Competition Gets a Pat on the Back

by Richard Rosenberg : Tuesday, February 7, 2012

People who favor a commercial approach to microfinance and people who oppose interest rate caps have argued for years that competition would bring meaningful reductions in microcredit interest rates. Others have been skeptical about this prediction. A new study by Guillermo Baquero, Malika Hamadi, and Andreas Heinen seems to shed considerable light on this question. The study uses a comprehensive data set including not only MIX Market information but also data on 329 microlenders provided by three rating agencies, for a total of 1335 institutions.

The headline finding is that increased competition in a country is associated with substantial declines in interest rates. Interestingly, this effect is driven by for-profit lenders; the study found a much smaller correlation between competition and changes in non-profit rates.

In for-profit institutions, the size of the effect is surprising. The opposite of competition is market concentration, typically measured by the Herfindahl Hirschman Index (HHI). If a given country moves from the 90th percentile on the distribution of HHIs down to the 50th percentile (i.e., the industry is becoming less concentrated, more competitive) the associated drop in the interest rate is almost eight percentage points. This is about triple the rate decrease reported in studies of the regular commercial banking sector. Competition seems to work better in microfinance than in conventional finance. Of course, these are average results. They di not mean that interest rates will decline in every microcredit market that becomes more competitive.

A second question is whether loan delinquency will get worse as competition increases. Does the availability of multiple lenders increase over-indebtedness? The study finds no empirical support for this proposition. There is actually some evidence that loan repayment in for-profit institutions will be better as competition increases.

Does this mean that we don’t need to be so worried about over-indebtedness in more advanced markets? Not really. Competition has to do with the number and relative market power of providers. Market saturation is a different issue: how close is the supply to catching up with the demand for microloans? Adrian Gonzalez reports  that when loans-per-population (a proxy for saturation) exceeds 10%, delinquency and default are higher.

Putting all these results together, competition comes out looking like a pretty good thing. It’s market saturation (rather than competition or growth rates) that seems to be associated with repayment problems.

TAGS: , , , ,

Comments: Comments and trackbacks are open.

6 Comments RSS 2.0

  1. February 7th, 2012 at 7:07 am, Daniel Rozas ()

    Rich,

    this is an interesting paper, though it’s too bad that only the abstract is publicly available — I couldn’t find the paper itself. My comments are more on the overindebtedness side, rather than interest rates.

    Taking the findings at face value, I have to wonder where exactly competition and saturation overlap. I know of no saturated microcredit market that did not feature extensive competition. So it appears that competition is a precondition for saturation. And Adrian’s 10% threshold, while rough, is a reasonable measure of levels where saturation risks becoming unsustainable (my calculations for AP penetration a year before the bubble suggested a 16% penetration level: http://www.microfinancefocus.com/content/there-microfinance-bubble-south-india). So is that where we should draw the line? If saturation is below 10%, then we’re ok? I don’t think so.

    There are several elements to bear in mind. First, competition and saturation tend to be very unevenly spread. Even within a region or a small country such as Bosnia, some locales feature a lot more competition (and saturation) than others. So the 10% rule of thumb needs to be applied with that in mind, which can be difficult given generally available data. I believe this is a point that Adrian emphasizes in his paper.

    Moreover, as you have pointed out in your paper with Jessica Schicks, repayment problems are only lagging indicators of overindebtedness. Greater competition can actually allow the lag to be extended, by giving struggling borrowers the ability to take on multiple loans in order to juggle repayments. Studies of markets below the 10% threshold demonstrate that overindebtedness can certainly manifest itself at far lower penetration levels (see http://centerforfinancialinclusionblog.wordpress.com/2011/11/14/cfi-publishes-over-indebtedness-of-microborrowers-in-ghana-report-by-jessica-schicks/). In that respect, the correlation between competition and high repayment rates is exactly what one would expect until penetration becomes unsustainable. Such high repayment says nothing about borrowers not becoming overindebted.

    My personal take would be that, yes, competition can be a net positive for reducing interest rates, but faced with a highly competitive credit market, I would start to be concerned with growing overindebtedness long before that market reached saturation.

  • February 9th, 2012 at 9:46 am, N.Srinivasan ()

    I endorse comments made by Daniel. while competition might bring down interest rates, it might be fueling excessive debt. AP debt levels – both in terms of number of loans and amounts had risen to very high levels – helped by competition between SHGs and MFIs and between MFIs. Kiteflying with loans from different sources creates a false impression of positive credit culture and low default risks…. Repayment rates improve in competitive markets as one can borrow from Paul to pay Peter and so on. More qualitative information is needed before we can convince ourselves that competition in microfinance does not lead to excessive debt or has a positive impact on repayments.
    Srinivasan

  • February 11th, 2012 at 4:12 pm, Chuck Waterfield ()

    I, too, would like to see the full study. The abstract indicates that they likely completely missed a very key factor as they leapt to the popular-but-very-wrong conclusion that “for-profits are better than non-profits”.

    It appears that their data set looks at price changes without taking into consideration changes in the average loan amounts of those institutions. My analysis of for-profit MFIs over time is that many of them push to higher and higher average loan sizes. Doing so results in lower operating cost ratios, and thus they have the potential to lower their prices. Prices come down not primarily from competition but more importantly from their shift to larger loans. Non-profit MFIs tend to be more concentrated on the very smallest loan sizes in their countries, and they generally stick to those very small loan sizes. Thus, they have higher prices than the for-profits that generally shift to the easier loans.

    You can see some example analysis of this in my Powerpoint. Look for the slides titled “What Causes Microloan Prices to Decline over Time?”, pages 24-35, available here:

    http://www.mftransparency.org/pages/wp-content/uploads/2012/01/MFT-PPT-036-RBI-training-Session-1-Interpreting-India-data-in-context-of-the-global-market-2011-02-ENG.pdf

    A couple of closing points on this topic:

    * I do agree that competition would be a good thing for the poor, but we do not yet have “fair competition” in most of microfinance. We don’t have full price competition because the pricing schemes we use are so confusing that even we can’t understand true prices, let alone our clients.

    * I’m not in favor of interest rate caps, but I would say that a main factor for forcing MFIs to think harder about their pricing has been pricing caps. Ecuador is an intriguing example of how pricing caps pressured MFIs to work harder at lowering costs, lowing prices, and lowering profits while not having significant mission drift. When there isn’t price competition, when there isn’t pricing transparency, when there aren’t pricing caps, MFIs — especially for-profit MFIs — don’t have much incentive to really think hard about reducing the prices they are charging the poor.

    Chuck Waterfield
    That Guy Who Obsesses about Interest Rates Even While He Sleeps

  • February 12th, 2012 at 1:16 pm, Dr V.Rengarajan ()

    I agree with Rich on the association of market saturation with repayment problems in principles. However, I have some moot questions on the validity of such correlation exercise in view of following rationale in general and more from demand side perspectives in particular as the entire exercise seems to be supply (institutional)oriented .
    Theoretical points
    1. Is it adequate to correlate ‘credit saturation’ with competition in credit market in supply side alone while it has more sensitive correlation with credit absorption capacity in the given potential of the area and capability of poor credit borrowers ?
    2. On similar account does ‘repayment’ seriously correlated with the degree of competition in the supply of credit or productive function of the credit with adequate income generation in a given environ in the demand side?
    Practical points (in agreement with N.Srinivasan)
    1. In the absence of ‘quality information’ in this industry due to the persistence factors such as ‘loans from different sources creating false positive credit culture’ and ‘low default risk’ and the phenomenon like borrowing ‘Paul to pay Peter’, unreliable client status data right from MIX to SHG, absence of level playing field in the market environ, how far the facts could be meaningfully adoptable in the typical market analysis using standard economic tools ? and further how far the findings from the correlation analysis would be useful for establishing a convincing conclusion
    In fine while from the demand side perspectives it still remains debatable whether the ‘cost’ or ‘access’ to the micro credit that matters, why one should have obsession with interest rate even during his sleeping?
    Thanks for sharing my views
    Dr Rengarajan

  • February 13th, 2012 at 8:32 am, Chuck Waterfield ()

    I found the link to the entire paper, rather than just the abstract. It is the document currently at the top of the list here:

    http://www.lsf.lu/eng/Research/Working-Papers/2011

    I’ve not had a chance to read it yet, but those interested can now do so.

    Chuck

  • February 21st, 2012 at 4:47 pm, Richard Rosenberg ()

    Apologies for the incomplete link to the paper, and thanks, Chuck, for providing it. (I was working from a hard copy of the complete paper.)

    I’m not good at reading regression tables, but my strong impression is that the authors DID control for loan size, meaning that their results cannot be written off as simply reflecting the lower costs of larger loans. This is consistent with what Adrian Gonzalez and I found in analyzing MIX data a few years ago: admin cost per borrower (which excludes the effect of loan size) has been declining significantly.

    Daniel raises the concern that competition may be causing over-indebtedness that has not shown up in the repayment statistics YET. I agree that this is worth worrying about, but regard it as a hypothesis that awaits confirmation.

    • Leave a Reply