n.V1-7.08 | MICROCREDIT
01| Small loans to poor entrepreneurs, mostly women, usually through joint-liability groups, at interest rates well below the moneylender and well above the bank. For two decades microcredit was not one solution among several. It was the solution — a Nobel Peace Prize, a global industry, hundreds of millions of borrowers, and a story simple enough to survive any retelling: the poor are entrepreneurs starved of capital, so lend it to them.
02| Six randomised evaluations, across six countries on four continents, urban and rural, were published together in 2015.
"a consistent pattern of modestly positive, but not transformative, effects"— Banerjee, Karlan & Zinman (Six Randomized Evaluations of Microcredit: Introduction and Further Steps, AEJ: Applied Economics, 2015, Vol. 7, No. 1, pp. 1–21)
03| No dramatic reductions in poverty. No dramatic gains in household wealth. Some improvement in business outcomes, and some improvement in women's empowerment measures, for subsets of borrowers in some settings. Six studies, four continents, one consistent shape.
04| The result deserves a fair reading rather than a triumphant one. Modestly positive is not nothing, and the studies did not find harm at the population level. A credit product that lets a household smooth a shock, buy stock in bulk, or avoid a moneylender at 20 percent a month is doing something real. What it is not doing is lifting people out of poverty, which is what was sold.
05| The instructive part of this case is not the finding. It is the sequence. Two decades of scaling came first. The evaluations came after the industry existed, after the prize, after the national policy frameworks, after hundreds of millions of people had taken loans. The question "does this work" became answerable only once the answer could no longer change very much.
06| That sequence is not a microcredit problem. It is the default in this sector, and it recurs in every chapter here: a model becomes institutionally established on the strength of its story, acquires staff, funders, conferences and a constituency, and is then evaluated by people with no power to unwind it. Microcredit is simply the clearest case, because the evaluation was unusually good and unusually collective, and because the gap between the claim and the finding was unusually wide.
07| One distinction is routinely lost and matters. Microcredit is not microfinance. Credit is one product; savings, insurance and payments are others, with thinner but genuinely separate evidence bases, and with a different logic. A savings product does not put a poor household into debt to test a business idea; it lets a household hold money away from the demands of the people around it, which in many settings is the binding problem. Payments infrastructure has changed what is possible for poor households in ways credit never did. The microcredit verdict has been generalised onto all of it, and that generalisation is not supported.
08| What is open: whether the modest business effects concentrate in the same way the asset-transfer results do — that is, whether borrowers above some threshold of existing assets or skill convert loans into occupational change while those below simply service debt. The six trials report averages. If the threshold logic holds here too, the average is hiding two different stories again, and the useful question is not whether microcredit works but which borrowers it works for and how a lender could tell in advance.
