The Evaluation of the Design Competition of Rebuild by Design
Jan Feb 2008 Review
1. Federal Reserve Bank of St. Louis
REVIEW
J A N UA R Y / F E B R UA R Y 2008 V O LU M E 9 0 , N U M B E R 1
Thinking Like a Central Banker
William Poole
The Microfinance Revolution: An Overview
Rajdeep Sengupta and Craig P. Aubuchon
A Primer on the Mortgage Market and Mortgage Finance
Daniel J. McDonald and Daniel L. Thornton
Changing Trends in the Labor Force: A Survey
Riccardo DiCecio, Kristie M. Engemann, Michael T. Owyang,
and Christopher H. Wheeler
2. 1
REVIEW Thinking Like a Central Banker
William Poole
Director of Research
Robert H. Rasche
Deputy Director of Research 9
Cletus C. Coughlin
The Microfinance Revolution:
Review Editor
William T. Gavin An Overview
Rajdeep Sengupta and Craig P. Aubuchon
Research Economists
Richard G. Anderson
Subhayu Bandyopadhyay
James B. Bullard 31
Riccardo DiCecio
A Primer on the Mortgage Market
Michael J. Dueker
Thomas A. Garrett
and Mortgage Finance
Carlos Garriga Daniel J. McDonald and Daniel L. Thornton
Massimo Guidolin
Rubén Hernández-Murillo
Kevin L. Kliesen
47
Natalia A. Kolesnikova
Christopher J. Neely Changing Trends in the Labor Force:
Edward Nelson A Survey
Michael T. Owyang
Riccardo DiCecio, Kristie M. Engemann,
Michael R. Pakko
Michael T. Owyang, and Christopher H. Wheeler
Rajdeep Sengupta
Daniel L. Thornton
Howard J. Wall
Yi Wen
Christopher H. Wheeler
David C. Wheelock
Managing Editor
George E. Fortier
Editor
Lydia H. Johnson
Graphic Designer
Donna M. Stiller
The views expressed are those of the individual authors
and do not necessarily reflect official positions of the
Federal Reserve Bank of St. Louis, the Federal Reserve
System, or the Board of Governors.
F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W J A N UA RY / F E B R UA RY 2008 i
5. Poole
casting effort. However, some private forecasters up to an FOMC meeting. Thus, I ordinarily do
have access to data and information the Fed does not give detailed answers to questions on the
not. Large financial firms in particular have access precise implications of the latest data for the
to data, such as credit card activity and prospec- economic outlook. In many cases, I just haven’t
tive borrowing by major clients, that the Fed does studied the implications thoroughly, although I
not have. Retail firms have extremely current certainly do so by the time the FOMC next meets.
information on sales and orders. Of course, the
Fed may obtain some of this information through
its business contacts, but private companies often DEALING WITH RISK
make much more systematic use of their own
internal business information than the Fed does. A private firm, especially a financial firm,
Forecasters continually provide updates based must have robust policies to address risk. To an
on the flow of current information, both statisti- economist, risk is a two-sided concept. Outcomes
cal and informal. In this regard, Fed and market may be above or below prior expectations. The
practice is essentially identical. possibility of an outcome far below expectation
There is, however, a difference between the deserves special attention, for such an outcome
Fed and the market in the use of forecast informa- may force a firm into bankruptcy. A financial firm
tion. Traders and portfolio managers base their models risk quantitatively, to the extent possible,
trades on the current flow of information, which and then examines with great care the extent to
needs to be updated throughout the trading day. which formal models may miss key risks, perhaps
Fed policymakers, on the other hand, do not con- because they were not observed during the sam-
tinuously adjust the stance of policy in the same ple period used to fit a model or because the eco-
way managers adjust portfolio holdings. For this nomic environment may be changing. A central
reason, my own practice is not to worry much as bank has a similar task. Quantification of risks to
to whether I have correctly absorbed the import the economy should be taken as far as possible
of each day’s, or each hour’s, data. I know that and then careful thought applied to risks beyond
some information will be irrelevant to my policy those that can be captured in models.
position because it will be superseded by new One important difference between a financial
information by the time of the next FOMC meet- firm and the central bank is that a firm has a much
ing. For example, I do not need hour-by-hour wider array of strategies available to mitigate risk
information on security prices. When I get to than does a central bank. A financial firm can
the next FOMC meeting, I’ll have the latest data, make careful calculations of the extent of duration
charts of how security prices have behaved mismatch between assets and liabilities and can
since the previous meeting, and analyses of price adjust its positions continuously to control the
behavior over a much longer period—indeed, for extent of mismatch. A financial firm can deal in
as far back in time as I find helpful. Given that many derivatives markets to control risk. A finan-
the FOMC does not adjust policy continuously, cial firm has wide latitude in choosing how much
updating my forecast with every data release risk to accept.
would not be an efficient use of my time. A central bank pretty much has to accept
A consequence of the fact that FOMC meet- policy risks to the economy arising from the
ings occur at six-week intervals, on average, is economy’s institutional structure and market
that when I give a speech and take questions I may environment. Market sentiment, bullish or bear-
not be completely up to date on the implications ish, can change quickly. Analytically, the central
of the latest data. In my speeches and discussions bank can explore implications of various possible
of policy with various audiences, I try to concen- scenarios and can engage in special information
trate on longer-run issues and general principles. collection to try to understand as quickly as pos-
I emphasize that I will be studying all the data sible what is happening in the economy. Beyond
and anecdotal information in the days leading that, what a central bank can do is to adopt from
2 J A N UA RY / F E B R UA RY 2 0 0 8 F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W
6. Poole
time to time a somewhat asymmetric policy directly in market prices, especially in the federal
stance in an effort to control risk, especially by funds futures and options markets and the similar
guarding against particularly costly possible out- markets in eurodollars. There is an important
comes. When inflation risk is the dominant con- policy purpose for the Fed to study these market
cern, policy should lean on the restrictive side expectations. Understanding how the flow of new
and policymakers should be more ready to tighten information affects market expectations can be
than to ease policy. Conversely, when deflation useful to policymakers. For example, suppose I
and/or recession risk predominates, policy should interpret a surprise change in employment to be
be asymmetric toward policy ease. However, an anomaly in the data but I observe a large market
there is always the danger of leaning in one reaction to the data release. In that case, I would
direction too far or too long; policymakers must reexamine my interpretation, and if I still believe
be prepared to reverse course and should try not I am correct I might comment during the Q&A
to allow the stance of policy to drift too far from session after a speech that my own personal take
a baseline approach. on the data differs from the market view. My
It is worth emphasizing that the central bank, aim would be to prompt market participants to
as the dominant player in the money market, is reexamine their interpretation of the data.
in a different situation than is a competitive firm. Consider another example of the importance
The central bank’s strategy in mitigating risk of tracking market expectations. When I examine
must work through the markets and by shaping the federal funds futures market, a large discrep-
accurate market expectations about future central ancy between market expectations and my “best
bank behavior. guess” of the FOMC’s future actions might suggest
The list of possible risks facing private firms to me the possibility of a Fed communications
and central banks is a long one. A risk that is failure. The ideal situation is one in which the
often incompletely understood by those outside market and the Fed have read available informa-
management is reputational risk. The issue is tion the same way. I am only one participant in
much more than simple embarrassment. Trust is the FOMC process, but I try not to contribute to
an essential capital asset for a financial firm, and market misunderstanding of monetary policy.
for a central bank. A damaged reputation can send The market is collating information from all FOMC
customers fleeing to competitors. For a central participants, paying especially close attention,
bank, a damaged reputation can lead market of course, to the Chairman’s views.
participants to question the bank’s policy consis- I also follow market data carefully as part of
tency, its motivations, and even its veracity. For ongoing research on how market expectations
these reasons, successful private sector firms and are formed. This research, conducted with econ-
central banks both invest heavily in programs omists in the St. Louis Fed’s Research Division,
and procedures to ensure fair dealing and high helps me to understand monetary policy at a
ethical standards. With regard to reputational risk, deeper level. My perspective in this research is
the issues inside and outside the central bank essentially the same as similar research conducted
are essentially identical. Financial firms and in universities and by active market participants.
central banks understand each other very well
on this dimension of managing risk.
OBJECTIVES
ASSESSING ODDS ON FED Private firms have the goal of profit maximiza-
tion, whereas the central bank is pursuing the
POLICY ACTION macroeconomic goals of price stability, employ-
Market participants are constantly assessing ment stability at a high level, and financial market
the odds on Fed policy actions at upcoming stability. The private sector and monetary policy
FOMC meetings. These assessments register goals are quite different, but that fact does not, in
F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W J A N UA RY / F E B R UA RY 2 0 0 8 3
7. Poole
my view, define an important difference in I pointed out earlier that both market partici-
approach. pants and policymakers try to understand the
Policymakers think in terms of a loss function implications of the flow of information for policy
that depends on departures of outcomes from actions. Now I want to emphasize the important
desired outcomes. Policy goals are quantifiable point that policymakers have the task of design-
and, as with profits, come with short and long ing systematic policy responses to new informa-
horizons. As already discussed, private firms tion. The design should advance achievement
and central banks must understand and control of policy goals, such as price stability. There are
risks to the extent possible. many dimensions to policy design. A simple
Private firm and central bank governing and example is that the Federal Reserve now adjusts
disciplining processes are, of course, quite differ- its target for the federal funds rate in multiples of
ent. Nevertheless, analytical approaches to achiev- 25 basis points. That may seem a trivial example,
ing goals are quite similar. I do not believe that but in the past the Fed sometimes adjusted its
differences of objectives and governing processes funds rate target by smaller amounts. Another
define an essential difference between the two example is disclosure of the policy decision
types of organizations. Thus, in this respect those promptly after the decision. That practice started
in the private sector and in the central bank only in 1994 and ever since the FOMC has almost
understand and relate to each other easily. constantly grappled with disclosure issues.
I could point to many other dimensions of
defining a policy rule, or response function
PRICE MAKERS VERSUS PRICE (Poole, 2005). My point is not to elaborate on the
TAKERS nature of the policy rule but instead to emphasize
What is a critically important difference how different that responsibility is from that of a
between a central bank and a private financial firm portfolio manager. Policymakers should shape
is that the central bank, in the short run anyway, their policy actions by conscious decisions about
sets a policy interest rate and importantly influ- how to guide market thinking not just in the con-
ences longer-term interest rates through effects on text of a particular policy action but also in the
market expectations. The central bank is a price future for policy actions in general. Put another
maker in the interbank funds market. Private way, when economic conditions recur, policy
financial firms are essentially price takers in that responses to the same set of conditions should
market and in the government securities market. also recur. If that were not the case, then policy
A typical trader or portfolio manager can actions could be interpreted only as random,
plan security purchases and sales with little or unpredictable responses to changes in economic
no regard to any effects on market prices or the conditions. It simply cannot be good policy for
behavior of other firms. Of course, this statement policy actions to be essentially random.
is not precisely true for very large portfolios, but The market interprets every policy action
the difference in market impact between a central and every policy statement in the context of past
bank and a large private portfolio is enormous. actions and statements. What is a surprise and
The fact that a central bank is a price maker what is expected depends on past practice. The
makes its strategy fundamentally different from implication of this obvious point is that every
that of a portfolio manager. To achieve policy policy action needs to be based on an understand-
goals, the central bank must think of its policy ing of how the action will be regarded in the
actions as following a predictable policy rule that future. Policy actions set precedents, and policy-
the private sector can observe. A portfolio man- makers must be careful about those precedents.
ager responds to the flow of new information Otherwise, what appears to be a policy success
partly as it affects probabilities of future central today could be the seed of a policy problem in
bank action. the future.
4 J A N UA RY / F E B R UA RY 2 0 0 8 F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W
8. Poole
Modern macroeconomics emphasizes the tions would justify particular appropriate policy
importance of policy predictability for good policy responses. One way to avoid misinformation is
outcomes (Taylor, 1984). The difference in per- to avoid providing any information. Put another
spective from standard practice 30 years ago is way, if my mouth is not open, I cannot put my
profound and incompletely recognized by many foot into it.
journalists and commentators. Even in the early In my view, however, it is important to try to
Greenspan years, many thought that monetary convey correct information. I do not believe that
policy worked by creating surprises. That perspec- I would be doing my job if fear of providing mis-
tive was natural because policy surprises had information led me to provide no information.
visible effects on security prices. For this reason, I have maintained an active speak-
Theoretical developments in macroeconomics ing schedule.
in the 1970s emphasized that policy surprises I do follow some general practices designed
were undesirable. Efficient planning in the private to reduce missteps. I try to schedule speeches,
sector requires that expectations about govern- and certainly press interviews, for times when
ment policies be accurate, or as accurate as the the markets are closed. That allows the market to
inherent uncertainty of the economic environ- digest what I say overnight. Another practice is
ment permits. Policymakers ought not to add to that I never predict the outcome of future FOMC
inherent economic uncertainty. It is desirable that, meetings. Given that I am only one participant in
to the maximum possible extent, the economy those meetings and that the Chairman’s opinion
be characterized by an expectational equilibrium carries great weight, predicting the outcome
in which the market behaves as policymakers would be foolish. That is obvious, but what is
expect and the central bank behaves as the market less obvious is that I do my best to avoid being
expects. There are certainly times, however, when committal even in my own mind about the policy
policy surprises are unavoidable. implications of recent data. Clearly, I could draw
So, much of my own thinking is driven by an conclusions from available data that would create
effort to help define a policy that will increase a certain presumption about the policy decision
policy predictability over time. In my speeches or at least about my policy position. I am very
and ensuing Q&A, I try to emphasize general
cautious about drawing firm implications about
policy principles rather than the current policy
policy from the data.
situation. What is important is not the policy
I emphasize that my policy position will
action at the next FOMC meeting, which is typi-
depend on all the information available at the
cally what people want to know, but the policy
time of the FOMC meeting, on the staff analysis,
regularity that will extend across many FOMC
and on the debate during the meeting. That
meetings, which is what people should want to
description of my attitude is literally correct. I
know.
noted earlier that I often do not focus on the data
arriving day by day because I know that new data
will supersede existing data and that I will benefit
AVOIDING POLICY from my own intensive preparation before each
DISTURBANCES meeting. I rely on the expert staff analysis pre-
An important corollary to the task of defining pared for each FOMC meeting. Given the com-
a policy rule is that the central bank ought not to plexity and dynamic nature of the issues, I find
be a source of random disturbances. All of us are it best not to form a settled policy position well
well aware of the potential for saying things in advance of the meeting.
inadvertently that will create market misunder- Moreover, what policy purpose would be
standing of likely future Fed policy actions. Or, served by my discussing publicly every twist and
more precisely, what needs to be understood is turn of my analysis between FOMC meetings?
how and why various possible economic condi- Market effects from doing so would not serve a
F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W J A N UA RY / F E B R UA RY 2 0 0 8 5
9. Poole
constructive policy purpose—indeed, they would We have tentative signs that the financial
violate one of the important findings in macro- markets are beginning to recover from the recent
economics that policy should not create random upset, but financial fragility is obviously still an
disturbances. issue. If the upset were to deepen in a sustained
way, it might have serious consequences for
employment stability. As of today, we just do not
BASICS OF POLICY STRATEGY know what the consequences may be. My best
guess is that the inherent resilience of the U.S.
I have emphasized the importance of the economy along with future policy actions, should
central banker perspective in conveying a policy they be desirable, will keep the economy on a
strategy. I will conclude by sketching the appro- track of moderate average growth and gradually
priate strategy as I see it. declining inflation over the next few years.
First, the central bank should be clear as to Similar bouts of financial market instability
its goals. The most fundamental goal is maximum in the nineteenth century on up to the financial
possible sustainable economic growth, which in panic of 1907 led Congress to pass the Federal
my mind motivates the dual mandate in the law Reserve Act in 1912. A fundamental responsibility
for the Federal Reserve to strive for price stability of the central bank is to contribute to orderly and
and high employment. Price stability, which is efficient functioning of financial markets. The
uniquely a central bank responsibility, contributes financial market upset of 2007 will join the his-
greatly to the goal of maximum sustainable growth. tory of upsets including those in 1970, 1984, 1987,
Price stability is not in conflict with high employ- and 1998. Each upset has different specifics but
ment but contributes to it. all of them have common characteristics, includ-
I personally believe, and have so stated on ing especially a flight to safe assets.
numerous occasions, that the inflation goal should I believe that part of the policy strategy ought
be quantified. I know that many disagree on this to be to convey as clearly as possible to the market
point. In today’s economy, I believe that a quan- what the central bank is doing and why. A policy
tified inflation goal is not critically important but strategy that is a mystery to the markets will not
quantification might be of great importance in serve the central bank well. Of course, the market
the future. I ask this question: If the Fed had had will observe what the central bank does and infer
a specific inflation goal in 1965, would that com- many aspects of the strategy from those observa-
mitment have helped to avoid the Great Inflation? tions. Nevertheless, central bank strategy always
I think the answer to the question is “yes.” If that relies in part on judgments about incoming infor-
is the correct answer, then the United States might mation, such as whether a particular data release
have avoided a very costly 15-year period of infla- has anomalous features and should be discounted.
tion, or the period might have been shorter. The strategy of a central bank should be institu-
A central bank cannot fix the level of tionalized and enduring. The strategy should not
employment or its rate of growth, or the average change just because the official roster changes. The
rate of unemployment. However, the central bank strategy should evolve as economic knowledge
can contribute to employment stability. Avoiding, improves and as economic conditions change.
or at least cushioning, recessions is an important I hope these remarks are useful. They do, in
goal. This goal should not be viewed as in conflict any event, explain something about how I have
with price stability. The most serious employment approached my responsibilities.
disaster in U.S. history was the Great Depression,
which was a consequence of monetary policy mis-
takes that led to ongoing serious deflation. Simi- REFERENCES
larly, the period of the Great Inflation saw four Poole, William. “The Fed’s Monetary Policy Rule.”
recessions in 14 years. Price stability is an essen- Federal Reserve Bank of St. Louis Review,
tial precondition for overall economic stability. January/February 2006, 88(1), pp. 1-12;
6 J A N UA RY / F E B R UA RY 2 0 0 8 F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W
10. Poole
http://research.stlouisfed.org/publications/review/
06/01/Poole.pdf.
Taylor, John B. “An Appeal for Rationality in the
Policy Activism Debate.” Federal Reserve Bank of
St. Louis Review, December 1984, 66(10), pp. 151-63;
http://research.stlouisfed.org/publications/review/
84/conf/taylor.pdf.
F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W J A N UA RY / F E B R UA RY 2 0 0 8 7
11. 8 J A N UA RY / F E B R UA RY 2 0 0 8 F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W
13. Sengupta and Aubuchon
will, for the most part, follow Yunus’s character- Microfinance” section of this paper.) The next
ization of Grameencredit. section offers a brief history of the Grameen Bank
Although the terms microcredit and micro- and a discussion of its premier innovation of
finance are often used interchangeably, it is group lending contracts; the following sections
important to recognize the distinction between describe the current state of microfinance and
the two. As mentioned before, microcredit refers provide a review of some of the common percep-
to the act of providing the loan. Microfinance, tions on microfinance. The final section outlines
on the other hand, is the act of providing these the future of microfinance, particularly in the
same borrowers with financial services, such as context of global capital markets.
savings institutions and insurance policies. In
short, microfinance encompasses the field of
microcredit. Currently, it is estimated that any- A BRIEF HISTORY OF THE
where from 1,000 to 2,500 microfinance institu- GRAMEEN BANK
tions (MFIs) serve some 67.6 million clients in
over 100 different countries.2 The story of the Grameen Bank is a suitable
Many MFIs have a dual mandate to provide point to begin a discussion of microcredit and
financial as well as social services, such as health microfinance. After obtaining a PhD in economics
care and educational services for the underprivi- in 1969 and then teaching in the United States
leged. In this sense, they are not always perceived for a few years, Muhammad Yunus returned to
as profit-maximizing financial institutions. At Bangladesh in 1972. Following its independence
the same time, the remarkable accomplishment from Pakistan in 1971 and two years of flooding,
of microfinance lies in the fact that some of the Bangladesh found itself in the grips of a terrible
successful MFIs report high rates of repayment, famine. By 1974, over 80 percent of the popula-
sometimes above 95 percent. This rate demon- tion was living in abject poverty (Yunus, 2003).
strates that lending to underprivileged borrow- Yunus, then a professor of economics at
ers—those without credit histories or the assets Chittagong University in southeast Bangladesh,
to post collateral—can be a financially sustainable became disillusioned with economics: “Nothing
venture. in the economic theories I taught reflected the life
Not surprisingly, philanthropy is not a around me. How could I go on telling my students
requirement of microfinance—not all MFIs are make believe stories in the name of economics?”
non-profit organizations. While MFIs such as (See Yunus, 2003, p. viii.) He ventured into the
Banco Sol of Bolivia operate with the intent to nearby village of Jobra to learn from the poor what
return a profit, other MFIs like the Grameen Bank causes their poverty. Yunus soon realized that it
charge below-market rates to promote social was their lack of access to credit that held them
equity.3 As will be discussed below, this distinc- in poverty. Hence, the origins of “microfinance”
emerged from this experience when Yunus lent
tion is important: As the microfinance industry
$27 of his own money to 42 women involved in
continues to grow and MFIs serve a wider client
the manufacturing of bamboo stools.4
base, the commercial viability of an MFI is often
viewed as crucial for its access to more main- 4
Yunus (2003) describes his conversation with Sufiya, a stool maker.
stream sources of finance. (We will return to this She had no money to buy the bamboo for her stools. Instead, she
and related queries in the “The Evidence of was forced to buy the raw materials and sell her stools through the
same middleman. After extracting interest on the loan that Sufiya
used to buy the bamboo that morning, the moneylender left her
2 with a profit of only 2 cents for the day. Sufiya was poor not for
Microfinance Information Exchange (MIX) lists financial profiles
and data for 973 MFIs. The high estimate of 2,500 comes from a lack of work or skills, but because she lacked the necessary credit
survey conducted by the Microcredit Summit Campaign in 2002. to break free from a moneylender. With the help of a graduate stu-
dent, Yunus surveyed Jobra and found 41 other women just like
3
The social objectives of the Grameen Bank are summarized by the Sufiya. Disillusioned by the poverty around him and questioning
16 decisions in their mission statement. The statement is available what could be done, Yunus lent $27 dollars to these 42 women
at http://grameen-info.org/bank/the16.html. and asked that he be repaid whenever they could afford it.
10 J A N UA RY / F E B R UA RY 2008 F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W
14. Sengupta and Aubuchon
Figure 1
Grameen Bank Membership
Millions of Persons
6
5
4
3
2
1
0
1976 1981 1986 1991 1996 2001
Through a series of trials and errors, Yunus the poor access to credit. The Grameen Bank has
settled on a working model and by 1983, under a challenged decades of thinking and received
special charter from the Bangladesh government, wisdom on lending to the poor. It has success-
founded the Grameen Bank as a formal and inde- fully demonstrated this in two ways: First, it has
pendent financial institution. Grameen is derived shown that poor households can benefit from
from the Bengali word gram, which means village; greater access to credit and that the provision of
grameen literally means “of the village,” an appro- credit can be an effective tool for poverty allevia-
priate name for a lending institution that requires tion. Second, it has proven that institutions do
the cooperation of the villagers. The Grameen not necessarily suffer heavy losses from lending
Bank targets the poor, with the goal of lending to the poor. An obvious question, though, is how
primarily to women. Since its inception, the the Grameen Bank succeeded where so many oth-
Grameen Bank has experienced high growth rates ers have failed. The answer, according to most
and now has more than 5.5 million members economists, lies in its unique group lending
(see Figure 1), more than 95 percent of whom are contracts, which enabled the Grameen Bank to
women.5 ensure repayment without requiring collateral
Lending to poor villagers involves a signifi- from the poor.
cant credit risk because the poor are believed to
be uncreditworthy: That is, they lack the skills The Group Lending Innovation
or the expertise needed to put the borrowed
This Grameen Bank lending model can be
funds to their best possible use. Consequently,
described as follows: Borrowers organize them-
mainstream banks have for the most part denied
selves into a group of five and present themselves
5
to the Bank. After agreeing to the Bank rules, the
Grameen Bank, annual reports (various years). Data can be viewed
at www.grameen-info.org/annualreport/commonElements/htmls/
first two members of the group receive a loan. If
index.html. the first two successfully repay their loans, then
F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W J A N UA RY / F E B R UA RY 2008 11
15. Sengupta and Aubuchon
four to six weeks later the next two are offered borrower. Market failure occurs because safe
loans; after another four to six weeks, the last per- borrowers (who are more likely to repay) have to
son is finally offered a loan. As long as all mem- subsidize risky borrowers (who are more likely
bers in the group repay their loans, the promise to default). Because the bank cannot tell a safe
of future credit is extended. If any member of the borrower from a risky one, it has to charge the
group defaults on a loan, then all members are same rate to all borrowers. The rate depends on
denied access to future credit. Furthermore, eight the mix of safe and risky borrowers in the popu-
groups of Grameen borrowers are organized into lation. When the proportion of risky borrowers
centers and repayment is collected during public is sufficiently large, the subsidy required (for the
meetings. While this ensures transparency, any lender to break even on all borrowers) is so high
borrower who defaults is visible to the entire that the lender has to charge all borrowers a sig-
village, which imposes a sense of shame. In rural nificantly high rate. If the rates are sufficiently
Bangladesh, this societal pressure is a strong dis- high, safe borrowers are unlikely to apply for a
incentive to default on the loan. Initial loans are loan, thereby adversely affecting the composition
small, generally less than $100, and require weekly of the borrower pool. In extreme cases, this could
repayments that amount to a rate of 10 percent lead to market failure—a situation in which
per annum.6 Weekly repayments give the borrow- lenders do not offer loans because only the risky
ers and lenders the added benefit of discovering types remain in the market!
problems early. Economic theory helps show how joint liabil-
Group lending—or the joint liability con- ity contracts mitigate adverse selection (Ghatak
tract—is the most celebrated lending innovation and Guinnane, 1999). Under group lending, bor-
by the Grameen Bank. Economies of scale moti- rowers choose their own groups. A direct way in
vated its first use, and Yunus later found that the which this might help is when a prospective
benefits of group lending were manifold. Under customer directly informs the bank about the
a joint liability contract, the members within the reliability of potential joiners. Perhaps a more
group (who are typically neighbors in the village) surprising result is that the lender can mitigate
can help mitigate the problems that an outside
the adverse selection problem even when cus-
lender would face. Outside lenders such as banks
tomers do not directly inform the bank but form
and government-sponsored agencies face what
themselves into like groups (peer selection). That
economists call agency costs. For example, they
is, given a joint liability clause, safe customers
cannot ensure that the borrowed money be put to
will more likely group together with other safe
its most productive use (moral hazard), cannot
verify success or failure of the proposed business customers, leaving the risky types to form groups
(costly state verification/auditing), and cannot by themselves. This “assortative matching” miti-
enforce repayment. It is not difficult to see how gates the adverse selection problem because now
peers within the group can help reduce these the risky borrowers are the ones who must bail out
costs, particularly in a situation where the prom- other risky borrowers, while the safe borrowers
ise of future credit depends on the timely repay- have to shoulder a lesser subsidy. Consequently,
ment of all members in the group. Joint liability all borrowers can be charged a lower rate, reduc-
lending thus transfers these agency costs from ing the likelihood of a market failure.
the bank onto the community of borrowers, who
can provide the same services more efficiently.
But perhaps the more difficult agency prob- CURRENT STATE OF
lem faced by lenders is that of adverse selection— MICROFINANCE
ascertaining the potential credit risk of the
Since the inception of the Grameen Bank,
6
See www.grameen-info.org/bank/GBGlance.htm. Other sources
microfinance has spread to cover five continents
put the annual rates charged by MFIs at around 30 to 60 percent. and numerous countries. The Grameen Bank has
12 J A N UA RY / F E B R UA RY 2008 F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W
16. Sengupta and Aubuchon
Figure 2
Savings by Region
953 MFIs Reporting, July 2007
61 Million Savers Africa, 8.4
S. Asia, 10.6
E. Asia Pacific, 32.6
Latin America, 6.8
Eastern Europe/C. Asia, 2.7
SOURCE: Microfinance Information Exchange Network; www.mixmarket.org.
been duplicated in Bolivia, Chile, China, Ethiopia, Bangladesh, which uses the idea that frequent
Honduras, India, Malaysia, Mali, the Philippines, small deposits will guard against the temptation
Sri Lanka, Tanzania, Thailand, the United States, of spending excess income. To keep the transac-
and Vietnam; the microfinance information tion costs of daily deposits low, SafeSave hires
exchange market (MIX) lists financial information poor workers from within the collection areas
for 973 MFIs in 105 different countries. Some (typically urban slums) to meet with clients on
MFIs have also begun to seek out public and a daily basis. By coming to the client, SafeSave
international financing, further increasing their makes it convenient for households to save; by
amount of working capital and expanding the hiring individuals from the given area, training
scope of their operations. As MFIs have become costs and wages are also kept low. With this effi-
more efficient and increased their client base, they cient model for both the bank and individuals,
have begun to expand their services through differ- SafeSave has accumulated over 7,000 clients in
ent product offerings such as micro-savings, flexi- six years.7 Not surprisingly, microfinance deposits
ble loan repayment, and insurance. We discuss (like microfinance loans) break from traditional
these three different product offerings below. commercial banking experiences. The example
At the time of their inception, many MFIs of Bank Rakyat Indonesia (BRI) suggests that the
included a compulsory savings component that poor often value higher liquidity over higher inter-
limited a borrower’s access to deposited funds. est rates on deposit products. In 1986, after a year
This promoted long-term savings, but ignored of field experiments, they offered two deposit
the fact that many poor save for the short term to products: The TABANAS product offered a 12
smooth consumption during seasonal lows of pro- percent interest rate but restricted withdrawals
duction. Figure 2 provides a look at the distribu- to twice monthly, whereas the SIMPEDES prod-
tion of voluntary MFI savings by region. As MFIs uct offered an interest rate of zero but allowed
have become better versed in the microfinance unlimited withdrawals. The SIMPEDES program
market, they have applied their innovations in saw the largest gain in popularity and to this day
lending to the collection of deposits. One of the
7
leading examples is SafeSave, located in Dhaka, See www.savesafe.org.
F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W J A N UA RY / F E B R UA RY 2008 13
17. Sengupta and Aubuchon
still offers a lower interest rate but maintains more either extend the life of the loan or pay only the
accounts than the TABANAS program.8 principle for an extended period of time. As a
The original Grameen Bank was one of the penalty, the dynamic incentives of her loan are
first MFIs that incorporated a compulsory savings reset; she cannot access larger (additional)
requirement into their lending structure. Every amounts of credit until the original loan is repaid.
client was required to make a deposit worth 5 Because her default now poses no threat to the
percent of their given loan, which was placed group promise of future credit, each member is
into a group fund with strict withdrawal rules accountable only up to their individual liabilities.
(generally no withdrawals before three years). In The third offering is the addition of insurance
2001, the Grameen Bank reviewed both its lend- to microfinance loans. The most basic insurance
ing and savings policy and reinvented itself as is debt relief for the death of a borrower, offered
Grameen II. At the heart of this change were more by many MFIs, including Grameen. Other MFIs
savings options and more flexible loans, which have begun experimenting with health insurance
act as a form of insurance. New to Grameen II is and natural disaster insurance. As with lending,
a pension fund, which allows clients with loans agency problems present a dilemma for micro-
greater than 8,000 taka ($138) to contribute at least insurance. To this end, some groups such as
50 taka ($0.86) per month. The client receives 12 FINCA Uganda require life insurance of all bor-
percent per year in compound interest, earning a rowers, including “risky” and “healthy” alike and
187 percent return after the mandatory 10-year thus avoid the adverse selection problem. Other
wait. This scheme allows Grameen II to earn more ideas include providing rain insurance to guard
money in the present and expand services, while against catastrophes. This relies on the assumption
delaying payment in the near future. that crop yields (and much of the developing
economy) are tied to seasonal rain cycles. This
Grameen II serves as a good example of a sec-
innovation eliminates the problem of moral hazard
ond innovation in microfinance: flexible loan
associated with a crop loan. By tying performance
repayment. Group lending still exists and is an
to rain cycles, a farmer has no incentive to take
integral part of the process, but Grameen II intro-
crop insurance and then fail to adequately pro-
duced a flexi-loan that allows borrowers multi-
duce a crop during a season of adequate rainfall.
ple options to repay their loan on an individual
A more recent phenomenon in microfinance
basis. Yunus (2002) stated that “group solidarity
is the emergence of foreign investment in MFIs.
is used for forward-looking joint actions for
As more and more MFIs establish positive returns,
building things for the future, rather than for the
microfinance is being seen by many professional
unpleasant task of putting unfriendly pressure on
investors as a profitable investment opportunity.
a friend.” The flexi-loan is based on the assump-
One of the most important developments for the
tion that the poor will always pay back a loan and
MFIs was the June 2007 release of Standard &
thus allows the poor to reschedule their loan
Poor’s (S&P) report on the rating methodology
during difficult periods without defaulting. If
for MFIs. By applying a common methodology,
the borrower repays as promised, then the flexi-
S&P will be able to send a stronger signal to poten-
loan operates exactly like the basic loan, using
tial investors about the quality of MFI investments.
dynamic incentives9 to increase the size of the The process of debt offerings and securitization
loan after each period. If the borrower cannot in the microfinance sector will be covered in
make her payments, she is allowed to renegotiate greater detail below.
her loan contract rather than default. She can
8
The SIMPEDES program does also use a lottery system to give
rewards, often worth 0.7 percent of deposits. More details are
MICROFINANCE AROUND THE
available at the BRI web page: www.bri.co.id/english/mikrobank-
ing/aboutmikrobanking.aspx.
WORLD
9
Dynamic incentives threaten to exclude defaulted borrowers from
As Yunus and the Grameen Bank began to
future loans. prove that microfinance is a viable method to
14 J A N UA RY / F E B R UA RY 2008 F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W
18. Sengupta and Aubuchon
Table 1
Characteristics of Select Microfinance Institutions
Enterprise
Grameen Bank, Banco Sol, Compartamos, Development Group,
Bangladesh Bolivia Mexico Washington, D.C.
Established 1983 1992 1990 1993
Membership 6,948,685 103,786 616,528 250
Average loan balance (US$) $69 $1,571 $440 $22,285**
Percent female 96.70% 46.40% 98.40% 30.00%
Group lending contracts? Yes Yes Yes No
Collateral required? No No No No
Portfolio at risk >30 days ratio 1.92% 2.91% 1.13% N/A
Return on equity 1.95%* 22.81% 57.35% N/A
Operational self-sufficiency 102.24%* 120.09% 181.22% 53%**
NOTE: *12/31/2005; **2004.
SOURCE: Data for this table come from the Microfinance Information Exchange (MIX) Network, which is a web-based platform:
www.mixmarket.org. Information was provided for the Enterprise Development Group because it is the only U.S.-based MFI that
reports data on the MIX network. Some of the information for EDG was taken from their 2003/2004 annual report, available at
www.entdevgroup.org. Comparable information is not available for the Southern Good Faith Fund, as the scope of their mission has
changed and expanded to more training-based programs. A more comprehensive summary chart exists in Morduch (1999).
alleviate poverty, their methodology and program activities. By 1992, PRODEM serviced 17,000
began to spread around the world. It is difficult clients and disbursed funds totaling $4 million
to know exactly how many MFIs there currently dollars. Constrained by the legal and financial
are, but Microfinance Information Exchange (MIX) regulations governing an NGO, the board of
estimates range from 1,000 to 2,500 serving some directors decided to expand their services and
67.6 million clients. Of these 67 million, more PRODEM became the commercial bank, Banco
than half of them come from the bottom 50 per- Solidario, later that year. Currently, Banco Sol
cent of people living below the poverty line. has 48 branches in seven cities with over 110,000
That is, some 41.6 million of the poorest people clients and a loan portfolio of more than $172
in the world have been reached by MFIs. MFIs million. As of March 31, 2007, Banco Sol reported
have expanded their operations into five differ- a past-due loans level of only 1.78 percent. An
ent continents and penetrated both rural and important distinction between Grameen and
urban markets. They have achieved success with Banco Sol is the latter’s emphasis on returning a
a variety of credit products and collection mech- profit with poverty alleviation stated only as a
anisms. Table 1 provides a comparison of several secondary goal.
groups from around the world. Banco Sol offers credit, savings, and a variety
of insurance products. Their initial loan offering
Banco Solidario (Bolivia) was based on Grameen-style joint-liability lend-
Banco Solidario originally existed as the ing, offering a maximum of $3,000 per client to
Fundacion para Promocion y el Desarrollo de la groups of three or four individuals with at least
Microempresa (PRODEM), a non-governmental one year of experience in their proposed occupa-
organization (NGO) in the mid-to-late 1980s and tion. Using dynamic incentives, the size of the
provided small capital loans to groups of three loan is gradually increased based on good repay-
or more people dedicated to entrepreneurial ment history. Annual interest rates average
F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W J A N UA RY / F E B R UA RY 2008 15
19. Sengupta and Aubuchon
between 12 and 24 percent and can be anywhere through the use of credit enhancements, allow-
from 1 to 60 months in length (120 months for a ing them to place the bond with institutional
housing loan).10 With these higher interest rates, investors. Their fifth issue to date was three times
Banco Sol does not rely on subsidies and, at the oversubscribed with 70 percent of the bond pur-
end of 2006, posted returns on equity of 22.8 chased by institutional investors. By accessing
percent. the commercial market, Compartamos has been
able to lower the cost of obtaining funds and, in
Compartamos (Mexico) turn, offer better services to their borrowers, such
as absorbing the costs of providing life insurance
Compartamos is the largest MFI in Mexico,
for all clients. Their efforts to improve operational
servicing some 630,000 clients with an active
efficiency have also created a self-sufficient MFI
loan portfolio of $285 million. Located in Mexico
that has existed without subsidies for over a
City, Compartamos is active in 26 Mexican states
decade.
throughout the country and services primarily
rural borrowers. Compartamos was founded in
Good Faith Fund (United States)
1990 and began by offering joint-liability loans
to female borrowers for income-generating activ- The Good Faith Fund was modeled after the
ities. Compartamos has only recently expanded Grameen Bank and was one of the first MFIs to
their services to allow men to borrow through be established in America. In 1986, while gover-
their solidarity group and their individual credit nor of Arkansas, Bill Clinton invited Muhammad
program; still, around 98 percent of their borrow- Yunus to visit and discuss microfinance. The
ers are female. In 1998, Compartamos formed a initial program was started as the Grameen Fund,
strategic alliance with Accion International and but the name was later changed to better reflect
transformed into a regulated financial institution, the fund’s commitment to providing loans to
called a Sociedad Financiera de Objeto Limitado micro-entrepreneurs. Loans weren’t securitized
(SFOL). In 2002, Compartamos took a unique step with collateral; rather, they were guaranteed on
for a MFI and became one of the first MFIs to issue “good faith” (Yunus, 2003, p.180).
public debt, listing themselves on the Mexican As the Good Faith Fund grew, practitioners
Stock Exchange. As an SFOL, Compartamos was and academics alike began to question the effec-
limited to only offering credit for working capital. tiveness of a pure Grameen-style program in the
In order to offer more services, such as savings United States. Much like the original Grameen
and insurance programs, Compartamos became a Bank, the Good Faith Fund has relied on innova-
commercial bank in 2006. tion and change to apply microlending to the
Compartamos was one of the first MFIs to rural economy of Arkansas. Taub (1998) argues
raise additional capital funds through the sale of that the Good Faith Fund is a successful poverty
domestic bond issuances. In 2002, Compartamos alleviation program, but that it is a poor eco-
was the first MFI in Mexico and one of the first nomic development program. In Taub’s words,
in Latin America to offer a bond sale. Because “the Good Faith Fund has never been able to
this was Standard and Poor’s first attempt at rat- deliver a meaningful volume of customers, pro-
ing a microfinance bond, they adapted their cur- vide substantial loan services to the really poor,
rent methodology and rated the bond using their or achieve anything close to institutional self-
Mexican scale and assumed local buyers. S&P was sufficiency.” He argues that important social dif-
impressed with the diversified portfolio of debt ferences arise because rural Arkansas is
and offered Compartamos an MXA+ (Mexican inherently different from rural Bangladesh and
AA) rating. Reddy and Rhyne (2006) report that that these social differences cause the group
their most recent bond was rated an MXAA lending model to fail.
Group lending failed for several reasons, but
10
Banco Sol, accessed July 27, 2007; www.bancosol.com.bo/en/ foremost was the inability of potential borrowers
intro.html. to form a group. In Bangladesh, where poverty
16 J A N UA RY / F E B R UA RY 2008 F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W
20. Sengupta and Aubuchon
rates and population density are much higher quently, moneylenders are typically perceived
than the those in the United States, potential as being exploitative, taking advantage of poor
borrowers can more readily find other entrepre- villagers who have no other recourse to loans.
neurs. However, a close network of social ties Therefore, it is not surprising that microfinance
among the poor does not exist in rural Arkansas. has been welcomed by most as an alternative to
In response to this problem, Good Faith Fund the abusive practices of village moneylenders.
personnel established a mandatory six-week However, this common perception requires a
training program for individual new members more careful study: Why don’t mainstream banks
and then created groups from the training pro- lend to the poor? In the banks’ absence, do local
grams. These newly formed groups of relative moneylenders have monopoly power? More
strangers lacked the social cohesion to enforce importantly, are these high interest rates charged
contract payments, unlike group members in rural by moneylenders welfare reducing?
Bangladesh, who often live in the same village We begin by listing the difficulties that arise
and have family/community histories together. in lending to the poor. First, early studies believed
Consequently, group lending was slowly phased that poor people often lack the resources needed
out of the Good Faith Fund. Today, the Good Faith to invest their borrowings to the most productive
Fund focuses mainly on career training through use. In short, the poor borrow mostly to finance
their Business Development Center and Asset consumption needs (Bhaduri, 1977; Aleem, 1990).
Builders program. They have also found a niche Second, even if loans could be earmarked for
in loaning larger amounts of money to small- and investment purposes, commercial banks would
medium-sized enterprises that are underserved find it difficult to lend: Lack of credit histories
by the commercial banking center. These loans and documented records on small entrepreneurs
provide the same service, but at $100,000 or more, or farmers make it difficult for the bank to assess
they can hardly be considered “micro” credit. the creditworthiness of the borrower. Finally, there
is the inability of the poor to post collateral on
the loans. This reduces the bank’s recourse to a
THE EVIDENCE ON saleable asset once the borrower defaults on the
MICROFINANCE loan. Therefore, it is not difficult to see why com-
mercial banks have avoided lending to the poor.
In this section, we review some of the impor-
On the other hand, it is believed that local
tant questions on microfinance. Our assessment
moneylenders could mitigate the problems faced
is based on numerous studies, technical surveys,
by outside banks in lending to the poor. Local
and newspaper reports on microfinance. The
moneylenders are arguably better informed of
attempt here is to be illustrative rather than pro-
borrower quality and have more effective means
vide a comprehensive review of microfinance.
of monitoring and enforcing contracts than out-
side banks. In short, because of their social ties,
Is Microfinance a Desirable Alternative
information, and location advantage, these mon-
to Informal, Exploitative Sources of eylenders are in a unique position to lend to the
Finance? poor. Some observers argue that usurious interest
The spread of microfinance and the success rates in these markets can be explained by this
of MFIs in various countries around the world “monopoly” that the local moneylenders enjoy.
prompts a question: Who served the poor before Several researchers have studied the market
the microcredit revolution? It is well known that structure of rural credit markets in developing
conventional banks, which act as creditors to most countries. Some argue that rural credit markets
entrepreneurial activity in the modern world, have are more competitive than previously imagined
largely avoided lending to the poor. Instead, credit because there is free entry for local moneylend-
to the poor has been provided mostly by local ers if not outside banks. While there is no broad
moneylenders, often at usurious rates. Conse- consensus yet, most observers believe that despite
F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W J A N UA RY / F E B R UA RY 2008 17
21. Sengupta and Aubuchon
free entry in these markets, moneylenders often charged by the moneylender will enter into these
enjoy some form of local monopoly power (in the contracts. Clearly, this situation can be improved
manner of monopolistic competition), at least in upon by offering lower rates: This would allow
the short run. more borrowers—i.e., those who expect to gener-
However, there are other reasons why money- ate a lower rate of return on their investment—to
lenders charge high interest rates. First, money- enter into loan contracts. However, this does not
lenders have to compensate for the high mean that a high interest rate per se reduces wel-
transaction costs of issuing and servicing a small fare. On the contrary, getting rid of moneylenders
loan. Second, some observers believe that these or preventing them from offering loans at these
funds have high “opportunity costs”—that is, high rates can be welfare reducing; in their
moneylenders can earn high returns by investing absence, entrepreneurs with the highest returns
in their own farms. Finally, and this is despite on their projects have no recourse to loans.
their local informational advantage, moneylenders In contrast, MFIs can often offer lower interest
face some of the same problems as commercial rates than local moneylenders because of their
banks in identifying risky borrowers and securing higher efficiency in screening and monitoring
collateral, particularly in poor rural areas. A borrowers, which results from both their economy
simple numerical example helps illustrate this of scale (serving more borrowers) and their use of
result11: Consider two lenders with the same cost joint liability lending mechanisms. This lowers
of funds. Suppose now that the first lender oper- the MFI’s cost of lending relative to that of the
ates in a prime market where borrowers faithfully local moneylender. To the extent that MFIs can
repay all of their loans at 10 percent, giving him provide loans at a lower rate than moneylenders,
an expected 10 percent return. However, the sec- enabling more and more borrowers to enter the
ond lender operates in a poor rural market where credit market, is an argument for both the effi-
borrowers arguably have a higher rate of default, ciency (because of the reduced cost of funds) and
say 50 percent.12 Consequently, her expected net welfare enhancement (because of an increase in
return is thus [ 1 + interest rate * 1 – probability the borrower pool) of microfinance.
of default – 1]. Therefore, for the second money-
lender to earn the same 10 percent return, she How High are the Repayment Rates
must charge an interest rate equal to 120 percent: for MFIs?
1 + 120% * 1 – 50% – 1 = 10%. This is not to
say that some moneylenders don’t engage in This is widely regarded as the greatest achieve-
price setting, but it does give a simple example ment of microfinance. Many MFIs report high
in which a moneylender can be competitive but rates of repayment, often greater than 90 percent.
still charge extremely high interest rates. These claims have driven considerable academic
Do moneylenders reduce welfare because interest in why and how microfinance works.
they charge high interest rates? To the extent that Furthermore, these repayment rates are widely
borrowers willingly accept these loan contracts, cited in popular media (Business Week, July 9
the answer is no.13 These loan contracts do gen- and 16, 2007; Wall Street Journal, September 23,
erate a positive surplus ex ante. That is, only those 2007) and have been one of the reasons for the
borrowers who expect to generate a rate of return recent interest generated by microfinance in finan-
from their investment that is higher than that cial markets worldwide. Although the theories of
joint liability contracts, progressive lending,14
11
This example in Armendáriz de Aghion and Morduch (2005) is frequent repayments, and flexible collateral ade-
drawn from the early work of Bottomley (1975). quately explain these high rates of repayment,
12
Of course, Yunus believes that this wrong assumption is the root Morduch (1999) raises the important issue of
of all the problems that the poor have in obtaining credit.
13 14
Bhaduri (1973) points to some degree of coercion in rural credit Progressive lending is a type of dynamic incentive in which
markets, particularly in situations where landlords double as access to larger amounts of credit becomes available after each
moneylenders. successfully repaid loan.
18 J A N UA RY / F E B R UA RY 2008 F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W
22. Sengupta and Aubuchon
validation. Because many of these repayment age interest rate on all products to 32 percent.
rates are self reported, it is important to under- Morduch is careful to point out that it is unknown
stand the methodology used to calculate these whether or not borrowers would defect, because
repayment rates. for most borrowers the alternative is either no
Morduch studies the repayment rates for the loan or an even higher interest rate on loans from
Grameen Bank for the 10-year period of 1985 to a moneylender.
1996. During this period, Grameen’s average loan Although there is an apparent disagreement
portfolio grew from $10 million to $271 million between Morduch’s adjusted rates of repayment
and membership expanded more than 12-fold to and the Grameen Bank’s self reported rates, this
include 2.06 million members in 1996. For this alone does not mean that Grameen is a financial
decade, Grameen reports an average overdue rate failure. In one case, the modest write-offs of bad
of only 1.6 percent.15 Morduch’s contention is loans offer proof of Yunus’s organizational com-
that the Grameen Bank does not follow conven- mitment to the poor and the belief that, given time,
tional accounting practices and calculates the they will repay a loan. The since-implemented
overdue rates as the value of loans overdue (for Grameen II Bank builds on this concept and
more than one year) divided by the current port- allows borrowers to restructure a loan into smaller
folio, instead of dividing by the size of the port- payments or to take a scheduled amount of time
folio when the overdue loans were issued. Because off, rather than default. Yunus describes the dif-
the size of the loan portfolio expanded 27-fold ference: “[The] overarching objective of the con-
during this 10-year period, the loan portfolio is ventional banks is to maximize profit. The
significantly larger at the end of any one year than Grameen Bank’s objective is to bring financial
at the beginning. Morduch finds the adjusted services to the poor, particularly women and the
average default rate to be 7.8 percent for the same poorest and to help them fight poverty, stay prof-
10-year period. He makes the point that “the rate itable and financially sound. It is a composite
is still impressive relative to the performance of objective, coming out of social and economic
government development banks, but it is high visions.” Given that the Grameen Bank’s focus is
enough to start creating financial difficulties” largely on social objectives and not profit maxi-
(Morduch, 1999, p. 1590). mization, some have argued that it is not obligated
As for these financial difficulties, Morduch to adopt standard accounting procedures. What
then focuses on reported profits, taking special is important is that Grameen is among the few
care to examine the provision of loan losses. He transparent microfinance organizations and
finds that the bank is slow to write off bad loans, researchers have been able to review and evaluate
dropping only a modest 3.5 percent of its portfolio their financial statements.
every year, again overstating the amount of profit. An important consideration here is that MFIs
are known to charge considerably higher rates
He calculates that instead of posting a total of
compared with similar loans from conventional
$1.5 million in profits, the bank would have
banks. In their celebrated work, Stiglitz and Weiss
instead lost a total of $18 million. The implica-
(1981) showed that the high interest rate that a
tions to Morduch’s findings are as follows: In the
lender charges may itself adversely affect repay-
early 1990s, to operate without subsidies, the
ment rates by either discouraging creditworthy
Grameen Bank would have had to raise interest
borrowers (adverse selection) or tempting the
rates on its general product from 20 percent to
borrowers to opt for riskier projects (moral hazard).
50 percent, and this would have raised the aver-
Consequently, the coexistence of high repayment
15
rates (around 95 percent) and higher interest rates
In comparison, nonperforming loans averaged between 1 and 1.5
percent for all U.S. commercial banks for the decade of 1995 to (a 30 to 60 percent interest rate is common) in
2005. (Source: Federal Financial Institutions Examination Council.) microfinance has “puzzled” economists.
Braverman and Gausch (1986) found that government credit pro-
grams in Africa, the Middle East, Latin America, South Asia, and
One explanation offered by some economists
Southeast Asia all had default rates between 40 and 95 percent. is that MFIs face an inelastic demand for loans.
F E D E R A L R E S E R V E B A N K O F S T. LO U I S R E V I E W J A N UA RY / F E B R UA RY 2008 19