The case for microfinance was originally put in strong terms. The poor were said to be entrepreneurial but capital-starved; lend them small sums at rates far below what the village moneylender charged, and businesses would grow, incomes would rise, and repayment would follow. Group liability, in which members of a small borrowing circle stood behind one another, was said to solve the problem of lending without collateral.
Randomised evaluations across several countries have produced a more modest picture. Access to microcredit reliably increases borrowing and business investment. It does not reliably increase household income, consumption or the schooling of children — the outcomes the strong case promised. The finding is remarkably consistent, and it has been accepted even by researchers sympathetic to the sector.
One response is that the outcome being measured is the wrong one. A household that can borrow to meet a medical emergency, or to smooth a harvest-to-harvest gap, has gained something real even if its annual income is unchanged; credit is useful as insurance, not only as capital. On this view microfinance succeeded, and was simply advertised under the wrong description.
A second response is less comfortable. Where lenders compete for the same borrowers and assess creditworthiness loosely, a household may hold four or five loans at once and service each by borrowing from the next. Group liability, which was meant to enforce repayment through mutual monitoring, can then turn coercive, with the pressure applied by neighbours rather than by the lender. Crises of over-indebtedness in Andhra Pradesh and elsewhere followed roughly this pattern, and the regulatory response — caps on interest, limits on the number of lenders per borrower, mandatory credit bureau reporting — was directed at exactly these features rather than at lending to the poor as such.
