Are Microcredit Interest Rates Excessive?

by Richard Rosenberg : Wednesday, February 25, 2009

High microfinance interest rates is a hotly debated topic in a lot of places. Obviously, the answer for a given MFI depends on its particular circumstances–see for example the CGAP study of the controversial Mexican MFI Compartamos. Now, can we say anything about the overall picture worldwide? CGAP has published a new study based on data from the Microfinance Information Exchange. Individual observers will have their own criteria for judging what is excessive, but empirical data do shed important light on the question:

The very high rates that get so much publicity are rare: for instance, less than 1 percent of microborrowers worldwide pay rates as high as the 85 percent for which Compartamos has been pilloried. In fact, the median rate for 2006 was about 26 percent.
Very high profits for MFIs are uncommon; the median profit on MFI owners’ equity was about 12 percent in 2006, compared with 18 percent for banks.
MFI interest rates have been dropping fast–about 2.3 percentage points a year in 2003-2007, probably due to some combination of learning curve and competition. Commercial bank interest rates have dropped, too, but far less.
Underlying that drop in interest rates are drops in administrative costs (about one percentage point per year) and profits (about 0.6 percentage points per year).
There is strong empirical support for the commonsense notion that it is inevitably more expensive to make many tiny loans than to make a few large ones.

I hope those interested in this question will take a look at the data we developed and form their own views. My own appraisal is that the worldwide picture of microcredit rates is an encouraging one. Given thousands of MFIs, there will always be some whose interest charges and profits might seem hard to square with the best interests of clients, but I can’t find any indication that this is a common problem. I have a list of things that might worry me about the development of microfinance, but price gouging is a long, long way from the top of that list.

TAGS: Interest Rates, mfis, Richard Rosenberg

Comments: Comments and trackbacks are open.

6 Comments

February 27th, 2009 at 12:54 pm, Shaun Campbell ()

The admin cost for MFIs is a truely challenging issue. The adoption of technology is the only way forward to automate processes, remove the paper, create real-time services and reach the remote customers, as follows:
1. m-MFI solutions: utilizing the mobile handset / adapted POS terminals to replace paper with electronic transaction handling
2. ASP for MFI banking applications: funded by the MFI on a very low cost per customer per year ($2 – $3 per end customer per year). Cost savings and efficiency gains within the MFI will far outweigh the $2-$3 cost per client. Coupled with the m-MFI and POS-MFI this nis a powerful end-to-end solution
3. Shared service centres in remote areas, such as a location in a small town that can service a range of MFIs, SACCOs, Banks from a purpose built shop/kiosk/container
4. Share agent networks! Leverage other party’s agents such as mCommerce project agents
5. Leverage other organisations’ office / POP facilities, such as joint working with Cooperatives, post offices

Bottom line – take the paperwork out of the system, create real-time electronic transactions, leverage shared services/agents = lower costs = ability to offer lower interest rates to the customers

This is commencing in Kenya, Rwanda, Tanzania and Nigeria (to name a few) right now.
See recent slides from M-MFI discussion at AITEC Nairobi last week:

http://aitec.usp.net/Banking%20&%20Payment%20Technologies%20East%20Africa,17-19Feb2009/Shaun%20Campbell%20Sybase365%20Compatibility%20Mode.pdf

Shaun

February 27th, 2009 at 3:25 pm, Jonathan ()

This is an excellent concise review of issues and a very useful exercise to ‘decompose’ interest rate charges into components.

One area where you might extend your analysis is to explore the connection between the components. Consider for instance an MFI, let’s call it ‘Seamos’ which is charging 100% interest rates. For a $100 loan they want you to repay $200. You look at the numbers and conclude that, for every 100 dollars of interest expense, say, 60 dollars are due to high ‘operating expense’ (10 to cost of funds and 50 to the cost of monitoring and processing loans) another 30 dollars are due to ‘profits,’ and the rest to loan loss provisions. This 30% is above normal industry profits but you point out that even if profits were squeezed to zero the firm would still have to charge 70% on account of its high costs which you have attributed mostly to small loans and the high cost of monitoring.

But this potentially misses the point that much of the reason that the monitoring has to be so intense (and the part of what may be keeping the loans small, and loan loss provisions high) is that the MFI is trying to extract such a high profit from these borrowers.

It’s well understood that heavy interest expenses tend to increase the risk of moral hazard (the classic reference is Stiglitz and Weiss, 1981). Costly monitoring is the way MFI’s typically choose to reduce this risk particularly of borrowers who cannot pledge collateral, and the high cost of monitoring is indeed an important reason why MFI interest rates are high. But consider an MFI that is earning zero economic profits (i.e. it’s equity holders earn a normal rate of profit). This MFI will use only as much monitoring as necessary to make sure its borrowers can keep up payments.

Now the MFI decides to pursue ‘high profits.’ This will require the borrower to now have to pledge a larger share of project cashflows to repayment. This creates a potential moral hazard (the borrower reasons, why should I continue to work as hard as before only to have the bank tax more of my returns away). If the terms of the contract did not change the most likely effect would be higher defaults (and loan loss provision costs would rise). More likely however the bank adjusts it’s lending methodology and what they have to do is assign more staff time to monitor these borrowers to try to avoid that moral hazard problem (giving out smaller loans might also help).

If the above is true then lowering profits from $30 on each dollar lent might also contribute to lowering loan monitoring costs by $15 and loan loss provisions by $5, and might allow the bank to slightly increase average loan sizes.

The larger point is that when you lower the ‘profit’ that the bank tries to extract, it ought to be able to reduce interest rates by proportionately more.

I have no idea how important this might be in practice, this is simply a theoretical musing, but one that strikes me as potentially important.

To the above argument one might also add that firms that enjoy market power often have higher costs because their employees and managers can demand more (i.e. total rents exceed measured profits because employees and managers are capturing part of the rents in the form of higher renumeration, more perks, and they can afford to be less efficient).

I wonder if your data would allow you test any of these hypotheses. One idea: The way you’ve analyzed the problem ‘profits’ are treated as being independent of operating costs. If the above is true, you ought to find a correlation in the data. You might also find interest rates falling faster than profits.