Debunking the microfinance bubble

Is there a “microfinance bubble”? The question recently hit the front page of the Wall Street Journal. A storm of responses followed, including one from CGD’s David Roodman and another from SKS’ Vikram Akula. Both Roodman and Akula question the evidence for this so-called bubble. This week, the Economist jumped in, adding a big dose of (much needed) perspective.

Here are a few thoughts. First, there's very little good evidence. We don't have a good measure of over-indebtedness, which is the real fear in a bubble.  There's talk about collecting and defining measures, but no real resolution on how to do it. Debt as a percentage of income is sometimes mentioned as a benchmark, but one issue is that it’s usually hard to define income as seasonality and informal employment complicate measurement. Plus, it’s typically hard to see the full extent of household debt and saving (one of the revelations of Portfolios of the Poor).

The WSJ purports that microfinance clients are being “carpet bombed” by loans. The fact is that the typical low-income person (with a profile like that of microfinance customers) lives a life involving multiple loans: some from informal sources (family, friends, moneylenders) and -- if they're available -- from microfinance or formal-sector banks.  The evidence from Portfolios of the Poor suggests that most people manage a handful of financial devices at any one time, patching the pieces together to get the timing and volume as close to what is needed as possible.  This is normal.  It isn't cause for alarm in itself.  It happens because none of the pieces solve the whole problem on their own.  Microfinance very rarely displaces informal finance.

In the places where microfinance is expanding fastest, like India, customers do have loans from a number of microfinance institutions.  Other rapidly growing markets include parts of Bangladesh, Peru and Uganda.  And while we haven't seen anything that looks like a crisis anywhere yet, there are three reasons for concern:

(1) Diminished incentives: the more access to multiple loan sources a customer has, the less incentive there is to repay a particular loan.

(2) Lack of transparency: loan officers need to assess whether households have adequate cash flow to service a loan, and that's hard to assess when staff can't see the full financial picture.

(3) Borrowing to repay another lender can hide and exacerbate the eventual crisis -- which is the bubble issue.

The first two problems are, quantitatively, more important than the bubble concern. But all three problems can be - and are being - addressed by credit bureaus that share information. Though credit bureaus don’t exist in Bangladesh or India yet, they do seem to work quite well in other places. (But new technology can help make credit bureaus more effective.) The one hitch is that they are unlikely to include much reliable information on moneylender loans and loans from friends, which can be a big gap in information.

The article also discusses how microfinance clients took loans for such “questionable” purposes as to pay overdue bills and buy food for their families. Similarly, Stuart Rutherford’s work reported in Portfolios of the Poor shows on how 43 Grameen Bank borrowers use their loans -- a small sample to be sure. These clients report to Grameen that the loans are used for productive purposes, but more than half of them actually use them for daily needs, like putting food on the table, paying for medicine, education, and paying off more expensive loans -- which many would agree, is not such a bad thing.

I don't believe consumption loans are necessarily a problem. In fact, I think people should generally have greater access to them.  I'd be sunk without my credit card, and that kind of cash flow management tool is exactly what many poor people need.  True, credit cards can get you into trouble -- as seen in the U.S. -- but that's partly because U.S. credit card companies make their money by not curtailing over-borrowing.

Microfinance loans are generally set up well for consumption purposes, involving repayment through a stream of small installments.  And even "business" loans are often repaid in part through "non-business" income -- a spouse's wage, a son's remittance, etc. One of the lessons from Bank Rakyat Indonesia, perhaps the world's most successful commercial microlender, is that smart loan officers pay attention to existing household cash flows, wherever they may come from.  The stated loan use is often of secondary importance in judging creditworthiness.

The problem of a microcredit bubble shouldn't be dismissed.  Lenders operate with limited information and do sometimes over-lend (aided by their competitors). Grameen got into a repayment crisis in the late 1990s in part because they were moving too much money out the door too quickly -- and their clients couldn't absorb it.  That’s not the full story, but my conversations with Grameen staff suggest it’s part of it. There will likely be a big microfinance failure someplace, bigger than what we've seen to date. Very little lending is secured, so the risks can't be dismissed.

But the problem will likely be relatively localized, and steps to address the problems are being implemented, albeit patchily.  I don't see any evidence at all for something like a global bubble.

Neither lending for consumption purposes nor the fact that customers are borrowing from multiple organizations are necessarily problems. Both features could -- and should -- be part of well-functioning financial systems.  The real problem happens when lenders have limited information on what's going on.

So, what to do:

  1. Push much harder for credit bureaus.
  2. Financial literacy programs might help, though the evidence so far is weak.  It seems like the time for more and different approaches.
  3. Lenders should introduce more flexibility and drop the practice of pressuring customers to take another loan (often of bigger size) immediately after one is paid off.
  4. Lenders should allow customers to borrow in a wider range of maturities and installment schedules - to better match their household cash flow.
  5.  Lenders should open up to non-business loans – money is fungible anyway, plus you can't really stop people from doing what they think is best with the money, and you may be encouraging people to be less than truthful. 

All this should be done anyhow -- whether or not there's a risk of bubbles.