You might know that I run the Commitment to Development Index, which rates rich countries on how much they are helping poor ones, and involves collecting lots of data (of variable quality) from around the world, making debatable judgments about how to aggregate it, then publishing it annually with somewhat artificial confidence and fanfare in order to broadcast an important message.So you can imagine how indignant I am that the Microcredit Summit (MCS) Campaign regularly collects lots of data (of variable quality) from around the world, makes debatable judgments about how to aggregate it, then publishes it annually with somewhat artificial confidence and fanfare in order to broadcast an important message. Their index is a tally of how many people are using microcredit. In January, the MCS announced that the figure has surpassed 100 million, the goal set at the organization's maiden summit in 1997.One thing my forthcoming chapter 4 draft does is survey the current microfinance landscape statistically. It's not easy, because several data sources are out there, and they are highly problematic and superficially contradictory in their big numbers (global totals). There's the MixMarket, to which self-selected and self-defined microfinance institutions (MFIs) voluntarily report data, and whose buggy "heat map" shows 26 million borrowers in 2007. There's CGAP's eye-opening, one-shot, circa-2000 report and data set, which find some 750 million savings and loan accounts at "alternative financial institutions" such as MFIs, cooperatives, postal savings banks, and state agricultural banks. Stephen Peachey says that if you add private savings banks, you get 1.4 billion.Part of the muddle here is about what we are talking about when we talk about microfinance. Seemingly, for instance, the MCS only counts loans, where others include savings. And those agricultural banks CGAP counts might not be lending just to poor people, which is perhaps why other sources don't count them. MCS includes India's self-help groups (SHGs), which are groups of a typical 15 women that pool savings in a bank account until it grows large enough to qualify them for a loan, which credit they then allocate among themselves. Since no MFI runs SHGs, no institution reports on their behalf to the MixMarket. In effect, they are not microfinance by that site's institution-based definition.I'm not attempting a grand unification of all these numbers and definitions. But staring at these data sets did highlight a few oddities in MCS's high-profile 100 million number:
- The Vietnam Bank for Social Policies accounts for 5.6 million of the total. (Page 44 of the report.) Alone among the big MFIs counted, it is heavily subsidized. So I wonder: why not count other heavily subsidized state banks targeting the poor, especially farmers?
- India's self-help groups account for nearly half the global total, at 49.7 million. In fact, data about SHGs are abysmal. As far as grand totals go, all we know is how many SHGs have ever taken loans (for 2007, see bottom numbers in the first two columns here). MCS multiplies that by an estimated 14 members/group. We don't know to what degree this is double-counting women who join two groups at once. Alok Misra at the Indian MFI rating firm M-CRIL has done extensive research on SHGs and estimates that official figures double-count by 20%, which isn't terrible. However the figures may also double-count over time: many old SHGs may have gone defunct and their members joined new groups. Or they may simply overcount by not excluding women who have dropped out of the program. Alok also told me that because of the top-down pressure in the Indian government to meet SHG expansion targets, many of the new SHGs are borrowing only token amounts. Yet the women are saving in order to win those loans. So the big SHG number, the one that launched the global "microcredit" total past 100 milion, may mostly be about savings. Oddly consistent with this confusion is this definition of microcredit from DSK Rao, MCS's SHG enumerator:
Microcredit Summit Campaign defines “Microcredit” as financial services to the poorest. It is not restricted to loans only. Whereas microfinance refers to financial services to all, including the poorest.
This definition is internally consistent, if unusual. So if the MSC wants to count savings too, why not include the 21 million savers at Bank Rakyat Indonesia and not just the 3.5 million borrowers? Ditto for other big deposit-taking institutions?
Quibbling aside, I don't doubt the Microcredit Summit Campaign's message about numbers. Microfinance is growing fast, whatever it is.
