Retort to the Malegam Report

In the last quarter, bombshells have been dropped on microfinance in both India and Bangladesh. I have criticized how both have been delivered---the Andhra Pradesh law for its extremity, its suddenness, and its drafting by officials with vested interests, the Tom Heinemann documentary for its apparently pure accent on the negative. But it has been interesting to watch the effects of both unfold, and they are not all bad. So complex are the sequelae that either explosion could yet do more good than harm.Case in point: In India, the imperious microcredit ordinance, while throwing the industry into disarray, did draw the eyes of the country and the world to serious problems. In direct response, the Reserve Bank of India (RBI), the country's central bank, just released a report of a committee hastily appointed to study the regulation of microcredit---a responsibility that the RBI had neglected. That is good in itself.As a piece of analysis done under high temporal and political pressure, the report of the Malegam Committee (as it informally known, named after its chair) is admirably organized, concise, and thorough. And as I have noted, India, quite unlike Bangladesh, enjoys a wealth of experts who are not only smart and logical but quick to say what they think. Already, M.S. Sriram, N. Srinivasan, and IFMR have chimed in, and more will follow. So it is with humility that I comment on this debate. I expect to learn and modify my views in the days ahead.Sriram's bottom line is harsh:

The Andhra Pradesh ordinance (as recommended by the sub-committee) can lapse because the panel has effectively completed the task that was started by the state government.

I.e., the AP law would be redundant because the Malegam recommendations would finish off the industry nationwide.However, the #1 microfinance institution, SKS, disagreed. And its shares jumped. Perhaps the new regime would eliminate the small fry, but the big fish could survive and might even benefit from reduced competition:

"These regulations will not have much impact on us. But if somebody doesn’t have enough capital, it might affect them," claims SKS chief financial officer, Mr S. Dilli Raj.

#2 Spandana concurred.Which is perhaps why the Andhra Pradesh government is disappointed in the report:

No proper implementation machinery has been suggested. Reliance is placed on the very same MFIs---who have been violating the RBI guidelines at will, and caused the current crisis---to self regulate. The committee failed to appreciate that State Governments are mandated to protect the people from money lending activities as per...the Indian Constitution.

The debate over what to learn from and do about the Indian microcredit crisis reminds me of the parallel one for the global financial crisis. Especially in the early days of the latter, people were struggling to understand what caused the crisis. (Then most of us gave up.) Many candidates were offered: cheap capital from China (digging deeper, the cheap yuan); too little regulation; too much regulation; Fannie Mae; ratings agencies paid by the firms issuing the securities they rated; greedy Wall Streeters who took the fees and ran before the loans came due; greedy Main Streeters who took the loans and defaulted when they came due; securitization; unscrupulous mortgage brokers....Contemplating the confusion back then, I realized that the search for a single cause was muddled, for two reasons. First, the focus on causes ignored the question of agency. Suppose it was determined that sunspots contributed to the mortgage meltdown. Blaming sunspots would not help. Better to blame the parts of the system that humans control---for not being robust to sunspots. That's a fanciful example, but consider whether it is a good metaphor for human greed, which also seems to be a fact of nature. Arguably, it does not do us much good to blame the crisis on greed. Better to blame it on rules that did not fully take into account the consequences of greedy behavior. Now, I realize that's a somewhat facile distinction. After all, the rules are made by self-interested people too; these agents are not angels. Still, the politicians and regulators do sometimes act in the public interest so I suppose it is on them that our best hopes rest for agency in the public interest. Thus, as a practical matter, the search for causes converges to a focus on what these public servants should have done differently, taking as given human greed and the yuan-dollar exchange rate.The second thing muddling the search for a single cause is that sometimes causal factors interact. Some people pointed out that the financial crisis hit many countries, with quite different regulatory systems, so you couldn't just blame a national factor such as Alan Greenspan. The real cause was the huge swell of capital, much from certain developing countries, which inflated a bubble. But others said, not so fast: thanks to firm banking regulations, Canada came through just fine even though it was tied to the same global capital markets. So actually domestic policy failures were the cause, you see. How to square this circle? Both sides were right. At the risk of oversimplifying, the crises in the U.S., U.K., and other countries arose from the combination of easy money and bad policies. If either had been eliminated (leaving aside the question of what agent would have done that), the crises would have been prevented. Thus we could blame---and adopt policies to redress---either factor alone and be sort of right.But ideally, policymakers would recognize both factors, survey possible policy changes that could affect either, then choose from among them all in light of what is known about costs, effectiveness, and political constraints. It appears to me that the Malegam Committee did not do the analogous thing for Indian microcredit. If the Indian microcredit industry is a poorly made tire that springs leaks at high pressure, then the committee decided without apparent reflection that its job is to aggressively patch the tire, not reduce the pressure. In fact, it wants to increase the pressure:

We...recommend that bank lending to the Microfinance sector both through the [Self-Help Group]-Bank Linkage programme and directly [through microcreditors] should be significantly increased.

In particular, it recommends that Indian banks still be able to count loans to microcreditors in satisfying the government’s "priority sector" lending quotas. Clearly the committee members haven't been reading the Financial Times, or they would have been won over by this eloquent letter to the editor:

Let us attack the problem of fast growth by slowing the spigots that inflated the bubble. Indian banks should no longer be able to count loans to microcreditors in satisfying the government’s "priority sector" lending quotas.

Conceptually, as I have already implied, the Malegam Committee and I can both be right. Restricting the money flow to microcreditors might prevent another crisis, as might restricting what microcreditors do with the money. The best combination of reforms under these two headings depends on specifics. I will not get extremely specific here since plenty of people more knowledgeable are doing that right now. But I will give the general flavor of the committee's recommendations and make a general comment.Malegam would confine microcredit to a tidy box. Microfinance institutions (MFIs) could lend only to people in the classic joint-liability groups, never to individuals outside them. People could belong to no more than one such group (or self-help group). Each group could borrow from at most two MFIs. Each person could borrow no more than Rs. 25,000 from MFIs (25,000 rupees = ~$550). The maximum household income of microcredit borrowers would be Rs. 50,000/year (just $240/person in a family of five). 75% of each MFI's loans would go for "income-generating activities"---investment rather than consumption. Borrowing the structure of Muhammad Yunus's rule and tightening the parameter, the spread between an MFI's cost of funds (such as the interest rate it pays big banks) and what it charges customers could not exceed 10%---or 12% for small MFIs. (Yunus proposes 15% for all MFIs.) MFIs could levy at most three charges: interest, an origination fee of at most 1%, and an at-cost premium for credit-life insurance, which cancels the loan if the borrowers dies. The effective interest rate on a loan could not exceed 24% (whether before or after compounding is unclear). No loans could be given on collateral. (I guess only the moneylenders should do that.) A lender would need to devote 90% of its portfolio to microcredit so confined in order to qualify as a legitimate, regulated MFI---or less than 10% and be regulated as something else. All MFIs would need Rs. 15 crores (150,000,000 rupees = ~$3.3 million) in capital on hand to absorb losses. They could sell credit only---no deposit-taking, no health insurance, no laundry soap.If these rules went into effect, and if they worked, they would certainly end the epidemic of multiple borrowing and apparent overborrowing. They would also prevent a lot of potentially constructive lending, such as to people who earn more than that quite low income threshold. Now, regulators always face such trade-offs. And there is a lot to said for conservatism. Back in the day, U.S. bankers joked that they worked under the "3-6-3 rule": pay depositors 3%, charge borrowers 6%, and get to the golf course by 3pm. Banking was boring. Good thing we deregulated banking...And there is much to like in the recommendations: the new regulatory category for MFIs, the simplification of fees, the requirement that interest rates be calculated honestly and disclosed prominently, the endorsement of the use of credit bureaus, and more.But it does look like the committee has projected a lot of the traditional banking approach onto a non-traditional kind of banking, and that creates risks. I worry that the proposals are not adequately informed by an understanding of how microcredit is done and how it is used, and thus make fatally impractical assumptions.For example, most poor people do not earn salaries. Ergo, their incomes are not fixed, precisely measurable quantities. One day a farmer is weeding his cotton crop, earning nothing but anticipating good earnings at harvest. The next day a flood destroys his crop. A computation of his income based on historical data would not reflect the change in status; one based on future income would be highly conjectural. Then too, measuring historical income is hard. Trained researchers in South Africa, who probably cost more than the income flows they measured, "found that it took half a dozen rounds of diary questionnaires, even after the initial questionnaires, before we felt confident that we had full information on the cash flows of a household." The definition of a household is probably also flexible when applying for loan. ("My brother and his income don't belong to our household anymore. He moved out last week.") What the Malegam Committee seems not to understand is that group microcredit targets the poor not by measuring their poverty, for that is impractically expensive, but by being so inconvenient that only poor people bother with it.No easier to gauge is whether 75% of loans are invested. Even if MFIs could spare the expense to monitor every borrower (while staying within the spread and interest rate caps), there is the problem of fungibility. Consciously or not, a borrower might use a loan for a goat she would have bought anyway. The money she would have spent on the goat in the absence of the loan instead pays for school. Ergo, the loan is really financing school, not the goat. Likewise, while everyone favors credit bureaus, and Indian MFIs are at this moment working to share data on borrowers, these too are easier said than done. Especially without a national ID system, it becomes difficult to detect and prevent multiple borrowings by the same person under different names. Poor people are very good at making the most of what comes their way. (Or am I exaggerating this problem? Could loan officers for two different MFIs in the same village quickly match up their loan records?)And while the Yunus spread limit is more credible when applied within India than when applied globally (in Latin America and much of Africa, microcredit just costs more to do), it still discriminates against poorer people in harder-to-serve areas. India is big and diverse. The proposed limits may work in parts of India where microcredit is common now; but presumably where microcredit is less common, that is because it is more expensive to do. So the limits on the spread and interest rate may confine microcredit to India's more fortunate regions. The 24% figure is not based on any analysis of what rate would best serve the poor, but rather on MFIs' current cost structures. Even if MFIs could break even in Bihar at 28%, and even if that rate would help people, it would become illegal. The report seems not to understand that the cost of microcredit delivery can vary substantially with demography and geography.Granted, if some of the proposed rules are impractical or misdirected, that alone does not invalidate the strategy of patching the tire. Perhaps conservative, imperfect restrictions on microcredit delivery are the best practical option. So the question is: what's the alternative?I have tended to view the Indian microcredit industry as a system and asked what characteristics of that system are dysfunctional. That has led me to focus on the dangerous positive feedback loop that drives all credit bubbles: fast growth makes all borrowers look creditworthy, which "justifies" more fast growth. In this view, pushing credit emerges as a core dysfunction. Credit markets are not normal: the more people buy, the more they want to buy, and the better everything looks. Shoving credit into such an inherently unstable system seems inherently irresponsible. That is why the priority sector lending rule concerns me. It undermines whatever fragile impulses there are in the system toward moderation. Would ending or restricting the consideration of microcredit as priority sector lending---or otherwise regulating the flow of credit at the upstream end---work better than restricting how microcredit is done? Would some of each be best? I don't know. I do know that focusing on the upstream end is more administratively practical, for that is where one finds the conventional banks, which the Reserve Bank knows well how to regulate. The Malegam committee's focus on the other end---the filigreed delta where loan officers meet clients---may send the RBI into waters far less navigable than it realizes. I would trust the report's guidance more if it were clearly based on a good map.Update: commenter Jonathan C reminds me just below that looking at the Indian microcredit industry as a system with potential dysfunctionalities should lead to a discussion of more than the priority sector lending rule. Beth Rhyne has written most cogently in this vein, from an international comparative perspective. E.g., also up for discussion should be whether MFIs could take savings and whether non-profits could be given more voice on MFI boards.