[For the full review of this paper, click here]

An important recent paper written by distinguished academics examines the impact of loans upon a range of variables. The authors use a sophisticated Randomized Control Trial (RCT) methodology and data from the controversial Mexican Bank, Compartamos. This bank is (in)famous for high interest rates, an extremely profitable IPO in 2007, and high salaries and payouts to certain individuals.

The results are surprisingly gloomy. Despite every effort to paint a positive picture, perhaps their best summary is simply “some good and little harm”.

The analysis takes place over approximately 2 years, thus it is not clear that the long-term impact of indebting poor people (either positive or negative) will be visible over such a short time-span. I have a few questions about the details of the experiment, particularly regarding spill-over effects from borrowers to non-borrowers, and the absence of much discussion on client desertion and delinquency. However, the paper is certainly worth studying in detail, and the results are intriguing. I will summarise a selection here:

• Interest rates charged are 110% APR, for 16 weeks, for an average loan of about $650.
• Most clients are not to microentrepreneurs. The minimum requirement is simply that the client would like a loan.
• Compartamos does not consider over-indebtedness data from the credit bureau important in the loan application process, but does report non-repaying clients to the bureau.
• Apparently returns on micro enterprises are 200%, and yet when offered loans door-to-door, less than 1 in 5 clients took a loan from Compartamos. 30% of these did not take a repeat loan.
• Taking a loan from Compartamos does not reduce loans taken from other sources, but quite the contrary – it encourages clients to take more loans from other institutions.
• Taking a loan from Compartamos does not have any impact on loans taken from moneylenders, shedding doubt on the myth that microfinance replaces borrowings from moneylenders.
• Borrowers’ satisfaction with access to financial services does not change following a loan with Compartamos, although they are less likely to join informal savings groups.
• There is no notable difference in impact upon those with or without a business.
• Loans did encourage clients to grow their businesses (when they had one), but the increase was marginal.
• Loans did not prompt any job creation. They increased business revenues, but also expenses. The net impact on profit was actually negative. There is a very slight improvement in the likelihood of a borrower facing a financial problem (in support of the income smoothing debate for microcredit), but it is not statistically significant.
• A few clients demonstrated a reduced tendency to have to sell assets to repay a loan. But equally they found no evidence that clients used loans to buy assets, thus the iconic use of a microfinance loan to acquire a goat or sewing machine seems challenged.
• There is no evidence that consumption of household groceries or expenses on family events increases as a result of borrowing. There is a minor increase in school and medical expenses, but they are not statistically significant. Borrowers spent less on “temptation” goods, but these excluded junk food despite the obesity crisis in Mexico.
• They examine the impact on various “soft” variables, such as happiness, stress, health, trust in institutions etc. Most results are not significant. Of those that are, there is evidence that credit reduces depression, increases trust in people (but not in institutions), and improves participation in decision-making.

The authors optimistically interpret this in a positive fashion, suggesting that microcredit leads to “some good and little harm” (abstract & page 3). “There is evidence of both increased investment and improved consumption smoothing” (page 3). “Happiness, trust in others, and female intra-household decision power also increase” (page 3). “Overall we do not find strong evidence that the credit expansion creates large numbers of ‘losers’ as well as winners” (page 29).

My principal criticism of this otherwise decent paper is that I was left with an overwhelming suspicion that every effort was made to extract something positive from the data, while brushing over the rather disturbing negative impacts. “Mediocre”, or “disappointing” would be the adjectives I would use to describe their findings. However much lipstick one applies to a pig, it is still a pig. However, towards the end of the paper the authors announce that:

“The main takeaway is some, albeit far from overwhelming, evidence that some people fare worse when faced with expanded access to credit. Several of the sub-groups have more negative treatment effects than positive ones, with the patterns of results for those who are poorer or without prior use of formal credit access perhaps the most eye-opening”. (page 27)

Thus it appears the most vulnerable are the most likely to suffer.

I then place the analysis in context, first by examining the astonishing profitability of Compartamos, identifying the real beneficiaries of the operation (certain individuals and the shareholders, thanks to the high interest rates charged), and reviewing some of the criticism of MFIs charging these rates at such levels of profitability. I then wonder how the results may differ were the analysis carried out in a country where such interest rates are considered exploitative, or were the analysis performed for a longer time-period. Finally, I refer to Dan Rozas’s excellent work on a looming crisis in Chiapas, at the extreme other end of Mexico. It appears the authors of the main paper have selected one of the comparatively more wealthy regions of Mexico suffering the least from over-indebtedness, and in fact the problem may be far worse in other regions.

I end the paper with a quote from one of the authors that may shed light on why such mediocre evidence is interspersed with such positive dialogue:

“Investors should be quite happy and sleep better at night knowing they’re making money and making the world a better place”.

A constant stream of (profitable) poor people is guaranteed for many years, the bottleneck in microfinance is the flow of investors. I speculate that this may be the motivation for the (at times almost desperate) attempt to extract anything positive from this dataset.