largely from non-experimental studies comparing borrowers to non-borrowers; critics noted that these comparisons were confounded, since people who choose to borrow may differ systematically in ambition or opportunity from those who do not.

To address this, researchers ran randomized controlled trials, randomly offering credit access to some communities and withholding it from others. The trials yielded a sobering consensus: access to microloans did increase business investment and gave households more flexibility in managing their finances, but it produced no detectable average gain in income, consumption, or other measures of well-being over the study horizons. Borrowers were not, on average, lifted out of poverty.

Proponents of microfinance have offered two rejoinders. The first holds that the trials' two-to-three-year windows are simply too short to capture benefits that accrue over a decade. The second concedes the absence of an average effect but argues that averages conceal heterogeneity: a minority of borrowers with genuine business acumen may gain substantially while others, who use the loans for consumption smoothing rather than investment, see no income change—and that this redistribution of opportunity is itself valuable. Skeptics counter that consumption smoothing, however useful, is a far more modest claim than the original promise of mass poverty eradication, and that absent evidence of the long-run gains, the first rejoinder remains a conjecture rather than a finding.