Print
Share
Grameen Bank, Which Pioneered Loans For the Poor, Has Hit a Repayment Snag
February 9, 2010
As I blogged before, one of the last articles Daniel Pearl wrote for the Wall Street Journal before he was abducted and murdered—coauthored with Michael Phillips—exposed financial woes at the Grameen Bank. Appearing on page 1 on November 27, 2001, under the headline Grameen Bank, Which Pioneered Loans For the Poor, Has Hit a Repayment Snag, the piece described how some Bangladeshis were juggling loans from several microcreditors at once, how others had banded together to protest and resist the Bank’s policies, and how the Bank’s loose accounting standards and slow disclosure hid a decline in loan repayments.
This post shares new data that suggest that history is repeating itself in important ways. The Grameen Bank, indeed all big microcreditors in Bangladesh, may be finding it harder to collect on loans. As far as the evidence goes, there has been no epidemic of default. But the combination of years of rapid growth and accelerating declines in key indicators of delinquency are so reminiscent of the lead-up to the global financial crisis that the broad implications hardly need explaining. A partial meltdown in the Mecca of microcredit would not sow the same economic destruction—microfinance is not the heart of Bangladesh’s economy in Schumpeter’s sense—but it could have lasting implications for microcredit worldwide.
In his autobiography, Banker to the Poor, Muhammad Yunus describes how Grameen grew from an idea, to a project with his students, to a formal branch of a state bank, to an independent bank. By the mid-1990s, the Grameen Bank was a national operation with a global reputation. However, growth then slowed and the Bank ran into trouble persuading its borrowers to pay back. Stuart Rutherford:
Arrears on loan repayments began to grow, and more and more clients stopped attending the village-level weekly meetings where bank business is conducted. Then in 1998 Bangladesh suffered its worst floods in living memory, disrupting the bank’s work in almost two thirds of the country and dealing its balance sheet another severe blow.
The system consisted of a set of well-defined standardised rules. No departure from these rules was allowed. Once a borrower fell off the track, she found it very difficult to move back on, since the rules which allowed her to return, were not easy for her to fulfill. More and more borrowers fell off the track. Then there was the multiplier effect. If one borrower stopped payments, it encouraged others to follow.
Grameen’s performance in recent years hasn’t lived up to the bank’s own hype. In two northern districts of Bangladesh that have been used to highlight Grameen’s success, half the loan portfolio is overdue by at least a year, according to monthly figures supplied by Grameen. For the whole bank, 19% of loans are one year overdue. Grameen itself defines a loan as delinquent if it still isn’t paid off two years after its due date. Under those terms, 10% of all the bank’s loans are overdue, giving it a delinquency rate more than twice the often-cited level of less than 5%.
Some of Grameen’s troubles stem from a 1998 flood, and others from the bank’s own success. Imitators have brought more competition, making it harder for Grameen to control its borrowers.
…
…microlending has lost its novelty. In Tangail, signboards for rival microlenders dot a landscape of gravel roads, jute fields and ponds with simple fishing nets. Shopkeepers playing cards in the village of Bagil Bazar can cite from memory the terms being offered by seven competing microlenders….Surveys have estimated that 23% to 40% of families borrowing from microlenders in Tangail borrow from more than one.Borrowers have also become more rebellious. “The experience was good in the beginning,” says Munjurani Sharkan, who became leader of a Grameen group in Tangail’s Khatuajugnie village in 1986. To put pressure on “lazy” group members who were slow making payments, she says she used to start removing the tin roofs of their homes. But one day, the whole group decided to stop making payments.
They were protesting Grameen’s handling of a fund it created for each group, using 5% of each loan and additional mandatory deposits. The “group fund” was meant for emergencies, but many borrowers wanted to withdraw money from the group fund. After a protest movement, complete with placards and amplified speeches, Grameen finally agreed to give borrowers easier access to the fund.
The troubles and the exposé left several legacies. First, even before Pearl and Phillips got onto the story, Grameen embarked upon a program of transformation eventually called “Grameen II.” Its hallmarks were simplification (of the various loan products), computerization, flexibility, and the taking of voluntary savings deposits. In response to the article, Yunus publicly regretted the Journal‘s failure to tell this story of positive change (and published all his e-mail exchanges with Pearl and Phillips!). He lacked credibility since Grameen’s forthrightness had just been called into question; yet he was substantially right. Grameen II confounded the critics. Today, the icon of tiny loans for the poor does more microsavings than microcredit. Portfolios of the Poor extols the flexible new “topping up” system that lets people borrow back loan balances after 26 weeks of on-time payment.
A second legacy: Grameen overhauled its metrics of loan delinquency and began posting them monthly on its web site. It thus set a standard of transparency that its main rivals, ASA and BRAC, are far from matching.
Third, and far less important, when I started at the Center for Global Development in early 2002 and listed topics I could work on, my new boss Nancy Birdsall lit on “microfinance”; I think the recent Pearl and Phillips article was in the back of her mind. So it’s one reason I am writing this today.
Grameen II also confounded the critics in embarking upon a new round of growth—from 2.4 million Grameen members at the end of 2001 to a stunning 6.9 million at end-2006 and just under 8.0 million today. ASA and BRAC kept pace—in fact, closed the gap—so that that all the big three clustered around 6 million borrowers at end-2008. So strong is the convergence that one wonders whether the three are lending to the same 6 million households. (Not all 8 million Grameen members borrow at a time.)
Indeed, multiple borrowing is widespread in Bangladesh now, and it has raised concerns that some Bangladeshis are juggling microcredit loans the way some Americans juggle credit card debt, in a merry-go-round that must one day stop. The worry, in other words, is that there is a microcredit bubble. In 2007, Shafiqual Haque Choudhury, founder and head of ASA, which is known as the most commercially savvy of the big three, worried aloud about a “train crash.” And that was before the global financial crisis, which has probably been transmitted into poor Bangladeshi households via lower exports of clothing made in Bangladeshi factories and fewer construction jobs in the Middle East for Bangladeshi workers.
Yet so far, at least from afar, tranquility seems to prevail. Repayment rates, to the extent they are reported, have remained high. In the last year, Portfolios of the Poor has depicted microcredit as a source of stability for Bangladeshi families more than instability. And Rich Rosenberg implicitly leaned on the high repayment rates in microcredit’s exemplar nation in arguing that since people keep repaying loans over many years, and there have been only scattered credit meltdowns, most poor people must be avoiding gross over-borrowing.
If Bangladeshi microcredit was approaching a train wreck—or at least a bumpy stretch—where would we see it first? On the website of the most transparent large microcreditor in the country. A few days ago, I visited Grameen’s site, and was surprised to find this trend (full spreadsheet):
[Technical note].
I think about this graph in two ways. One is by focusing on the ending level of 96.54%. That seems high in absolute terms. But it is low by historical standards. And Rich Rosenberg, in his authoritative field guide to delinquency metrics (quoted by Pearl and Phillips) and in a recent post, explains that a 95% collection rate can spell disaster. On a one-year loan with weekly repayments, lent taka (the Bangladeshi currency) return to the microcreditor over 0–12 months, so the average taka makes a lender-borrower round trip in 6 months, and can be immediately relent. Thus a 95% collection rate can lose 5% of capital in 6 months, and 10% in a year—maybe manageable if interest rates are high enough, but not trivial. In addition, non-payment is contagious, as Yunus noted. Once delinquency starts feeding on itself, the costs of cajoling and pressuring for repayment can skyrocket.
Another way to analyze the graph is by focusing on the recent change. Whether or not Grameen Bank is yet in the red zone, it seems likely that something bad is happening. In Rosenberg’s language, the on-time collection rate graphed above is an excellent red flag indicator because it plummets as soon as borrowers start struggling. It is a leading indicator of trouble, a canary in the coal mine.
The on-time collection rate does have a weakness, though: paying late is not paying never. Delinquencies do not automatically harm a creditor greatly in the long run. This is why Rosenberg recommends the microlenders also compile indicators of more protracted difficulties—and why Grameen does so. Its monthly tables list total amounts owed by people that have missed 5–9 consecutive payments (1–2 months of weekly installments). By this measure, if you miss the first 5 payments on a $100,000 mortgage, the entire $100,000 is counted as at risk. Grameen does the same for those missing 10 or more payments.
This graph shows all three of Grameen’s delinquency indicators. The blue line is the same as that above, but flipped for comparability from a collection to a non-collection rate. The two added indicators are for Grameen’s core credit product, the basic loan, which can last from 3 months to 3 years and has weekly payments. These too point to degradation:
So what is the story behind the numbers? I imagine two main possibilities. One is that the global financial crisis is mainly to blame and little is fundamentally wrong with microfinance in Bangladesh. The other is that a “train crash” is indeed occurring, however mild or severe. The global crisis may merely have popped the bubble. Both of these stories are about Bangladeshi microcredit in general, not just the Grameen Bank. I focus on the Bank only because it does such a good job of publishing delinquency indicators. [Update: a bit more analysis in this post.]
I spoke last night with the leading foreign observer of Bangladeshi microfinance, Stuart Rutherford, and he strongly favored the second explanation. More than 650 microcreditors operate in Bangladesh. Multiple borrowing is widespread. At least until very recently, there has been no national ID system, so creditors have been unable to share information in order to track how much debt people are juggling, the way creditors in rich countries do through credit bureaus. The microcreditors are flying blind. Some winnowing seems inevitable and healthy.
How will the big three fare? ASA, according to Rutherford’s Pledge has more money than it knows what to do with. It is lean and mean, financially strong. It stopped opening new branches in 2007 and went into reverse, closing or amalgamating some. BRAC borrows its capital more than the others, which may make it vulnerable, but also receives more support from donors for the way it integrates microcredit with other activities to help the poor.
Finally, Grameen holds 45.4 billion taka in savings for its members, against 56.1 billion in outstanding loans. To the (unclear) extent that the individuals who owe the bank the most are those who save the most with it, Grameen should be in reasonable shape: defaulters will lose their savings. (That said, I wonder whether recent drops in Bangladesh’s commercial interest rates will place another strain on the Grameen. It has committed to paying 12%/year over 10 years (10% over 5) on its Grameen Pension Scheme savings plan. Adding the 38.4 billion taka in savings of non-members puts Grameen in a net savings position, and the earnings on its net savings are going down. See the bottom of the “Monthly data” tab in my spreadsheet.)
More graphs may shed more light. This one shows what may be a problematic pattern in the evolution of the Bank, and perhaps microcredit generally in the country. Whenever the Bank has stopped growing, delinquency has flared:
This graph shows that despite the current plateau in membership, the Bank has maintained growth in taka terms by increasing lending per member:
I tried to ask the Grameen Bank about the story behind the statistics. But the phone number on its Contact Us page has yet to answer, messages to the posted e-mail address bounce, and the web-based e-mail form produces an error message. I hope that this post will elicit an informative response.
Concluding thoughts
- I do not know whether Bangladesh’s microcreditors are in major trouble. But if they are, the denouement will manifest much as in the graphs above.
- It is tempting to link the degradation in the Bank’s portfolio to the mysterious dismissal of Deputy Managing Director Dipal Barua in December. I have no evidence for a link. However, when a financial institution forces out its head of operations while key indicators are going south, it raises questions. Shareholders—mainly the Bank’s members—donors, and other stakeholders deserve an explanation.
- A crisis in Bangladesh, akin to recent ones in Bosnia, Morocco, Pakistan, and India, would tarnish the image of microcredit worldwide, perhaps permanently.
- Other microcreditors should quickly match Grameen’s standards of financial disclosure so that we can get a better read on the extent of trouble.
- The Grameen Bank has repeatedly defied the skeptics and flourished for 34 years. This history should instill humility in any who would declare Grameen compromised now.
- Yet the Bank and its competitors have not quite proved that they can thrive without growth, as they must for permanence. A history of ending lending problems by outgrowing them is not entirely reassuring.
19 Comments on “Grameen Bank, Which Pioneered Loans For the Poor, Has Hit a Repayment Snag”
Post a Comment
We value frank and constructive exchanges and encourage you to use your real name in your comments.
February 10th, 2010 at 12:55 am
I think it is because the bank is growing too rapidly. At one time, the bank was opening two branches per day–more than Starbucks. The comparison with Starbucks is rather ironic–it is in serious trouble.
Global financial crisis had little effect on Bangldeshi MFIs. The only indirect effect is via remittances. Surprisingly, remittances to Bangladesh did not slow down during the financial crisis.
The concern with multiple borrowing has been with Bangladesh’s microfinance industry since the 1980s. People has been predicting disaster for a long time and are still waiting. What people forget is that it could be a sign of unsatisfied demand for credit and Portfolios of the Poor clearly suggests that the demand for credit by the poor in Bangladesh is multifaceted and rather dynamic.
But what you are noticing is worrisome, no doubt. I am hopeful that “history will instill humility in any who would declare Grameen compromised now.” I remember vividly the schadenfreude in the DevFinance listserv after publishing of the story by the WSJ.
February 10th, 2010 at 12:12 pm
Asif, it looks like you are right about the (lack of) financial shock for Bangladesh:
2009 remittance sets new benchmark
And according to the IMF’s International Financial Statistics, its exports jumped from 700 billion to 800 billion taka between 2007 and 2008. (No data yet for 2009.)
February 10th, 2010 at 1:19 pm
This has been waiting to happen.
As we noted in our Grameen Bank II Briefing Note # 8:
This excess liquidity (which is growing in absolute terms year on year) took the bank into lending more to its client and starting individual lending too. The former risks repeating the same error the bank made in the early 1990s (in response to having to raise salaries by 66% – since they were tied to the Government’s salaries). The latter is a whole new territory for Grameen Bank and one that it may struggle to manage effectively.
I would also note that when I visited Grameen Bank a couple of years ago and reviewed the data input sheets from (a very few) Kendras in the Tangail, they were a blizzard of red ink on the pre-printed repayment forms … most clients were making variable payments – not just on their loans but also on their GPS recurring deposit accounts … which led me to worry about the erosion of discipline through the GBII system (even though Credit Officers resisted implementing it in the field).
February 10th, 2010 at 1:52 pm
Has the canary keeled over in the coal mine? I don’t have the quantitative skills or the background to be able to analyze the data, but they don’t look good. In addition, if Grameen, an exemplary MFI, is experiencing these kinds of delinquency and default rates, what must the less established, less efficient MFIs be facing?
February 10th, 2010 at 3:22 pm
Ryan, as far as I can understand, the numbers are not so terrible that the Grameen Bank is obviously in huge trouble. So I suspect that one has to go beyond the numbers to answer your question, to find out exactly what is happening at Grameen and the other microfinance groups. We need a good investigative reporter.
February 10th, 2010 at 4:09 pm
This is quite fascinating. An interesting point is that the 5-9 and 10+ overdues appear to have anticipated the sudden jump in the one-time collect rate (if that’s what these numbers really are). If you look at just the 2-year period since Jan 2008, you note a relative flatness in all 3 curves over 2008, and then the 5-9s and 10+s start rather fast climbs in 2009, with the most recent delinquents (blue line) waiting until Sept 2009 to start climbing…
Ok, so it’s tough to explain graphs with words, but if true, this seems a bit unusual — rather then seeing an overall increase in short-term delinquency and subsequently greater long-term delinquency as the past-due loans begin aging, the pattern is reversed, with those in delinquency seeming to stay there longer, and new loans starting to falter only later. I know nothing of Bangladesh economy or possible events that could explain this, but this is not what I’d generally expect to see.
Also, the sudden spikes as seen in Sept 09 generally have a specific trigger associated with them. If this were a train wreck, I’d expect a rapid, but still gradual decline, rather than a one-month spike. If there’s nothing in the external environment to explain it, I wonder whether there may not have been some internal issues at Grameen during that month — the firing of the COO perhaps?
BTW, the absolute numbers are still generally low and not a cause for concern. The real question is where does this trend go, which we won’t know until we get there.
February 10th, 2010 at 4:18 pm
Interesting points, Daniel. Yes, I noticed that funny timing relationship. I didn’t know what to make of it. It calls for talking to human beings who actually know what is going on. I did notice too that the green line has several discontinuities. I wonder if certain overdues were accumulating in the accounts but were not recognized as soon as they formally ought to have been, and were then recognized after certain administrative decisions were made. If there is a sort of “administrative error term” in these overdues indicators, then perhaps the timing of their movements should not be interpreted literally.
February 10th, 2010 at 11:46 pm
The delinquency graph is an interesting one. After continuously rising, the defaults of both 5+ and 10+ instalment defaulters seem to be turning – do I detect a small decline over the last three months.
I am not clear about the data underlying the graph. The 5+ instalment defaults should be more than 10+ instalment defaults if the 5+ defaults also includes the 10+ defaults. If these two mutually exclude each other category, then we get a default level of more than 3.8% (1.5 from 5+ and 2.3 from 10+). But the overall defaults are just 3.5%. Is there some more information available on this?
I would argue that we should look at more pragmatic recovery rates than we are willing to in case of microfinance. If higher transaction costs are acceptable to the sector as it serves remote and small clients, why cannot the same logic be applied to look at a more realistic default rate? And let pricing factor in the higher risk level and ensure institutional sustainability.
The offer of 12% interest on long term savings has to be seen from a customer protection point of view as well. The lack of high-interest lending opportunities might compromise saver’s interests and the very existence of the institution.
Grameen has to be complimented for making information available in the public domain which makes such analysis possible.
N.Srinivasan
February 11th, 2010 at 1:56 am
I believe the two overdues categories are mutually exclusive, based on the descriptions in the Grameen statistical reports and on the fact that sometimes one is larer and sometimes the other is. Grameen just posted the January 2010 data, so I added it to the graphs in the post. For “basic loans” the two categories now add to 4.26% of outstanding amounts. Both rose a bit in January.
Of course there are two ways for overdues to go down, with very different implications: people catch up on their loans or debts are permanently written off by the Bank. A footnote to the statistical reports states that for “10+” overdues, “Entire outstanding amount of overdue loans are written off one year after they became overdue.” So it’s clear what rising overdues means, but less clear what falling overdues means.
February 11th, 2010 at 3:23 am
I spent 2 years following the bank and the other Grameen companies, and I have to say during that time Bangladesh was hit with several severe floods – one that displaced 27 million people for up to 6 weeks, and one of the largest cyclones, which wiped out huge agricultural areas. That combined with the world economic collapse which sent certain prices soaring was a huge problem for so many who are on the brink of poverty could all have been factors in this current decline.
The borrowers can dip into their savings accounts for a while and during these natural disasters the bank often stops collecting repayments for a while while they are recovering. It is really complicated when there are so many factors.
There certainly seemed to be a concern that some members were borrowing with several agencies and they did their best to avoid that, but it’s difficult to tell.
And of course, you have to factor in the problem that MOST bankruptcies around the world are due to health problems. When there are floods and cyclones and people have damage to their homes, their risks of disease increases. If someone is sick in the family, people will spend everything to save them.
That is precisely why they started Grameen Kalyan.
Just to note, the new bank branch that I was following in the far north was profitable and self-sufficient within 1 year and 4 months.
Also of note, they have implemented over 50,000 “struggling member” loans to women who are not yet capable of being “regular members” because they may be too old or have other life circumstances that makes them seem to be beyond starting a successful entrepreneurial business. These women are given 500 Taka instead of the initial 5000 Taka loan and it’s at 0% interest instead of 10% and they can pay back in any amount of time they wish, many taking years.
I saw so many people’s lives being vastly improved, it was really deeply inspiring. Certainly you can never have one solution that helps everyone, but here you have a solution that helps so many, it’s value to society is immeasurable.
I would wait to see what happens in the world, to know if this is a trend that matches the world’s economy and the other factors specific to Bangladesh. But I think it is too soon to tell.
February 12th, 2010 at 1:25 am
It is interesting to observe increasing trend of non-performing assets in the books of MFIs especially because these institutions were reckoned with respect because of their almost zero percent bad debts. However the missing link in the study is the linkage of growth of income earning both capacity and opportunity with the savings and repayment capacity. It is imperative comprehend this linkage as borrowing is not borrowing sake but enable the borrower to earn more and improve their living standard and also to help them to become self reliant and not look for loan unless he or she is looking for upgrading their standard and likes to seize the growing opportunity in their surroundings because of stimulus given by the state or even by larger corporate entities under their social mission program. It is obvious therefore necessary to study the correlation of this two objectives and locate the trend of the impact on each other to identify the success or failure of MFIs.
February 12th, 2010 at 8:40 am
Hi David,
This comment may be a bit off the mark but could not put it anywhere else in your blog!!
while doing my PhD research (looking at the ‘institutional entrepeneurship’ aspects of microfinance), I have come across evidence that that group lending existed beyond Grameen in the mid 1970s; this forces me to speculate the Dr Yunus did not ‘invent’ group lending (NOT cooperative credit, I am talking about plain vanilla class group lending as popularised by GB).There were World Bank projects which had ‘group lending components’ almost at the same time as Yunus was experimenting with GB. I have a copy of a Ohio State University thesis on “Repayment and Group Lending in the Province of Camarines Sur, Philippines”.
Please comment!
Thanks
Amer S.Khan
PhD Student
Faculty of Economics and Business
University of Sydney
February 12th, 2010 at 12:12 pm
Amer, there is no question group lending has been around for a long time. The British introduced one version of it into what are now India, Pakistan, Bangladesh, and Burma just over 100 years ago. And this certainly was a source of ideas for Yunus and his students. Still, I credit them with devised a model that worked–that got high repayment and scaled to millions. Ford didn’t invent the car, if you see what I mean. For more read my draft chapter 3 and chapter 4.
February 13th, 2010 at 10:14 pm
Graham Wright’s comment about Grameen’s need to push the money out the door to guarantee the interest rate on the popular GPS savings scheme points out the need for enabling macro environment for the success of mature microcredit industry. Theoretically, GB should be investing the savings in the financial market. Unfortunately, the financial market in Bangladesh is underdeveloped. The stock market is rather thin. GB has parked some of the savings in two widely subscribed mutual fund schemes. The total money invested in the mutual funds are only a small fraction of the total savings poor. If the financial market was well-developed with various options for investing the large savings pool, GB would not have to grow so rapidly. This suggests that a matured microcredit industry will need a mature macro environment for its success. In our book, The Poor Always Pay Back, we already anticipated the problem and suggested that the Government should allow GB to park some of its savings outside the country. We were thinking out loud; we didn’t beleive the Govt will allow such move.
February 14th, 2010 at 1:36 am
I would like to pick up two topics touched upon in this article:
a) The Pearl-Philipps quote describes the change in attitude and behavior of clients over time. I am reminded of some freshly formed women groups a colleague of mine had formed for their joint producing (mushroom growing, trade, tailoring, and the like). The groups were very enthusiastic, and proud of their sites and savings they practiced, so were we. Only my local colleague (this anecdote is from Uganda) “spilled the beans” by dryly remarking that all groups would start like that; then after two cycles they would become “stubborn” and delinquency kick in; at that point, one would best refer them to the “tough” MFIs (FINCA, PRIDE, UML [now Equity Bank] have that reputation in Uganda), at NGO-level they would become unmanageable. I do not know if there is research that looks at this change over time and disentangles its elements? The economist in me would immediately say “here you have the empowerment all these NGOs like to brag about”. Then there is probably the conflict between individual and systemic rationality – individuals taking advantage out of the given system may destroy the same; however, it is the job of the system-providers, not the individuals (and certainly not moral education), to give a system within which individuals’ choices are confined to those that cannot destroy the system (apparently, in finance that is notoriously tricky; I think Paul Krugman summed the dilemma up nicely when he pointed out that financial service providers always have an uneven balance sheet: Liabilities side short, assets [loans] long; in MFIs,both tend to be short but that probably doesn’t help). Yet another element, I suppose, is about self-selection as a system (like our NGO, or Grameen Bank, etc) grows; it attracts a certain kind of people; who can make themselves comfortable with the choices it provides (e. g. Al Fernandez pointed out that some people are just not for groups). Like Windows attracts Hackers, it probably always attracts a share of people who are particularly comfortable with the weaknesses of that system. Wouldn’t it be worthwhile to study these issues for some MFIs over the 1.5 or so decades they have existed? Some lessons for designing the organizations could be there, I guess.
2) I feel the theme “multiple borrowing” is rushed over unduly in this article. It is a theme that deserves a more thoughtful treatment than just waking the usual prejudices and moving on. I would like to recommend Sa-Dhan’s Side-by-Side report 2007 for a more differentiated discussion of the issue. Also, the cited “Portfolios of the Poor” is full of stories about “multiple borrowing”, giving an array of reasons why it appears. The “lights and shades” study by EDA/APMAS also has something to tell; which by the way links into the previous point, because it looks at how people plan their repayment towards different “systems” (“Portfolios” implies the same). The bottom line is that multiple borrowing as an observable – and ubiquitous – fact does not explain anything; less so for a clientele that is poor and thus almost by definition over-indebted. If we are to strengthen our “systems”, we have to isolate those portions of the fact which carry undue risk. There we find that it is the vulnerability of clients’ businesses and households that pushes over-geared ones over the edge. Of course, behavioral economics have shown that all people (not confined to MF clients) tend to under-estimate the risk of credit, and thus to over-borrow (I highly recommend Thaler/Sunstein for the presentation of that bit). But again, with the people we are targeting, under-estimating their future risks is truly the least of their problems. Sure, large MFIs (Bangladesh and elsewhere) should be made (by the regulators) subscribe to a mitigation mechanism – i. e. credit bureau – to limit “credit-push”. How to make that mechanism work in favor of a target group that borrows mostly informal, nobody has answered yet. One thing should not be overlooked: A classical credit bureau is likely to have the same effect on the target group as Grameen’s procedures end of the 1990ties had (see Yunus’ quote above). Fresh thinking is required!
Cheers
Oliver
February 14th, 2010 at 2:36 pm
I think this analysis has not looked into the breakup of what kinds of loans are overdue and I think that is a key missing piece.
As per Grameen’s provisioning policy, if loans are overdue for 10 weeks, then the loan is restructured as a flexible loan.
Further the loan loss provisioning on flexible loans is very significant.
Here is the summary of their overdues by type of loan as of January 2010.
Flexible loans: $19.279M out of $27.01M is overdue (71.38%)
Basic loans : $12.95M out of $756.75 is overdue (2.5%)
This means that flexible loans constitute a disproportionately high level of all overdue loans.
A worrying factor would be if flexible loans are increasing dramatically in size on the overall loan portfolio.
I looked at the proportion of flexible loans in December 2006, which seems to be a logical starting point based on the overdue graph.
Flexible loans were $24.72M out of a $475M portfolio ( i.e 5.2%).
Currently they are $27.01M out of $811M (i.e. 3.3%).
Had overdues been rising and if there were a bubble forming, then the percentage of flexible loans would have increased. On the other hand, they have come down. There is however a small piece of data missing in this analysis and that is the overdue amount in flexible loans as of December 2006.
Nevertheless, the fact that flexible loans percentage has come down and it seems to indicate that all is well at Grameen and there is no bubble or even froth in the horizon.
This confirms what Prof. Yunus has been stated all along, despite the financial crisis the poor have been paying back.
And further despite the financial crisis, poor borrowers in Bangladesh can enjoy 12% on their savings account (My three banks accounts in the US give me 0.1%, 0.2% and 1% respectively). Thus through the financial crisis Grameen has managed their asset as well their liability side in an exemplary manner. Hats off to them!
Bhalchander
February 14th, 2010 at 3:59 pm
Bhalchander, I think you make a good point. You raise a technicality that I did not fully discuss in my post. Indeed, when people have trouble keeping up on a standard (“basic”) loan, it can be converted to a “flexible” loan. The recovery rate (the blue line in the first two graphs) includes flexible loans. But the red and green lines in the second graph, as I mentioned, are for basic loans only, which in fact account for 93% of outstanding credit.
I left flexible loans out for those two lines because I wasn’t sure how to interpret the stock of flexible loans. Are these loans just a way of making default look good? Or they a great way to help struggling members get back on the “microcredit highway” of the basic loan?
But you are right that they matter. In response to you comment, I just did a version of the second graph that counts flexible loans for the red and green lines too. To be precise these are, for all loan types: total amounts owed by people who have missed 5–9 (or 10+) payments as a % of total amounts outstanding. To me it sends pretty much the same message.
Now, I could do: (total amounts owed on non-flexible loans by people who have missed 5–9 (or 10+) payments + all amounts outstanding on flexible loans) as a % of total amounts outstanding. Your comment seems to suggest that. I lean against doing it this way in part because most people with flexible loans appear to be in perfect standing with the bank. In other words, what I just did is consistent with the Grameen Bank’s own definitions.
And here is a graph of flexible loans outstanding as a share of flexible+basic loans. You’re right that the statistic is historically low. But is has been rising since the middle of last year.
What do you think?
February 15th, 2010 at 2:36 am
David, overall I think there are a lot of things happening and it is very difficult for me to analyze the situation on the ground looking at graphs alone. Nevertheless based on the information I see some remarkable trends:
1. Average outstanding per member has dramatically gone up after 12/07 from 5000 Taka to 7000 Taka (40%). While the amount was almost constant prior to that for nearly 5 years ( maybe even longer but graph 4 does not have all the historical data).
2. There has been a tremendous increase in savings from both members and non-members. e.g.
From 12/07 to 1/10, member deposit has gone up from $430M to $656M ( $226M – 52% growth). Non-member deposit has increased from $326M to $555M ($229M – 101% growth).
The average saving has also increased per member, however the exact growth in savings per member is unclear from the data. It is possibly at the same level since the growth in borrowers has plateaued over the last few years.
What I infer from this:
1. Since savings are increasing and average loan size is increasing this perhaps implies that borrowers are moving into bigger, more riskier, higher gestation period businesses with less predictable cash flows in the short term. Since loan sizes are linked to savings, I think Grameen knows the risks they are taking.
2. The less predictable cash flows from businesses are perhaps leading to higher outstanding amounts in the short term i.e. 0 to 10+ weeks.
3. However by and large the businesses seem to be working out. We could speculate this since the percentage of flexible loans have not gone up in the period despite the financial crisis.
4. With the flexible loan product Grameen has identified a remarkable model to support struggling members while maintaining their dignity as ‘full’ members rather than as ‘defaulters’. They perhaps also have additional capacity building measures for such members. They have also in some ways smartly figured out a way to isolate and treat delinquency contagions with this product.
An ancillary point, non-members savings are growing much faster than member savings (101% versus 52% growth over the same period). This seems to imply that Grameen management is comfortable mobilizing non-member savings at the current growth levels. They perhaps see it is as strategic and are anticipating even bigger increases in the outstanding amount per borrower in the future.
One more point and this is purely my gut feel – Members are also perhaps more regular in the savings for the pension scheme which gives them 12% return than on paying their loan outstanding due to irregularity in their short term cash flows. (In some ways it is counter intuitive and irrational. But it is very similar to what happens in the US – people go delinquent on their credit card debt which charges them 25% – 30% rather than go delinquent on their mortgage which costs 5%). This may be one more factor leading to higher overdue amounts in the 1-10+ weeks.
Of course one would need field interaction and more data to really know what is happening. I rest my case here.
Bhalchander
February 16th, 2010 at 6:18 pm
Some geekish glosses on David’s excellent post:
Grameen’s most recently reported recovery rate—96.5%—is defined in a footnote as “amount repaid as percentage of amount due.” Unfortunately they don’t tell us what “amount due” means. Is it the amount that fell due for the first time during the month? Or does it also include amounts that fell due in previous months and are carried over (because they are still “due”)? If the figure includes such carry-overs, then it is nearly useless for estimating the percentage of the loan portfolio that is being (or is likely to be) lost.
But let’s suppose for the moment that they’re measuring the most useful way, which is to count all cash payments received during the month in the numerator of the fraction (either principal alone or principal plus interest) and to include in the denominator all amounts that fell due during the month FOR THE FIRST TIME, UNDER THE TERMS OF THE ORIGINAL LOAN CONTRACT (so re-scheduling loans doesn’t change the figure.) Recovery rates tend to bounce around from month to month, so it’s better to look at them over six months or a year. So let’s further assume that Grameen’s recovery rate, measured the most meaningful way, stays at 96.5%. This implies a loss of a bit under 7% of loan portfolio to bad loans every year. The conventional wisdom for years has been that when annual loan loss rates get above 5%, you have to fix the problem right away or things will spin out of control. When borrowers see so many of their fellow borrowers defaulting, credit discipline tends to go up in smoke.
I certainly hope that Grameen is using the meaningless recovery rate definition rather than the meaningful one.