How to Tell Good MFIs from Bad MFIs

Most of us working in microfinance want microloan clients to be paying interest rates that are as low as possible. While we have the same vision, there is disagreement about how to determine whether an interest rate is an appropriate one.

Some people, including Mohammed Yunus, are worried about the growing commercialization of microfinance, including the entry of profit-motivated owners and managers. They are concerned, reasonably enough, about possible “mission drift,” especially in the form of interest rates rising to (or staying at) excessive levels. In his book and in many presentations, Professor Yunus offers a straightforward formula for judging MFIs and their objectives:

• If you’re a real microlender who cares about the poor, then your interest margin (the difference between the rate you charge when lending to your clients and the rate you have to pay when you borrow from your funding sources) should be no more than 10%. That’s the “green zone” where true microlenders operate.

• If your interest margin is 10-15%, a big warning sign is flashing because you’re in the yellow zone.

• Anything above 15% is the red zone, where you’ve left true microcredit behind and joined the loan sharks.

Unfortunately, when you look at the evidence, this appealingly direct formula turns out to be pretty far off the mark.

To begin with the conceptual problem, the formula doesn’t allow enough room for legitimate differences in administrative costs among MFIs. For an MFI that makes especially small loans or serves a sparse rural clientele, administrative costs will inevitably be a higher percentage of loan portfolio, and the lion’s share of the interest rate spread goes to cover those costs. Application of the proposed formula could actually discourage outreach by such MFIs to poorer clients.

But concepts aside, how does the formula match up against actual MFI experience? It turns out that this formula would place most of the world’s MFIs in the red zone—the average interest rate spread for MIX MFIs in 2008 was over 20%. But to be fair to Prof. Yunus, that shouldn’t end the discussion. After all, maybe plenty of the MFIs in the MIX are charging their borrowers rates that are way too high.

Now let’s test the green-yellow-red formula against a group of Grameen-approved MFIs. Christoph Kneiding and I analyzed MIX data on Grameen along with several dozen MFIs that received support from the Grameen Foundation and reported to MIX. In 2007, for instance, 33 MFIs (representing about two-thirds of the Grameen Foundation recipients) reported to the MIX. The only one in the green zone that year (interest spread below 10%) was Grameen Bank itself. Seven were in the yellow warning zone (10-15%). All the other 25 were up in the red zone (above 15%) and most of them way up in the red zone (between 30 and 55%). The three preceding years looked pretty much the same.

The proportion of Grameen affiliates in the red zone was about the same as the worldwide proportion: for instance, 75% of all MIX MFIs were in the red zone in 2008, according to a new study by Adrian Gonzalez of MIX. NGOs were more likely to be in the red zone than for-profit MFIs, suggesting that interest spreads may be driven more by the higher costs of smaller loans than by profit maximization objectives. (Average loan size in NGOs is about a third of what it is in for-profit MFIs.)

Has the Grameen Foundation has been fooled into working with a bunch of red-zone partner MFIs that are wolves in sheep’s clothing? Far from it. The Grameen partner MFIs that look so terrible on the green-yellow-red test actually appear quite strong—in fact, well above average—on indicators normally thought to be associated with commitment to the poor, such as average loan size. Nor do they appear to be inefficient: they average considerably lower on cost per borrower than the other MFIs in their countries.

It’s disappointing that simple formulas can’t help much when it comes to appraising things like mission drift or fairness of interest rates. It takes a more complex analysis (see, for example, the CGAP papers on microcredit interest rates and Banco Compartamos). I hope we see a lot more MFI-by-MFI analysis, in which the reasonableness of interest rates is judged by the reasonableness of the costs and profits that produce those interest rates. We all want to see MFIs charging clients rates that are as low as possible, so we need analytic tools that can do a credible job of separating the sheep from the goats in that regard.