Jan Feb 2008 Review

RE

VIE

WJANUARY/FE

BRUARY2008

•VOLU

ME90

NUMBER1

REVIEWJANUARY/FEBRUARY 2008

Federal Reserve Bank of St. Louis

VOLUME 90, NUMBER 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

1Thinking Like a Central Banker

William Poole

9The Microfinance Revolution:

An Overview


31A Primer on the Mortgage Market

and Mortgage Finance


47Changing Trends in the Labor Force:

A Survey

Riccardo DiCecio, Kristie M. Engemann,

Michael T. Owyang, and Christopher H. Wheeler

Director of Research

Robert H. Rasche

Deputy Director of Research

Cletus C. CoughlinReview Editor

William T. Gavin

Research Economists

Richard G. AndersonSubhayu Bandyopadhyay

James B. BullardRiccardo DiCecioMichael J. DuekerThomas A. Garrett

Carlos GarrigaMassimo Guidolin

Rubén Hernández-MurilloKevin L. Kliesen

Natalia A. KolesnikovaChristopher J. Neely

Edward NelsonMichael T. OwyangMichael R. PakkoRajdeep SenguptaDaniel L. Thornton

Howard J. WallYi Wen

Christopher H. WheelerDavid C. Wheelock

Managing Editor

George E. Fortier

Editor

Lydia H. Johnson

Graphic Designer

Donna M. Stiller

The views expressed are those of the individual authorsand do not necessarily reflect official positions of theFederal Reserve Bank of St. Louis, the Federal Reserve

System, or the Board of Governors.

REVIEW

FEDERAL RESERVE BANK OF ST. LOUIS REVIEW JANUARY/FEBRUARY 2008 i

ii JANUARY/FEBRUARY 2008 FEDERAL RESERVE BANK OF ST. LOUIS REVIEW

Review is published six times per year by the Research Division of the Federal Reserve Bank ofSt. Louis and may be accessed through our web site: research.stlouisfed.org/publications/review.All nonproprietary and nonconfidential data and programs for the articles written by FederalReserve Bank of St. Louis staff and published in Review also are available to our readers on this website. These data and programs are also available through Inter-university Consortium for Politicaland Social Research (ICPSR) via their FTP site: www.icpsr.umich.edu/pra/index.html. Or contactthe ICPSR at P.O. Box 1248, Ann Arbor, MI 48106-1248; 734-647-5000; [email protected].

Single-copy subscriptions are available free of charge. Send requests to: Federal Reserve Bank ofSt. Louis, Public Affairs Department, P.O. Box 442, St. Louis, MO 63166-0442, or call (314) 444-8809.

General data can be obtained through FRED (Federal Reserve Economic Data), a database providingU.S. economic and financial data and regional data for the Eighth Federal Reserve District. You mayaccess FRED through our web site: research.stlouisfed.org/fred.

Articles may be reprinted, reproduced, published, distributed, displayed, and transmitted in theirentirety if copyright notice, author name(s), and full citation are included. Please send a copy of anyreprinted, published, or displayed materials to George Fortier, Research Division, Federal ReserveBank of St. Louis, P.O. Box 442, St. Louis, MO 63166-0442; [email protected]. Please note:Abstracts, synopses, and other derivative works may be made only with prior written permission ofthe Federal Reserve Bank of St. Louis. Please contact the Research Division at the above address torequest permission.

© 2008, Federal Reserve Bank of St. Louis.

ISSN 0014-9187

FEDERAL RESERVE BANK OF ST. LOUIS REVIEW JANUARY/FEBRUARY 2008 1

Thinking Like a Central Banker

William Poole

This article was originally presented as a speech to Market News International in New York, NewYork, September 28, 2007.

Federal Reserve Bank of St. Louis Review, January/February 2008, 90(1), pp. 1-7.

System but also may not reflect the views of any-one else at the Fed, past or present. I thank mycolleagues at the Federal Reserve Bank of St. Louisfor their comments, but I retain full responsibilityfor errors.

ASSESSING THE ECONOMYAn area where Fed practice and market prac-

tice are essentially identical is in assessing thestate of the economy and the outlook. Privatesector and Fed forecasters use similar methodsand rely on the same statistical information.Obviously, there are professional differences ofopinion and of approaches, but these do not createa divide between Fed and private forecasts. As Ihave often put it, economists inside and outsidethe Fed studied at the same universities underthe same professors and read the same journalarticles. There is substantial movement of econo-mists into and out of the Federal Reserve System.Fed economists attend many university seminars,and academic economists attend Fed seminars.Disagreements about forecasts are similar insideand outside of the Fed.

There is a difference in the informal or anec-dotal information available inside and outsidethe Fed. The Fed has a large network of businesscontacts and relies on them to augment the fore-

Everyone looks at the world throughlenses colored by his or her own expe-riences and background. Over my nineplus years at the Fed, I have been struck

by misunderstandings of why the Fed acts as itdoes—misunderstandings from vantage pointsthat are quite different from that of a Fed official.Those with Fed experience do know things thatothers do not. Some of what we know is confi-dential, but such information is in most casesdisclosed with a lag. There are few permanentsecrets. Still, there is a central-banker way ofthinking that can be described and analyzed;doing so may help others to avoid mistakes inassessing Fed policy. That is my topic in theseremarks.

Obviously, all I can do is to describe how oneparticular central banker with the initials W.P.thinks about what he does. And my perspectiveis that from a particular central bank, the FederalReserve. My Fed colleagues might put things dif-ferently and might believe that I am off base withsome of my comments. Nevertheless, I think theeffort is worthwhile, for the degree of success ofmonetary policy is positively correlated with howcompletely the market understands the Fed. Mydisclaimer is that the views I express here aremine. These views not only do not necessarilyreflect official positions of the Federal Reserve

William Poole is the president of the Federal Reserve Bank of St. Louis. The author appreciates comments provided by his colleagues at theFederal Reserve Bank of St. Louis. The views expressed are the author’s and do not necessarily reflect official positions of the FederalReserve System.

© 2008, The Federal Reserve Bank of St. Louis. Articles may be reprinted, reproduced, published, distributed, displayed, and transmitted intheir entirety if copyright notice, author name(s), and full citation are included. Abstracts, synopses, and other derivative works may be madeonly with prior written permission of the Federal Reserve Bank of St. Louis.

casting effort. However, some private forecastershave access to data and information the Fed doesnot. Large financial firms in particular have accessto data, such as credit card activity and prospec-tive borrowing by major clients, that the Fed doesnot have. Retail firms have extremely currentinformation on sales and orders. Of course, theFed may obtain some of this information throughits business contacts, but private companies oftenmake much more systematic use of their owninternal business information than the Fed does.

Forecasters continually provide updates basedon the flow of current information, both statisti-cal and informal. In this regard, Fed and marketpractice is essentially identical.

There is, however, a difference between theFed and the market in the use of forecast informa-tion. Traders and portfolio managers base theirtrades on the current flow of information, whichneeds to be updated throughout the trading day.Fed policymakers, on the other hand, do not con-tinuously adjust the stance of policy in the sameway managers adjust portfolio holdings. For thisreason, my own practice is not to worry much asto whether I have correctly absorbed the importof each day’s, or each hour’s, data. I know thatsome information will be irrelevant to my policyposition because it will be superseded by newinformation by the time of the next FOMC meet-ing. For example, I do not need hour-by-hourinformation on security prices. When I get tothe next FOMC meeting, I’ll have the latest data,charts of how security prices have behavedsince the previous meeting, and analyses of pricebehavior over a much longer period—indeed, foras far back in time as I find helpful. Given thatthe FOMC does not adjust policy continuously,updating my forecast with every data releasewould not be an efficient use of my time.

A consequence of the fact that FOMC meet-ings occur at six-week intervals, on average, isthat when I give a speech and take questions I maynot be completely up to date on the implicationsof the latest data. In my speeches and discussionsof policy with various audiences, I try to concen-trate on longer-run issues and general principles.I emphasize that I will be studying all the dataand anecdotal information in the days leading

up to an FOMC meeting. Thus, I ordinarily donot give detailed answers to questions on theprecise implications of the latest data for theeconomic outlook. In many cases, I just haven’tstudied the implications thoroughly, although Icertainly do so by the time the FOMC next meets.

DEALING WITH RISKA private firm, especially a financial firm,

must have robust policies to address risk. To aneconomist, risk is a two-sided concept. Outcomesmay be above or below prior expectations. Thepossibility of an outcome far below expectationdeserves special attention, for such an outcomemay force a firm into bankruptcy. A financial firmmodels risk quantitatively, to the extent possible,and then examines with great care the extent towhich formal models may miss key risks, perhapsbecause they were not observed during the sam-ple period used to fit a model or because the eco-nomic environment may be changing. A centralbank has a similar task. Quantification of risks tothe economy should be taken as far as possibleand then careful thought applied to risks beyondthose that can be captured in models.

One important difference between a financialfirm and the central bank is that a firm has a muchwider array of strategies available to mitigate riskthan does a central bank. A financial firm canmake careful calculations of the extent of durationmismatch between assets and liabilities and canadjust its positions continuously to control theextent of mismatch. A financial firm can deal inmany derivatives markets to control risk. A finan-cial firm has wide latitude in choosing how muchrisk to accept.

A central bank pretty much has to acceptpolicy risks to the economy arising from theeconomy’s institutional structure and marketenvironment. Market sentiment, bullish or bear-ish, can change quickly. Analytically, the centralbank can explore implications of various possiblescenarios and can engage in special informationcollection to try to understand as quickly as pos-sible what is happening in the economy. Beyondthat, what a central bank can do is to adopt from

Poole

2 JANUARY/FEBRUARY 2008 FEDERAL RESERVE BANK OF ST. LOUIS REVIEW

time to time a somewhat asymmetric policystance in an effort to control risk, especially byguarding against particularly costly possible out-comes. When inflation risk is the dominant con-cern, policy should lean on the restrictive sideand policymakers should be more ready to tightenthan to ease policy. Conversely, when deflationand/or recession risk predominates, policy shouldbe asymmetric toward policy ease. However,there is always the danger of leaning in onedirection too far or too long; policymakers mustbe prepared to reverse course and should try notto allow the stance of policy to drift too far froma baseline approach.

It is worth emphasizing that the central bank,as the dominant player in the money market, isin a different situation than is a competitive firm.The central bank’s strategy in mitigating riskmust work through the markets and by shapingaccurate market expectations about future centralbank behavior.

The list of possible risks facing private firmsand central banks is a long one. A risk that isoften incompletely understood by those outsidemanagement is reputational risk. The issue ismuch more than simple embarrassment. Trust isan essential capital asset for a financial firm, andfor a central bank. A damaged reputation can sendcustomers fleeing to competitors. For a centralbank, a damaged reputation can lead marketparticipants to question the bank’s policy consis-tency, its motivations, and even its veracity. Forthese reasons, successful private sector firms andcentral banks both invest heavily in programsand procedures to ensure fair dealing and highethical standards. With regard to reputational risk,the issues inside and outside the central bankare essentially identical. Financial firms andcentral banks understand each other very wellon this dimension of managing risk.

ASSESSING ODDS ON FEDPOLICY ACTION

Market participants are constantly assessingthe odds on Fed policy actions at upcomingFOMC meetings. These assessments register

directly in market prices, especially in the federalfunds futures and options markets and the similarmarkets in eurodollars. There is an importantpolicy purpose for the Fed to study these marketexpectations. Understanding how the flow of newinformation affects market expectations can beuseful to policymakers. For example, suppose Iinterpret a surprise change in employment to bean anomaly in the data but I observe a large marketreaction to the data release. In that case, I wouldreexamine my interpretation, and if I still believeI am correct I might comment during the Q&Asession after a speech that my own personal takeon the data differs from the market view. Myaim would be to prompt market participants toreexamine their interpretation of the data.

Consider another example of the importanceof tracking market expectations. When I examinethe federal funds futures market, a large discrep-ancy between market expectations and my “bestguess” of the FOMC’s future actions might suggestto me the possibility of a Fed communicationsfailure. The ideal situation is one in which themarket and the Fed have read available informa-tion the same way. I am only one participant inthe FOMC process, but I try not to contribute tomarket misunderstanding of monetary policy.The market is collating information from all FOMCparticipants, paying especially close attention,of course, to the Chairman’s views.

I also follow market data carefully as part ofongoing research on how market expectationsare formed. This research, conducted with econ-omists in the St. Louis Fed’s Research Division,helps me to understand monetary policy at adeeper level. My perspective in this research isessentially the same as similar research conductedin universities and by active market participants.

OBJECTIVESPrivate firms have the goal of profit maximiza-

tion, whereas the central bank is pursuing themacroeconomic goals of price stability, employ-ment stability at a high level, and financial marketstability. The private sector and monetary policygoals are quite different, but that fact does not, in

Poole


my view, define an important difference inapproach.

Policymakers think in terms of a loss functionthat depends on departures of outcomes fromdesired outcomes. Policy goals are quantifiableand, as with profits, come with short and longhorizons. As already discussed, private firmsand central banks must understand and controlrisks to the extent possible.

Private firm and central bank governing anddisciplining processes are, of course, quite differ-ent. Nevertheless, analytical approaches to achiev-ing goals are quite similar. I do not believe thatdifferences of objectives and governing processesdefine an essential difference between the twotypes of organizations. Thus, in this respect thosein the private sector and in the central bankunderstand and relate to each other easily.

PRICE MAKERS VERSUS PRICETAKERS

What is a critically important differencebetween a central bank and a private financial firmis that the central bank, in the short run anyway,sets a policy interest rate and importantly influ-ences longer-term interest rates through effects onmarket expectations. The central bank is a pricemaker in the interbank funds market. Privatefinancial firms are essentially price takers in thatmarket and in the government securities market.

A typical trader or portfolio manager canplan security purchases and sales with little orno regard to any effects on market prices or thebehavior of other firms. Of course, this statementis not precisely true for very large portfolios, butthe difference in market impact between a centralbank and a large private portfolio is enormous.

The fact that a central bank is a price makermakes its strategy fundamentally different fromthat of a portfolio manager. To achieve policygoals, the central bank must think of its policyactions as following a predictable policy rule thatthe private sector can observe. A portfolio man-ager responds to the flow of new informationpartly as it affects probabilities of future centralbank action.

I pointed out earlier that both market partici-pants and policymakers try to understand theimplications of the flow of information for policyactions. Now I want to emphasize the importantpoint that policymakers have the task of design-ing systematic policy responses to new informa-tion. The design should advance achievementof policy goals, such as price stability. There aremany dimensions to policy design. A simpleexample is that the Federal Reserve now adjustsits target for the federal funds rate in multiples of25 basis points. That may seem a trivial example,but in the past the Fed sometimes adjusted itsfunds rate target by smaller amounts. Anotherexample is disclosure of the policy decisionpromptly after the decision. That practice startedonly in 1994 and ever since the FOMC has almostconstantly grappled with disclosure issues.

I could point to many other dimensions ofdefining a policy rule, or response function(Poole, 2005). My point is not to elaborate on thenature of the policy rule but instead to emphasizehow different that responsibility is from that of aportfolio manager. Policymakers should shapetheir policy actions by conscious decisions abouthow to guide market thinking not just in the con-text of a particular policy action but also in thefuture for policy actions in general. Put anotherway, when economic conditions recur, policyresponses to the same set of conditions shouldalso recur. If that were not the case, then policyactions could be interpreted only as random,unpredictable responses to changes in economicconditions. It simply cannot be good policy forpolicy actions to be essentially random.

The market interprets every policy actionand every policy statement in the context of pastactions and statements. What is a surprise andwhat is expected depends on past practice. Theimplication of this obvious point is that everypolicy action needs to be based on an understand-ing of how the action will be regarded in thefuture. Policy actions set precedents, and policy-makers must be careful about those precedents.Otherwise, what appears to be a policy successtoday could be the seed of a policy problem inthe future.

Poole


Modern macroeconomics emphasizes theimportance of policy predictability for good policyoutcomes (Taylor, 1984). The difference in per-spective from standard practice 30 years ago isprofound and incompletely recognized by manyjournalists and commentators. Even in the earlyGreenspan years, many thought that monetarypolicy worked by creating surprises. That perspec-tive was natural because policy surprises hadvisible effects on security prices.

Theoretical developments in macroeconomicsin the 1970s emphasized that policy surpriseswere undesirable. Efficient planning in the privatesector requires that expectations about govern-ment policies be accurate, or as accurate as theinherent uncertainty of the economic environ-ment permits. Policymakers ought not to add toinherent economic uncertainty. It is desirable that,to the maximum possible extent, the economybe characterized by an expectational equilibriumin which the market behaves as policymakersexpect and the central bank behaves as the marketexpects. There are certainly times, however, whenpolicy surprises are unavoidable.

So, much of my own thinking is driven by aneffort to help define a policy that will increasepolicy predictability over time. In my speechesand ensuing Q&A, I try to emphasize generalpolicy principles rather than the current policysituation. What is important is not the policyaction at the next FOMC meeting, which is typi-cally what people want to know, but the policyregularity that will extend across many FOMCmeetings, which is what people should want toknow.

AVOIDING POLICYDISTURBANCES

An important corollary to the task of defininga policy rule is that the central bank ought not tobe a source of random disturbances. All of us arewell aware of the potential for saying thingsinadvertently that will create market misunder-standing of likely future Fed policy actions. Or,more precisely, what needs to be understood ishow and why various possible economic condi-

tions would justify particular appropriate policyresponses. One way to avoid misinformation isto avoid providing any information. Put anotherway, if my mouth is not open, I cannot put myfoot into it.

In my view, however, it is important to try toconvey correct information. I do not believe thatI would be doing my job if fear of providing mis-information led me to provide no information.For this reason, I have maintained an active speak-ing schedule.

I do follow some general practices designedto reduce missteps. I try to schedule speeches,and certainly press interviews, for times whenthe markets are closed. That allows the market todigest what I say overnight. Another practice isthat I never predict the outcome of future FOMCmeetings. Given that I am only one participant inthose meetings and that the Chairman’s opinioncarries great weight, predicting the outcomewould be foolish. That is obvious, but what isless obvious is that I do my best to avoid beingcommittal even in my own mind about the policyimplications of recent data. Clearly, I could drawconclusions from available data that would createa certain presumption about the policy decisionor at least about my policy position. I am verycautious about drawing firm implications aboutpolicy from the data.

I emphasize that my policy position willdepend on all the information available at thetime of the FOMC meeting, on the staff analysis,and on the debate during the meeting. Thatdescription of my attitude is literally correct. Inoted earlier that I often do not focus on the dataarriving day by day because I know that new datawill supersede existing data and that I will benefitfrom my own intensive preparation before eachmeeting. I rely on the expert staff analysis pre-pared for each FOMC meeting. Given the com-plexity and dynamic nature of the issues, I findit best not to form a settled policy position wellin advance of the meeting.

Moreover, what policy purpose would beserved by my discussing publicly every twist andturn of my analysis between FOMC meetings?Market effects from doing so would not serve a

Poole


constructive policy purpose—indeed, they wouldviolate one of the important findings in macro-economics that policy should not create randomdisturbances.

BASICS OF POLICY STRATEGYI have emphasized the importance of the

central banker perspective in conveying a policystrategy. I will conclude by sketching the appro-priate strategy as I see it.

First, the central bank should be clear as toits goals. The most fundamental goal is maximumpossible sustainable economic growth, which inmy mind motivates the dual mandate in the lawfor the Federal Reserve to strive for price stabilityand high employment. Price stability, which isuniquely a central bank responsibility, contributesgreatly to the goal of maximum sustainable growth.Price stability is not in conflict with high employ-ment but contributes to it.

I personally believe, and have so stated onnumerous occasions, that the inflation goal shouldbe quantified. I know that many disagree on thispoint. In today’s economy, I believe that a quan-tified inflation goal is not critically important butquantification might be of great importance inthe future. I ask this question: If the Fed had hada specific inflation goal in 1965, would that com-mitment have helped to avoid the Great Inflation?I think the answer to the question is “yes.” If thatis the correct answer, then the United States mighthave avoided a very costly 15-year period of infla-tion, or the period might have been shorter.

A central bank cannot fix the level ofemployment or its rate of growth, or the averagerate of unemployment. However, the central bankcan contribute to employment stability. Avoiding,or at least cushioning, recessions is an importantgoal. This goal should not be viewed as in conflictwith price stability. The most serious employmentdisaster in U.S. history was the Great Depression,which was a consequence of monetary policy mis-takes that led to ongoing serious deflation. Simi-larly, the period of the Great Inflation saw fourrecessions in 14 years. Price stability is an essen-tial precondition for overall economic stability.

We have tentative signs that the financialmarkets are beginning to recover from the recentupset, but financial fragility is obviously still anissue. If the upset were to deepen in a sustainedway, it might have serious consequences foremployment stability. As of today, we just do notknow what the consequences may be. My bestguess is that the inherent resilience of the U.S.economy along with future policy actions, shouldthey be desirable, will keep the economy on atrack of moderate average growth and graduallydeclining inflation over the next few years.

Similar bouts of financial market instabilityin the nineteenth century on up to the financialpanic of 1907 led Congress to pass the FederalReserve Act in 1912. A fundamental responsibilityof the central bank is to contribute to orderly andefficient functioning of financial markets. Thefinancial market upset of 2007 will join the his-tory of upsets including those in 1970, 1984, 1987,and 1998. Each upset has different specifics butall of them have common characteristics, includ-ing especially a flight to safe assets.

I believe that part of the policy strategy oughtto be to convey as clearly as possible to the marketwhat the central bank is doing and why. A policystrategy that is a mystery to the markets will notserve the central bank well. Of course, the marketwill observe what the central bank does and infermany aspects of the strategy from those observa-tions. Nevertheless, central bank strategy alwaysrelies in part on judgments about incoming infor-mation, such as whether a particular data releasehas anomalous features and should be discounted.The strategy of a central bank should be institu-tionalized and enduring. The strategy should notchange just because the official roster changes. Thestrategy should evolve as economic knowledgeimproves and as economic conditions change.

I hope these remarks are useful. They do, inany event, explain something about how I haveapproached my responsibilities.

REFERENCESPoole, William. “The Fed’s Monetary Policy Rule.”

Federal Reserve Bank of St. Louis Review,January/February 2006, 88(1), pp. 1-12;

Poole


http://research.stlouisfed.org/publications/review/06/01/Poole.pdf

http://research.stlouisfed.org/publications/review/06/01/Poole.pdf.

Taylor, John B. “An Appeal for Rationality in thePolicy Activism Debate.” Federal Reserve Bank ofSt. Louis Review, December 1984, 66(10), pp. 151-63;http://research.stlouisfed.org/publications/review/84/conf/taylor.pdf.

Poole


http://research.stlouisfed.org/publications/review/84/conf/taylor.pdf

http://research.stlouisfed.org/publications/review/84/conf/taylor.pdf



The Microfinance Revolution: An Overview


The Nobel Prize committee awarded the 2006 Nobel Peace Prize to Muhammad Yunus and theGrameen Bank “for their efforts to create economic and social development from below.” Themicrofinance revolution has come a long way since Yunus first provided financing to the poor inBangladesh. The committee has recognized microfinance as “an important liberating force” andan “ever more important instrument in the struggle against poverty.” Although several authorshave provided comprehensive surveys of microfinance, our aim is somewhat more modest: Thisarticle is intended as a non-technical overview on the growth and development of microcreditand microfinance. (JEL I3, J41, N80)


In its broadest sense, microcredit includesthe act of providing loans of small amounts, often$100 or less, to the poor and other borrowers thathave been ignored by commercial banks; underthis definition, microcredit encompasses alllenders, including the formal participants (suchas specialized credit cooperatives set up by thegovernment for the provision of rural credit)and those of a more informal variety (such as thevillage moneylender or even loan sharks). Yunus(2007) argues that it is important to distinguishmicrocredit in all its previous forms from thespecific form of credit adopted at the GrameenBank, which he calls “Grameencredit.” Yunusargues that the “most distinctive feature ofGrameencredit is that it is not based on any col-lateral, or legally enforceable contracts. It is basedon ‘trust,’ not on legal procedures and system.”For the purposes of this article and unless men-tioned otherwise, our use of the term microcredit

In 2006, the Grameen Bank and its founderMuhammad Yunus were awarded theNobel Peace Prize for their efforts to reducepoverty in Bangladesh. By providing small

loans to the extremely poor, the Grameen Bankoffers these recipients the chance to becomeentrepreneurs and earn sufficiently high incometo break themselves free from the cycle of poverty.Yunus’s pioneering efforts have brought renewedattention to the field of microfinance as a tool toeliminate poverty; and, since 1976 when he firstlent $27 to 42 stool makers, the Grameen Bankhas grown to include more than 5.5 million mem-bers with greater than $5.2 billion in dispersedloans. As microfinance institutions continue togrow and expand, in both the developing anddeveloped world, social activists and financialinvestors alike have begun to take notice. In thisarticle we seek to explain the rise in microfinancesince its inception in the early 1980s and thevarious mechanisms that make microfinancean effective tool in reducing poverty.1 We alsoaddress the current problems facing microfinanceand areas for future growth.

1 Other, more technical surveys of microfinance include Ghatak andGuinnane (1999), Morduch (1999), and Armendáriz de Aghionand Morduch (2005).

Rajdeep Sengupta is an economist and Craig P. Aubuchon is a research associate at the Federal Reserve Bank of St. Louis.


will, for the most part, follow Yunus’s character-ization of Grameencredit.

Although the terms microcredit and micro-finance are often used interchangeably, it isimportant to recognize the distinction betweenthe two. As mentioned before, microcredit refersto the act of providing the loan. Microfinance,on the other hand, is the act of providing thesesame borrowers with financial services, such assavings institutions and insurance policies. Inshort, microfinance encompasses the field ofmicrocredit. Currently, it is estimated that any-where from 1,000 to 2,500 microfinance institu-tions (MFIs) serve some 67.6 million clients inover 100 different countries.2

Many MFIs have a dual mandate to providefinancial as well as social services, such as healthcare and educational services for the underprivi-leged. In this sense, they are not always perceivedas profit-maximizing financial institutions. Atthe same time, the remarkable accomplishmentof microfinance lies in the fact that some of thesuccessful MFIs report high rates of repayment,sometimes above 95 percent. This rate demon-strates that lending to underprivileged borrow-ers—those without credit histories or the assetsto post collateral—can be a financially sustainableventure.

Not surprisingly, philanthropy is not arequirement of microfinance—not all MFIs arenon-profit organizations. While MFIs such asBanco Sol of Bolivia operate with the intent toreturn a profit, other MFIs like the Grameen Bankcharge below-market rates to promote socialequity.3 As will be discussed below, this distinc-tion is important: As the microfinance industrycontinues to grow and MFIs serve a wider clientbase, the commercial viability of an MFI is oftenviewed as crucial for its access to more main-stream sources of finance. (We will return to thisand related queries in the “The Evidence of

Microfinance” section of this paper.) The nextsection offers a brief history of the Grameen Bankand a discussion of its premier innovation ofgroup lending contracts; the following sectionsdescribe the current state of microfinance andprovide a review of some of the common percep-tions on microfinance. The final section outlinesthe future of microfinance, particularly in thecontext of global capital markets.

A BRIEF HISTORY OF THEGRAMEEN BANK

The story of the Grameen Bank is a suitablepoint to begin a discussion of microcredit andmicrofinance. After obtaining a PhD in economicsin 1969 and then teaching in the United Statesfor a few years, Muhammad Yunus returned toBangladesh in 1972. Following its independencefrom Pakistan in 1971 and two years of flooding,Bangladesh found itself in the grips of a terriblefamine. By 1974, over 80 percent of the popula-tion was living in abject poverty (Yunus, 2003).Yunus, then a professor of economics atChittagong University in southeast Bangladesh,became disillusioned with economics: “Nothingin the economic theories I taught reflected the lifearoundme. How could I go on telling my studentsmake believe stories in the name of economics?”(See Yunus, 2003, p. viii.) He ventured into thenearby village of Jobra to learn from the poor whatcauses their poverty. Yunus soon realized that itwas their lack of access to credit that held themin poverty. Hence, the origins of “microfinance”emerged from this experience when Yunus lent$27 of his own money to 42 women involved inthe manufacturing of bamboo stools.4

Sengupta and Aubuchon


2 Microfinance Information Exchange (MIX) lists financial profilesand data for 973 MFIs. The high estimate of 2,500 comes from asurvey conducted by the Microcredit Summit Campaign in 2002.

3 The social objectives of the Grameen Bank are summarized by the16 decisions in their mission statement. The statement is availableat http://grameen-info.org/bank/the16.html.

4 Yunus (2003) describes his conversation with Sufiya, a stool maker.She had no money to buy the bamboo for her stools. Instead, shewas forced to buy the raw materials and sell her stools through thesame middleman. After extracting interest on the loan that Sufiyaused to buy the bamboo that morning, the moneylender left herwith a profit of only 2 cents for the day. Sufiya was poor not forlack of work or skills, but because she lacked the necessary creditto break free from a moneylender. With the help of a graduate stu-dent, Yunus surveyed Jobra and found 41 other women just likeSufiya. Disillusioned by the poverty around him and questioningwhat could be done, Yunus lent $27 dollars to these 42 womenand asked that he be repaid whenever they could afford it.

Through a series of trials and errors, Yunussettled on a working model and by 1983, under aspecial charter from the Bangladesh government,founded the Grameen Bank as a formal and inde-pendent financial institution. Grameen is derivedfrom the Bengali word gram, which means village;grameen literally means “of the village,” an appro-priate name for a lending institution that requiresthe cooperation of the villagers. The GrameenBank targets the poor, with the goal of lendingprimarily to women. Since its inception, theGrameen Bank has experienced high growth ratesand now has more than 5.5 million members(see Figure 1), more than 95 percent of whom arewomen.5

Lending to poor villagers involves a signifi-cant credit risk because the poor are believed tobe uncreditworthy: That is, they lack the skillsor the expertise needed to put the borrowedfunds to their best possible use. Consequently,mainstream banks have for the most part denied

the poor access to credit. The Grameen Bank haschallenged decades of thinking and receivedwisdom on lending to the poor. It has success-fully demonstrated this in two ways: First, it hasshown that poor households can benefit fromgreater access to credit and that the provision ofcredit can be an effective tool for poverty allevia-tion. Second, it has proven that institutions donot necessarily suffer heavy losses from lendingto the poor. An obvious question, though, is howthe Grameen Bank succeeded where so many oth-ers have failed. The answer, according to mosteconomists, lies in its unique group lendingcontracts, which enabled the Grameen Bank toensure repayment without requiring collateralfrom the poor.

The Group Lending Innovation

This Grameen Bank lending model can bedescribed as follows: Borrowers organize them-selves into a group of five and present themselvesto the Bank. After agreeing to the Bank rules, thefirst two members of the group receive a loan. Ifthe first two successfully repay their loans, then



5 Grameen Bank, annual reports (various years). Data can be viewedat www.grameen-info.org/annualreport/commonElements/htmls/index.html.

0

1

2

3

4

5

6

1976 1981 1986 1991 1996 2001

Millions of Persons

Figure 1

Grameen Bank Membership

four to six weeks later the next two are offeredloans; after another four to six weeks, the last per-son is finally offered a loan. As long as all mem-bers in the group repay their loans, the promiseof future credit is extended. If any member of thegroup defaults on a loan, then all members aredenied access to future credit. Furthermore, eightgroups of Grameen borrowers are organized intocenters and repayment is collected during publicmeetings. While this ensures transparency, anyborrower who defaults is visible to the entirevillage, which imposes a sense of shame. In ruralBangladesh, this societal pressure is a strong dis-incentive to default on the loan. Initial loans aresmall, generally less than $100, and require weeklyrepayments that amount to a rate of 10 percentper annum.6 Weekly repayments give the borrow-ers and lenders the added benefit of discoveringproblems early.

Group lending—or the joint liability con-tract—is the most celebrated lending innovationby the Grameen Bank. Economies of scale moti-vated its first use, and Yunus later found that thebenefits of group lending were manifold. Undera joint liability contract, the members within thegroup (who are typically neighbors in the village)can help mitigate the problems that an outsidelender would face. Outside lenders such as banksand government-sponsored agencies face whateconomists call agency costs. For example, theycannot ensure that the borrowed money be put toits most productive use (moral hazard), cannotverify success or failure of the proposed business(costly state verification/auditing), and cannotenforce repayment. It is not difficult to see howpeers within the group can help reduce thesecosts, particularly in a situation where the prom-ise of future credit depends on the timely repay-ment of all members in the group. Joint liabilitylending thus transfers these agency costs fromthe bank onto the community of borrowers, whocan provide the same services more efficiently.

But perhaps the more difficult agency prob-lem faced by lenders is that of adverse selection—ascertaining the potential credit risk of the

borrower. Market failure occurs because safeborrowers (who are more likely to repay) have tosubsidize risky borrowers (who are more likelyto default). Because the bank cannot tell a safeborrower from a risky one, it has to charge thesame rate to all borrowers. The rate depends onthe mix of safe and risky borrowers in the popu-lation. When the proportion of risky borrowersis sufficiently large, the subsidy required (for thelender to break even on all borrowers) is so highthat the lender has to charge all borrowers a sig-nificantly high rate. If the rates are sufficientlyhigh, safe borrowers are unlikely to apply for aloan, thereby adversely affecting the compositionof the borrower pool. In extreme cases, this couldlead to market failure—a situation in whichlenders do not offer loans because only the riskytypes remain in the market!

Economic theory helps show how joint liabil-ity contracts mitigate adverse selection (Ghatakand Guinnane, 1999). Under group lending, bor-rowers choose their own groups. A direct way inwhich this might help is when a prospectivecustomer directly informs the bank about thereliability of potential joiners. Perhaps a moresurprising result is that the lender can mitigatethe adverse selection problem even when cus-tomers do not directly inform the bank but formthemselves into like groups (peer selection). Thatis, given a joint liability clause, safe customerswill more likely group together with other safecustomers, leaving the risky types to form groupsby themselves. This “assortative matching” miti-gates the adverse selection problem because nowthe risky borrowers are the ones whomust bail outother risky borrowers, while the safe borrowershave to shoulder a lesser subsidy. Consequently,all borrowers can be charged a lower rate, reduc-ing the likelihood of a market failure.

CURRENT STATE OFMICROFINANCE

Since the inception of the Grameen Bank,microfinance has spread to cover five continentsand numerous countries. The Grameen Bank has

6 See www.grameen-info.org/bank/GBGlance.htm. Other sourcesput the annual rates charged by MFIs at around 30 to 60 percent.



been duplicated in Bolivia, Chile, China, Ethiopia,Honduras, India, Malaysia, Mali, the Philippines,Sri Lanka, Tanzania, Thailand, the United States,and Vietnam; the microfinance informationexchange market (MIX) lists financial informationfor 973 MFIs in 105 different countries. SomeMFIs have also begun to seek out public andinternational financing, further increasing theiramount of working capital and expanding thescope of their operations. As MFIs have becomemore efficient and increased their client base, theyhave begun to expand their services through differ-ent product offerings such as micro-savings, flexi-ble loan repayment, and insurance. We discussthese three different product offerings below.

At the time of their inception, many MFIsincluded a compulsory savings component thatlimited a borrower’s access to deposited funds.This promoted long-term savings, but ignoredthe fact that many poor save for the short term tosmooth consumption during seasonal lows of pro-duction. Figure 2 provides a look at the distribu-tion of voluntary MFI savings by region. As MFIshave become better versed in the microfinancemarket, they have applied their innovations inlending to the collection of deposits. One of theleading examples is SafeSave, located in Dhaka,

Bangladesh, which uses the idea that frequentsmall deposits will guard against the temptationof spending excess income. To keep the transac-tion costs of daily deposits low, SafeSave hirespoor workers from within the collection areas(typically urban slums) to meet with clients ona daily basis. By coming to the client, SafeSavemakes it convenient for households to save; byhiring individuals from the given area, trainingcosts and wages are also kept low. With this effi-cient model for both the bank and individuals,SafeSave has accumulated over 7,000 clients insix years.7 Not surprisingly, microfinance deposits(like microfinance loans) break from traditionalcommercial banking experiences. The exampleof Bank Rakyat Indonesia (BRI) suggests that thepoor often value higher liquidity over higher inter-est rates on deposit products. In 1986, after a yearof field experiments, they offered two depositproducts: The TABANAS product offered a 12percent interest rate but restricted withdrawalsto twice monthly, whereas the SIMPEDES prod-uct offered an interest rate of zero but allowedunlimited withdrawals. The SIMPEDES programsaw the largest gain in popularity and to this day



7 See www.savesafe.org.

Africa, 8.4

S. Asia, 10.6

Latin America, 6.8

Eastern Europe/C. Asia, 2.7

E. Asia Pacific, 32.6

953 MFIs Reporting, July 200761 Million Savers

Figure 2

Savings by Region

SOURCE: Microfinance Information Exchange Network; www.mixmarket.org.

still offers a lower interest rate but maintains moreaccounts than the TABANAS program.8

The original Grameen Bank was one of thefirst MFIs that incorporated a compulsory savingsrequirement into their lending structure. Everyclient was required to make a deposit worth 5percent of their given loan, which was placedinto a group fund with strict withdrawal rules(generally no withdrawals before three years). In2001, the Grameen Bank reviewed both its lend-ing and savings policy and reinvented itself asGrameen II. At the heart of this change were moresavings options and more flexible loans, whichact as a form of insurance. New to Grameen II isa pension fund, which allows clients with loansgreater than 8,000 taka ($138) to contribute at least50 taka ($0.86) per month. The client receives 12percent per year in compound interest, earning a187 percent return after the mandatory 10-yearwait. This scheme allows Grameen II to earn moremoney in the present and expand services, whiledelaying payment in the near future.

Grameen II serves as a good example of a sec-ond innovation in microfinance: flexible loanrepayment. Group lending still exists and is anintegral part of the process, but Grameen II intro-duced a flexi-loan that allows borrowers multi-ple options to repay their loan on an individualbasis. Yunus (2002) stated that “group solidarityis used for forward-looking joint actions forbuilding things for the future, rather than for theunpleasant task of putting unfriendly pressure ona friend.” The flexi-loan is based on the assump-tion that the poor will always pay back a loan andthus allows the poor to reschedule their loanduring difficult periods without defaulting. Ifthe borrower repays as promised, then the flexi-loan operates exactly like the basic loan, usingdynamic incentives9 to increase the size of theloan after each period. If the borrower cannotmake her payments, she is allowed to renegotiateher loan contract rather than default. She can

either extend the life of the loan or pay only theprinciple for an extended period of time. As apenalty, the dynamic incentives of her loan arereset; she cannot access larger (additional)amounts of credit until the original loan is repaid.Because her default now poses no threat to thegroup promise of future credit, each member isaccountable only up to their individual liabilities.

The third offering is the addition of insuranceto microfinance loans. The most basic insuranceis debt relief for the death of a borrower, offeredby many MFIs, including Grameen. Other MFIshave begun experimenting with health insuranceand natural disaster insurance. As with lending,agency problems present a dilemma for micro-insurance. To this end, some groups such asFINCA Uganda require life insurance of all bor-rowers, including “risky” and “healthy” alike andthus avoid the adverse selection problem. Otherideas include providing rain insurance to guardagainst catastrophes. This relies on the assumptionthat crop yields (and much of the developingeconomy) are tied to seasonal rain cycles. Thisinnovation eliminates the problem of moral hazardassociated with a crop loan. By tying performanceto rain cycles, a farmer has no incentive to takecrop insurance and then fail to adequately pro-duce a crop during a season of adequate rainfall.

A more recent phenomenon in microfinanceis the emergence of foreign investment in MFIs.As more andmoreMFIs establish positive returns,microfinance is being seen by many professionalinvestors as a profitable investment opportunity.One of the most important developments for theMFIs was the June 2007 release of Standard &Poor’s (S&P) report on the rating methodologyfor MFIs. By applying a common methodology,S&P will be able to send a stronger signal to poten-tial investors about the quality of MFI investments.The process of debt offerings and securitizationin the microfinance sector will be covered ingreater detail below.

MICROFINANCE AROUND THEWORLD

As Yunus and the Grameen Bank began toprove that microfinance is a viable method to

8 The SIMPEDES program does also use a lottery system to giverewards, often worth 0.7 percent of deposits. More details areavailable at the BRI web page: www.bri.co.id/english/mikrobank-ing/aboutmikrobanking.aspx.

9 Dynamic incentives threaten to exclude defaulted borrowers fromfuture loans.



alleviate poverty, their methodology and programbegan to spread around the world. It is difficultto know exactly how many MFIs there currentlyare, but Microfinance Information Exchange (MIX)estimates range from 1,000 to 2,500 serving some67.6 million clients. Of these 67 million, morethan half of them come from the bottom 50 per-cent of people living below the poverty line.That is, some 41.6 million of the poorest peoplein the world have been reached by MFIs. MFIshave expanded their operations into five differ-ent continents and penetrated both rural andurban markets. They have achieved success witha variety of credit products and collection mech-anisms. Table 1 provides a comparison of severalgroups from around the world.

Banco Solidario (Bolivia)

Banco Solidario originally existed as theFundacion para Promocion y el Desarrollo de laMicroempresa (PRODEM), a non-governmentalorganization (NGO) in the mid-to-late 1980s andprovided small capital loans to groups of threeor more people dedicated to entrepreneurial

activities. By 1992, PRODEM serviced 17,000clients and disbursed funds totaling $4 milliondollars. Constrained by the legal and financialregulations governing an NGO, the board ofdirectors decided to expand their services andPRODEM became the commercial bank, BancoSolidario, later that year. Currently, Banco Solhas 48 branches in seven cities with over 110,000clients and a loan portfolio of more than $172million. As of March 31, 2007, Banco Sol reporteda past-due loans level of only 1.78 percent. Animportant distinction between Grameen andBanco Sol is the latter’s emphasis on returning aprofit with poverty alleviation stated only as asecondary goal.

Banco Sol offers credit, savings, and a varietyof insurance products. Their initial loan offeringwas based on Grameen-style joint-liability lend-ing, offering a maximum of $3,000 per client togroups of three or four individuals with at leastone year of experience in their proposed occupa-tion. Using dynamic incentives, the size of theloan is gradually increased based on good repay-ment history. Annual interest rates average



Table 1Characteristics of Select Microfinance Institutions

EnterpriseGrameen 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 thatreports data on the MIX network. Some of the information for EDG was taken from their 2003/2004 annual report, available atwww.entdevgroup.org. Comparable information is not available for the Southern Good Faith Fund, as the scope of their mission haschanged and expanded to more training-based programs. A more comprehensive summary chart exists in Morduch (1999).

between 12 and 24 percent and can be anywherefrom 1 to 60 months in length (120 months for ahousing loan).10 With these higher interest rates,Banco Sol does not rely on subsidies and, at theend of 2006, posted returns on equity of 22.8percent.

Compartamos (Mexico)

Compartamos is the largest MFI in Mexico,servicing some 630,000 clients with an activeloan portfolio of $285 million. Located in MexicoCity, Compartamos is active in 26 Mexican statesthroughout the country and services primarilyrural borrowers. Compartamos was founded in1990 and began by offering joint-liability loansto female borrowers for income-generating activ-ities. Compartamos has only recently expandedtheir services to allow men to borrow throughtheir solidarity group and their individual creditprogram; still, around 98 percent of their borrow-ers are female. In 1998, Compartamos formed astrategic alliance with Accion International andtransformed into a regulated financial institution,called a Sociedad Financiera de Objeto Limitado(SFOL). In 2002, Compartamos took a unique stepfor a MFI and became one of the first MFIs to issuepublic debt, listing themselves on the MexicanStock Exchange. As an SFOL, Compartamos waslimited to only offering credit for working capital.In order to offer more services, such as savingsand insurance programs, Compartamos became acommercial bank in 2006.

Compartamos was one of the first MFIs toraise additional capital funds through the sale ofdomestic bond issuances. In 2002, Compartamoswas the first MFI in Mexico and one of the firstin Latin America to offer a bond sale. Becausethis was Standard and Poor’s first attempt at rat-ing a microfinance bond, they adapted their cur-rent methodology and rated the bond using theirMexican scale and assumed local buyers. S&P wasimpressed with the diversified portfolio of debtand offered Compartamos an MXA+ (MexicanAA) rating. Reddy and Rhyne (2006) report thattheir most recent bond was rated an MXAA

through the use of credit enhancements, allow-ing them to place the bond with institutionalinvestors. Their fifth issue to date was three timesoversubscribed with 70 percent of the bond pur-chased by institutional investors. By accessingthe commercial market, Compartamos has beenable to lower the cost of obtaining funds and, inturn, offer better services to their borrowers, suchas absorbing the costs of providing life insurancefor all clients. Their efforts to improve operationalefficiency have also created a self-sufficient MFIthat has existed without subsidies for over adecade.

Good Faith Fund (United States)

The Good Faith Fund was modeled after theGrameen Bank and was one of the first MFIs tobe established in America. In 1986, while gover-nor of Arkansas, Bill Clinton invited MuhammadYunus to visit and discuss microfinance. Theinitial program was started as the Grameen Fund,but the name was later changed to better reflectthe fund’s commitment to providing loans tomicro-entrepreneurs. Loans weren’t securitizedwith collateral; rather, they were guaranteed on“good faith” (Yunus, 2003, p.180).

As the Good Faith Fund grew, practitionersand academics alike began to question the effec-tiveness of a pure Grameen-style program in theUnited States. Much like the original GrameenBank, the Good Faith Fund has relied on innova-tion and change to apply microlending to therural economy of Arkansas. Taub (1998) arguesthat the Good Faith Fund is a successful povertyalleviation program, but that it is a poor eco-nomic development program. In Taub’s words,“the Good Faith Fund has never been able todeliver a meaningful volume of customers, pro-vide substantial loan services to the really poor,or achieve anything close to institutional self-sufficiency.” He argues that important social dif-ferences arise because rural Arkansas isinherently different from rural Bangladesh andthat these social differences cause the grouplending model to fail.

Group lending failed for several reasons, butforemost was the inability of potential borrowersto form a group. In Bangladesh, where poverty

10 Banco Sol, accessed July 27, 2007; www.bancosol.com.bo/en/intro.html.



rates and population density are much higherthan the those in the United States, potentialborrowers can more readily find other entrepre-neurs. However, a close network of social tiesamong the poor does not exist in rural Arkansas.In response to this problem, Good Faith Fundpersonnel established a mandatory six-weektraining program for individual new membersand then created groups from the training pro-grams. These newly formed groups of relativestrangers lacked the social cohesion to enforcecontract payments, unlike groupmembers in ruralBangladesh, who often live in the same villageand have family/community histories together.Consequently, group lending was slowly phasedout of the Good Faith Fund. Today, the Good FaithFund focuses mainly on career training throughtheir Business Development Center and AssetBuilders program. They have also found a nichein loaning larger amounts of money to small- andmedium-sized enterprises that are underservedby the commercial banking center. These loansprovide the same service, but at $100,000 or more,they can hardly be considered “micro” credit.

THE EVIDENCE ONMICROFINANCE

In this section, we review some of the impor-tant questions on microfinance. Our assessmentis based on numerous studies, technical surveys,and newspaper reports on microfinance. Theattempt here is to be illustrative rather than pro-vide a comprehensive review of microfinance.

Is Microfinance a Desirable Alternativeto Informal, Exploitative Sources ofFinance?

The spread of microfinance and the successof MFIs in various countries around the worldprompts a question: Who served the poor beforethe microcredit revolution? It is well known thatconventional banks, which act as creditors to mostentrepreneurial activity in themodernworld, havelargely avoided lending to the poor. Instead, creditto the poor has been provided mostly by localmoneylenders, often at usurious rates. Conse-

quently, moneylenders are typically perceivedas being exploitative, taking advantage of poorvillagers who have no other recourse to loans.Therefore, it is not surprising that microfinancehas been welcomed by most as an alternative tothe abusive practices of village moneylenders.However, this common perception requires amore careful study: Why don’t mainstream bankslend to the poor? In the banks’ absence, do localmoneylenders have monopoly power? Moreimportantly, are these high interest rates chargedby moneylenders welfare reducing?

We begin by listing the difficulties that arisein lending to the poor. First, early studies believedthat poor people often lack the resources neededto invest their borrowings to the most productiveuse. In short, the poor borrow mostly to financeconsumption needs (Bhaduri, 1977; Aleem, 1990).Second, even if loans could be earmarked forinvestment purposes, commercial banks wouldfind it difficult to lend: Lack of credit historiesand documented records on small entrepreneursor farmers make it difficult for the bank to assessthe creditworthiness of the borrower. Finally, thereis the inability of the poor to post collateral onthe loans. This reduces the bank’s recourse to asaleable asset once the borrower defaults on theloan. Therefore, it is not difficult to see why com-mercial banks have avoided lending to the poor.

On the other hand, it is believed that localmoneylenders could mitigate the problems facedby outside banks in lending to the poor. Localmoneylenders are arguably better informed ofborrower quality and have more effective meansof monitoring and enforcing contracts than out-side banks. In short, because of their social ties,information, and location advantage, these mon-eylenders are in a unique position to lend to thepoor. Some observers argue that usurious interestrates in these markets can be explained by this“monopoly” that the local moneylenders enjoy.Several researchers have studied the marketstructure of rural credit markets in developingcountries. Some argue that rural credit marketsare more competitive than previously imaginedbecause there is free entry for local moneylend-ers if not outside banks. While there is no broadconsensus yet, most observers believe that despite



free entry in these markets, moneylenders oftenenjoy some form of local monopoly power (in themanner of monopolistic competition), at least inthe short run.

However, there are other reasons why money-lenders charge high interest rates. First, money-lenders have to compensate for the hightransaction costs of issuing and servicing a smallloan. Second, some observers believe that thesefunds have high “opportunity costs”—that is,moneylenders can earn high returns by investingin their own farms. Finally, and this is despitetheir local informational advantage, moneylendersface some of the same problems as commercialbanks in identifying risky borrowers and securingcollateral, particularly in poor rural areas. Asimple numerical example helps illustrate thisresult11: Consider two lenders with the same costof funds. Suppose now that the first lender oper-ates in a prime market where borrowers faithfullyrepay all of their loans at 10 percent, giving himan expected 10 percent return. However, the sec-ond lender operates in a poor rural market whereborrowers arguably have a higher rate of default,say 50 percent.12 Consequently, her expected netreturn is thus [�1 + interest rate� * �1 – probabilityof default� – 1]. Therefore, for the second money-lender to earn the same 10 percent return, shemust charge an interest rate equal to 120 percent:�1 + 120%� * �1 – 50%� – 1 = 10%. This is not tosay that some moneylenders don’t engage inprice setting, but it does give a simple examplein which a moneylender can be competitive butstill charge extremely high interest rates.

Do moneylenders reduce welfare becausethey charge high interest rates? To the extent thatborrowers willingly accept these loan contracts,the answer is no.13 These loan contracts do gen-erate a positive surplus ex ante. That is, only thoseborrowers who expect to generate a rate of returnfrom their investment that is higher than that

charged by the moneylender will enter into thesecontracts. Clearly, this situation can be improvedupon by offering lower rates: This would allowmore borrowers—i.e., those who expect to gener-ate a lower rate of return on their investment—toenter into loan contracts. However, this does notmean that a high interest rate per se reduces wel-fare. On the contrary, getting rid of moneylendersor preventing them from offering loans at thesehigh rates can be welfare reducing; in theirabsence, entrepreneurs with the highest returnson their projects have no recourse to loans.

In contrast, MFIs can often offer lower interestrates than local moneylenders because of theirhigher efficiency in screening and monitoringborrowers, which results from both their economyof scale (serving more borrowers) and their use ofjoint liability lending mechanisms. This lowersthe MFI’s cost of lending relative to that of thelocal moneylender. To the extent that MFIs canprovide loans at a lower rate than moneylenders,enabling more and more borrowers to enter thecredit market, is an argument for both the effi-ciency (because of the reduced cost of funds) andwelfare enhancement (because of an increase inthe borrower pool) of microfinance.

How High are the Repayment Ratesfor MFIs?

This is widely regarded as the greatest achieve-ment of microfinance. Many MFIs report highrates of repayment, often greater than 90 percent.These claims have driven considerable academicinterest in why and how microfinance works.Furthermore, these repayment rates are widelycited in popular media (Business Week, July 9and 16, 2007;Wall Street Journal, September 23,2007) and have been one of the reasons for therecent interest generated by microfinance in finan-cial markets worldwide. Although the theories ofjoint liability contracts, progressive lending,14

frequent repayments, and flexible collateral ade-quately explain these high rates of repayment,Morduch (1999) raises the important issue of



14 Progressive lending is a type of dynamic incentive in whichaccess to larger amounts of credit becomes available after eachsuccessfully repaid loan.

11 This example in Armendáriz de Aghion and Morduch (2005) isdrawn from the early work of Bottomley (1975).

12 Of course, Yunus believes that this wrong assumption is the rootof all the problems that the poor have in obtaining credit.

13 Bhaduri (1973) points to some degree of coercion in rural creditmarkets, particularly in situations where landlords double asmoneylenders.

validation. Because many of these repaymentrates are self reported, it is important to under-stand the methodology used to calculate theserepayment rates.

Morduch studies the repayment rates for theGrameen Bank for the 10-year period of 1985 to1996. During this period, Grameen’s average loanportfolio grew from $10 million to $271 millionand membership expanded more than 12-fold toinclude 2.06 million members in 1996. For thisdecade, Grameen reports an average overdue rateof only 1.6 percent.15 Morduch’s contention isthat the Grameen Bank does not follow conven-tional accounting practices and calculates theoverdue rates as the value of loans overdue (formore than one year) divided by the current port-folio, instead of dividing by the size of the port-folio when the overdue loans were issued. Becausethe size of the loan portfolio expanded 27-foldduring this 10-year period, the loan portfolio issignificantly larger at the end of any one year thanat the beginning. Morduch finds the adjustedaverage default rate to be 7.8 percent for the same10-year period. He makes the point that “the rateis still impressive relative to the performance ofgovernment development banks, but it is highenough to start creating financial difficulties”(Morduch, 1999, p. 1590).

As for these financial difficulties, Morduchthen focuses on reported profits, taking specialcare to examine the provision of loan losses. Hefinds that the bank is slow to write off bad loans,dropping only a modest 3.5 percent of its portfolioevery year, again overstating the amount of profit.He calculates that instead of posting a total of$1.5 million in profits, the bank would haveinstead lost a total of $18 million. The implica-tions to Morduch’s findings are as follows: In theearly 1990s, to operate without subsidies, theGrameen Bank would have had to raise interestrates on its general product from 20 percent to50 percent, and this would have raised the aver-

age interest rate on all products to 32 percent.Morduch is careful to point out that it is unknownwhether or not borrowers would defect, becausefor most borrowers the alternative is either noloan or an even higher interest rate on loans froma moneylender.

Although there is an apparent disagreementbetween Morduch’s adjusted rates of repaymentand the Grameen Bank’s self reported rates, thisalone does not mean that Grameen is a financialfailure. In one case, the modest write-offs of badloans offer proof of Yunus’s organizational com-mitment to the poor and the belief that, given time,they will repay a loan. The since-implementedGrameen II Bank builds on this concept andallows borrowers to restructure a loan into smallerpayments or to take a scheduled amount of timeoff, rather than default. Yunus describes the dif-ference: “[The] overarching objective of the con-ventional banks is to maximize profit. TheGrameen Bank’s objective is to bring financialservices to the poor, particularly women and thepoorest and to help them fight poverty, stay prof-itable and financially sound. It is a compositeobjective, coming out of social and economicvisions.” Given that the Grameen Bank’s focus islargely on social objectives and not profit maxi-mization, some have argued that it is not obligatedto adopt standard accounting procedures. Whatis important is that Grameen is among the fewtransparent microfinance organizations andresearchers have been able to review and evaluatetheir financial statements.

An important consideration here is that MFIsare known to charge considerably higher ratescompared with similar loans from conventionalbanks. In their celebrated work, Stiglitz andWeiss(1981) showed that the high interest rate that alender charges may itself adversely affect repay-ment rates by either discouraging creditworthyborrowers (adverse selection) or tempting theborrowers to opt for riskier projects (moral hazard).Consequently, the coexistence of high repaymentrates (around 95 percent) and higher interest rates(a 30 to 60 percent interest rate is common) inmicrofinance has “puzzled” economists.

One explanation offered by some economistsis that MFIs face an inelastic demand for loans.



15 In comparison, nonperforming loans averaged between 1 and 1.5percent for all U.S. commercial banks for the decade of 1995 to2005. (Source: Federal Financial Institutions Examination Council.)Braverman and Gausch (1986) found that government credit pro-grams in Africa, the Middle East, Latin America, South Asia, andSoutheast Asia all had default rates between 40 and 95 percent.

However, in a recent empirical study on theSafeSave program in Dhaka slums, Dehejia,Montgomery, and Morduch (2005) show that theelasticity of demand for microcredit may be sig-nificantly negative even though certain groupsof borrowers (particularly the wealthier ones) donot reduce their demand when faced with higherinterest rates. However, Emran, Morshed, andStiglitz (2006) offer a more promising explanationfor this puzzle. Departing from the traditionalfocus on credit markets in studies of microfinance,the authors examine the implications of missingor imperfect labor markets for poor women indeveloping countries (the typical customers ofMFIs in Bangladesh). Emran, Morshed, andStiglitz (2006, p. 4) demonstrate “the critical roleplayed by the structure of the labor market inmaking the small-scale household-based invest-ment projects ‘credit worthy’ in the face of veryhigh interest rates, especially for the poor house-holds with little or no collaterizable assets.”

Is There More to Microfinance thanGroup Lending or Joint LiabilityContracts?

The success of microfinance in generatinghigh repayment rates led many economists toinvestigate the reasons behind this success. Themid-to-late 1990s witnessed a large increase inthe number of journal articles on group lendingcontracts, as economists sought to explain howmicrofinance “succeeded” where traditionalforms of lending had failed. Joint liability con-tracts were seen as the break from traditionallending mechanisms and economic theory wasused to readily explain how these contractshelped to improve repayment rates. The growthof the literature on group lending contracts in themid-1990s offers the impression that all MFIsoperate as such, but the reality is that MFIs use avariety of lending techniques, such as dynamicand progressive loans, frequent repayment sched-ules, and nontraditional collateral to ensure highrepayment rates among poor, underserved borrow-ers. These mechanisms were either introducedindependently or in conjunction with joint liabil-ity programs such as Grameen’s and in many cases

operate alongside group contracts. Practitionersand theorists alike have now realized that thesemechanisms can operate with individual contractsand in certain cases (e.g., in areas of low popula-tion density) offer better repayment results thangroup lending schemes.

The mechanism of progressive lending guardsagainst the borrower’s strategic default at the endof a loan cycle, because by definition she has littleor no collateral to be seized in the event of default.Instead, MFIs have offered small initial loans,with the promise of future credit for timely repay-ment. The offer of future credit serves as a power-ful incentive for a micro-entrepreneur trying togrow her business. In this scenario, a borrowerwill default only if her current income is greaterthan her future expected profits. With a small ini-tial loan for a beginning entrepreneurial venture,this is unlikely. To further increase the likelihoodof repayment, MFIs use dynamic lending, inwhich the size of the loan is gradually increasedwith each successive loan repayment. Now, theexpected future profits are almost certainlygreater than current earned income because thesize of the loan continues to grow.

Another mechanism used by MFIs is that offrequent repayments, which often begin even theweek after the loan is disbursed. By requiringsmall repayments before the business venturehas reach maturity, MFIs are essentially requiringthat borrowers have a second source of incomeand, hence, borrow against their current consump-tion. This allows MFIs to screen against high-riskborrowers from the beginning because borrowerswill be able to repay the loan even if their venturefails. Indeed, weekly repayments give the borrow-ers and lenders the added benefit of discoveringproblems early. Armendáriz de Aghion andMorduch (2005) also suggest that frequent repay-ments provide better customer service, contraryto the belief that more repayments raise the trans-action costs for the borrower by requiring moretravel to and from payment centers. Instead, fre-quent repayments help borrowers with savingsconstraints such as seasonality of income, familymembers dropping by to borrow funds, or discre-tionary spending by one or more of the familymembers. When coupled with dynamic incen-



tives, frequent loan repayments begin to resemblesavings deposits that will be paid with interest(the graduated size of the next loan). This allowsfamilies to break free of certain savings constraints(such as those noted above) because the loan ispaid each week, before the money can be spenton anything else.

The final mechanism is the requirement ofnontraditional collateral, which was introducedby banks such as Bank Rakyat Indonesia (BRI).This feature breaks from the commercial practicethat collateral submitted must have a resale valueequal to the loan. In a group lending contract,joint liability often serves as collateral, but BRIoperates on the “notional value” of an item andallows collateral to be any item that is importantto the household, regardless of market value. Thismay include the family’s sole domestic animal,such as a cow, or it may be land that is not securedby title. Neither item could be sold for much of aprofit without significant transaction costs to thebank, but both items would be even more difficultand costly for the family to do without.

Armendáriz de Aghion and Morduch (2000)offer evidence of the success of individual loansthat use progressive/dynamic incentives, frequentrepayments, and nontraditional collateral toguarantee a loan. Using data from Eastern Europeand Russia, they demonstrate that individualloans can generate repayment rates greater than90 percent (and above 95 percent in Russia). Inindustrialized settings, borrowers are more likelyto face more competition, making it more costlyto form a borrowing group. In this scenario, loanproducts will go to different entrepreneurs, withdifferent expected payoffs—hence, necessitatingdifferent loan amounts. A group contract can beinefficient because it imposes a ceiling on theloan size equal to that given to the smallest mem-ber of any potential group. They conclude bysuggesting that in areas that are relatively indus-trialized, individual loan models may performbetter than traditional group lending models.

Is Microfinance an Important Tool forPoverty Alleviation?

Microfinance started as a method to fightpoverty, and although microfinance still fulfills

this goal, several institutions have sought to makea distinction between the “marginally poor” andthe “very poor.” The broadest definition distin-guishing these two groups comes from theConsultative Group to Assist the Poorest (CGAP),which defines the poor as individuals livingbelow the poverty line and the poorest as thebottom half of the poor. TheWorld Bank estimatesthat in 2001, some 1.1 billion people had con-sumption levels below $1 and another 2.7 billionlived on less than $2 per day.16 As microfinancecontinues to grow, questions have started to focuson who is the optimal client. Should microfinancetarget the marginally poor or the extremely poor?

Morduch (1999) tries to answer this questionby considering two representative microfinanceclients, one from each poverty group describedabove. The first client belongs to a subsidizedmicrofinance program and her income is only 50percent of the poverty line. The second clientbelongs to a financially sustainable program thataccordingly charges higher interest rates. Toensure repayment of the loan at the higher rate,the second borrower is chosen to be marginallypoor, that is, with an income of 90 percent of thepoverty line. Using the widely used “squaredpoverty gap” (Foster, Greer, and Thorbecke, 1984)measure of poverty, Morduch suggests that adollar increase in income for the very poor bor-rower has a five times greater impact than thesame dollar for the marginally poor borrower.

This simple example would suggest that, interms of poverty alleviation, MFIs should focuson the poorest borrowers first, but this is notalways the case. As MFIs seek to become finan-cially independent, they find themselves servingonly the marginally poor. This is an importantdistinction between Grameen and Banco Sol ofBolivia: The latter’s emphasis is on returning aprofit, and alleviating poverty is seen only as asecondary goal. Not surprisingly, Banco Solcharges higher interest rates,17 does not rely on



16 World Bank, “Poverty Analysis”; data can be viewed athttp://web.worldbank.org.

17 Annual interest rates average between 12 and 24 percent and canbe anywhere from 1 to 60 months in length (120 months for ahousing loan). The data are from Banco Sol, accessed 7/27/07;www.bancosol.com.bo/en/intro.html.

subsidies, and at the end of 2006 posted returnson equity of 22.8 percent.18

This apparent dichotomy between financialindependence and poverty alleviation also getsto the heart of a different problem. At what pointdoes a successful MFI begin to look like a regularbank? If the MFI successfully serves poor clients,then those clients should be able to use their loansto lift themselves out of poverty. Because of thenature of progressive and dynamic loans, success-ful borrowers earn access to larger loans, helpingthem break free of poverty even faster.

The Grameen Bank has found a way to makethis dichotomy work for them and now uses theireconomy of scale to create a financially independ-ent bank without raising interest rates. In 1995,the Grameen Bank decided not to request anymore funds from donors and instead began tofund the bank from collected deposits. With morethan two decades of successful borrowers behindthem, Grameen has had a chance to build up sav-ings deposits slowly, to the point that it is nowself-sustainable, based on the amount of fundsprovided by members. In a rough sense, it is nowthe more-successful poor that are subsidizingnew clients. This is a significant step, especiallyconsidering that, from the decade of 1985 to 1996,Armendáriz de Aghion and Morduch (2005) cal-culate that Grameen accepted $175 million insubsidies, including both direct donations and“soft” donations such as soft loans, implicit sub-sidies through equity holdings, and delayed loanloss provision.

Is Microfinance Sustainable or EvenProfitable?

With all of the positive publicity surroundingmicrofinance, it may be surprising to learn thatnot all MFIs are sustainable or able to return aprofit. Despite their rapid growth and soundoperations based on strong theoretical platforms(such as using group loans, dynamic incentives,and frequent repayments), less than half of all

MFIs return a profit and most still require thehelp of donors and subsidies. A lack of financialsustainability doesn’t necessarily indicate a failingMFI, but rather raises questions about the missionand direction of that particular MFI. Even withsubsidies, many MFIs remain the most cost-effective method to alleviate poverty; and, as weargued previously, subsidies can help change theprofile of the targeted client from the poor to theextremely poor.

For an MFI to be sustainable can mean one oftwo things: The organization can be operationallysustainable or it can be financially sustainable.An MFI that is operationally sustainable raisesenough revenue to cover the cost of operatingthe business—paying loan supervisors, openingbranch offices, etc. Subsidies might still be usedto issue loans or cover defaulted loans. An insti-tution that is financially sustainable does notrequire any subsidized inputs or outside fundsto operate. Instead, it raises money through itslending operations. TheMicroBanking Bulletin(2003) surveyed 124 MFIs with a stated commit-ment to becoming financially sustainable. In theirsurvey, the Bulletin found that only 66 operationswere sustainable, a rate just slightly above 50percent. As Armendáriz de Aghion and Morduch(2005, p. 232) note, all 124 programs asked forhelp in managing their accounting standards and,hence, “in terms of financial management, [these124 programs] are thus skimmed from the creamof the crop.” Similar sustainability data do notexist for the other 2,000+ MFIs; but, without simi-larly strong commitments to financial sustain-ability, the percentage of sustainable operationsis likely to be much lower than 50 percent.

Subsidized credit is financed in a variety offorms, some of which have been discussed brieflywith the Grameen Bank example. MFIs also securefunds from donors, many of whom want to alle-viate poverty but have not seen strong returns inthe nongovernmental organization (NGO) or gov-ernment sector. For many, donations and subsidiesare intended as a method to get MFIs started. Butwithout any accountability or empirical research,it is difficult for donors to decide at what pointan MFI should forgo its dependence on outsidefunds. Lacking in this debate is a clear under-



18 MIX Market financial data are from BancoSol, accessed 8/2/07;www.mixmarket.org/en/demand/demand.show.profile.asp?token=&ett=280.

standing of how subsidies affect the supply anddemand of loans. Without subsidies, interest ratesmay rise; and, as standard demand theory sug-gests, fewer loans will be requested. Moreover,rising interest rates without subsidies may excludepoorer projects, thus raising average returns. But,they may also increase the moral hazard problem;at higher interest rates, only risky borrowersapply for a loan, thus increasing the default rateand lowering returns. Finally, it is unclear whataffect subsidized lenders have on the overallcredit supply. Do they segment the credit marketwhile serving the very poor or do they squeezeout other lenders, reducing overall efficiency forthe market?

In some instances, government institutionscollaborate with local MFIs; but, more often thannot, government organizations and MFIs are atodds with one another, despite the fact that bothshare the stated goal of reducing poverty. A primeexample of the failure of government subsidizedinitiatives in the market for microcredit is theIntegrated Rural Development Program (IRDP),which allocated credit based on social targets inrural India, giving 30 percent of credit to sociallyexcluded groups and 30 percent to women.Armendáriz de Aghion and Morduch (2005)report that between 1979 and 1989 IRDP offeredover $6 billion in subsidized credit but generatedloan repayment rates below 60 percent, with only11 percent of borrowers taking out a second loan.During the same decade, the Grameen Bank alsoaccepted subsidies in a variety of forms, but didnot change their lending model to include socialtargets. During this time, the Grameen Bank sawits membership grow to half a million members,with repayment rates above 90 percent. The expe-rience of the Grameen Bank and IRDP during thelate 1970s and early 1980s is important becauseof the similarities between regions. BothBangladesh and India are densely populated,rural, agrarian economies with high rates ofpoverty. Therefore, it is likely that the GrameenBank's comparative success during this period isindicative of a more efficient lending model ratherthan variances in their lending environment.

In sum, even if many MFIs are not financiallysustainable, the microfinance movement may

still be the best per-dollar investment for alleviat-ing poverty. Further research is needed to showwhether financial sustainability is even a desiredobjective, and future work could help understandhow different subsidy mechanisms can best bal-ance financial sustainability with the desiredsocial objectives.

Could Competition Among MFIs Leadto Better Results?

At first glance, standard economic theorysuggests that competition should improve theperformance of MFIs and lead to better serviceand lower interest rates. With such a large poorpopulation and high rates of growth, there is alsoa large market to support more MFIs. Historically,though, competition has failed to increase servicesand often decreases the rate of repayment. Whenclients have access to alternative sources of credit,MFIs lose the leverage they gain from dynamicincentives and progressive loans (i.e., future loansare contingent on repayment).

During the late 1990s, Bolivia and Banco Solexperienced a microfinance crisis. As the successof Banco Sol increased and commercial banksbegan to see the profitability in an MFI model,competition increased. General economic theorysuggests that competition is inherently good, butfor the early MFIs, competition reduced efficiencyby weakening the incentives: As credit optionsincreased for borrowers, the incentives inherentin a dynamic or progressive loan became weaker.This proved difficult for Banco Sol, whose modelrelies on group lending and dynamic incentives.The competition mainly came from Acceso FFP,a Chilean finance company that paid its employ-ees on an incentive system. Within three years,Acceso had 90,000 loans, and Banco Sol lost 11percent of its clients. Regulated MFIs in Boliviasaw their loan overdue rates increase from 2.4percent to 8.4 percent in just over two years.Because of the increased competition, Banco Solsaw its return on equity fall by 20 percentagepoints to only 9 percent in 1999 (Armendáriz deAghion and Morduch, 2005, p. 127).

In their study of 2,875 households from 192villages in Thailand, Ahlin and Townsend (2007,



p. F43) reach a similar conclusion. They note that,with increased access to credit, borrowers do notrespond to dynamic incentives. Moreover, strongsocial ties, such as the clustering of relatives in avillage, can also lower repayment rates in the samemanner of competition. In their words, “thisresult has not been seen in the previous empiricalresearch, nor focused on in the theoretical models.”

In the early years of competition in the micro-finance sector, MFIs struggled to maintain a credi-ble threat of denying future credit on default. Inrecent times, however, new regulation has helpedto promote competition in Bolivia as lendersstarted to share more information on borrowers.By law, Banco Sol and other regulated financialintermediaries are now required to report thename and national identification number ofdelinquent borrowers to the Superintendent ofBanks and Financial Institutions. This informa-tion is available to all financial intermediariesthrough both formal and informal agreements.This agreement helped to strengthen the threatof dynamic incentives, and, as a result, competi-tion among lenders has led to an increase in theirclient base.

Does Microfinance Have Any SocialImpact in Terms of FemaleEmpowerment and Education?

Any review of microfinance is incompletewithout a discussion of its impact on women.The Microcredit Summit Campaign Report (2000)lists over a thousand programs in which 75 per-cent of the clients were women. Yunus (2003)recounts the initial difficulties overcoming thesocial mores in rural Bangladesh and lending towomen in this predominantly Islamic nation.However, his efforts were rewarded and 95 per-cent of the Grameen Bank’s current clients arewomen.

This focus on women follows largely fromYunus’s conviction that lending to women has astronger impact on the welfare of the householdthan lending to men. This has been confirmedby a large volume of research on microfinance.In countries where microfinance is predominant,country-level data reveal signs of a social trans-

formation in terms of lower fertility rates andhigher literacy rates for women. Pitt andKhandker (1998) show that loans to women havea positive impact on outcomes such as children'seducation, contraceptive use, and the value ofwomen's non-land assets. Khandker (2005) findsthat borrowing by a woman has a greater impacton per capita household expenditure on bothfood and non-food items than borrowing by aman. Among other things, this also improvesnutrition, health care, and educational opportu-nities for children in these households. Smith(2002) validates this assertion using empiricaldata from Ecuador and Honduras to comparemicrofinance institutions that also offer healthservices with institutions that offer only credit.He notes that, “in both countries, health bankparticipation significantly raises subsequenthealth care over credit-only participation.” Inparticular, he found that participation in MFIsthat offer health services reduces the tendency toswitch to bottle feeding as incomes rise. He notesthat breast-feeding children under age two is akey health-enhancing behavior.

A pro-female bias in lending works well forthe MFIs. Practitioners believe that women tendto be more risk averse in their choice of invest-ment projects, more fearful of social sanctions,and less mobile (and therefore easier to monitor)than men—making it easier for MFIs to ensure ahigher rate of repayment. Various studies fromboth Asia and Latin America have shown thatthe repayment rates are significantly higher forfemale borrowers compared with their malecounterparts.

However, critics have argued that microfi-nance has done little to change the status ofwomen within the household. A much-citedpaper by Goetz and Gupta (1996) points to evi-dence that it is mostly the men of the householdand not the women borrowers who actually exer-cise control over the borrowings. Moreover, micro-finance does little to transform the status ofwomen in terms of occupational choice, mobility,and social status within the family. Therefore,microfinance hardly “empowers” women in anymeaningful sense. Although this may truly be



the case, there is no denying the fact that micro-finance has provided heretofore unrealized work-ing opportunities for women with limited skillsin traditional activities.

Can the Microfinance Experiment BeSuccessfully Replicated Anywhere inthe World?

Although the microfinance revolution hasrecorded success in most developing countriesof the world, it has achieved little success insome of the more developed nations. The mostnotable example here is the Good Faith Fund inArkansas, where microfinance has failed to deliverthe same rapid growth and poverty alleviation asit has in the developing world. This seems reason-able given the relatively smaller percentage ofthose living in poverty and the much larger safetynet afforded the poor through welfare and unem-ployment programs. As Yunus (2003, p. 189)states, “In the developed world, my greatestnemesis is the tenacity of the social welfare sys-tem…[M]any calculate the amount of welfaremoney and insurance coverage they would loseby becoming self-employed and conclude therisk is not worth the effort.” Yunus correctlyaddresses a motivating factor for the relativelyweak success of microfinance in the UnitedStates, but studies have found other reasons whymicrofinance has failed to deliver: e.g., a lack ofentrepreneur opportunities for the poor, lack ofgroup structure, and the multitude of optionsfacing the U.S. poor.

Why Did Microfinance Initiatives Fail inthe United States? In their study of U.S. micro-finance, Edgcomb, Klein, and Clark (1996) findthat micro-enterprise accounts for only 8 to 20percent of all jobs—because of the availabilityof wage jobs and public assistance. When com-pared with the 60 to 80 percent of jobs suppliedby micro-enterprise in the developing world,the pool of potential microfinance beneficiariesin the United States is substantially smaller.Schreiner and Woller (2003) make the point thatthe characteristics of the poor are different inthe two regions. In the developing world, jobsare relatively scarce and hence the unemployed

are more likely on average to include individualsthat are highly skilled or better motivated tobecome entrepreneurs. In contrast, in the UnitedStates, where poverty is much less prevalent,most individuals with the aforementioned char-acteristics can find jobs. Furthermore, the amountof small business regulation in the United Statesposes problems; a micro-entrepreneur must knowtheir proposed business but must also under-stand local and federal tax laws and regulations.To compete with much larger national markets,small business owners must further understandand excel at marketing their products in bothlocal and larger markets. The lack of highlyskilled or better-motivated workers among thepoor in the United States, combined with thehigher entry costs for successful micro-enterprise,makes successful microfinance initiatives moredifficult. Schreiner (1999) finds that, in absoluteterms, only one person in a hundred was able tomove from unemployment to self-employmentthrough micro-enterprise.

Taub (1998) offers a slightly different expla-nation: He found that the markets for the borrow-ers differed between regions. In Bangladesh, mostsmall entrepreneurs engage in goods-producingactivities that, when combined with their smalllocal markets, offers an almost immediate streamof revenue. This feature allows the Grameen Bankand others to require weekly repayments, whichis often cited as a primary reason for their highrepayment rates. In the United States, most entre-preneurs engage in service-producing activitiesbecause it is difficult to compete against theeconomies of scale in goods production and dis-tribution within the U.S. market. These servicebusinesses provide a relatively unreliable sourceof income, particularly in the early stages. Thisrisk, combined with the safety net afforded to thepoor through welfare, discourages many potentialentrepreneurs from starting a new venture. Insupport of this point, Taub found that the likelyborrower comes from a family with at least onesource of steady income, so that their new ven-ture is unlikely to substantially hurt their familyresources.

In the late 1980s, the Good Faith Fund demon-strated the difficulty of forming a cohesive group



structure to enforce joint liability loans. SchreinerandWoller (2003) offer four basic failures of groupformation in the United States. First, they suggestthat the impersonal nature of U.S. market inter-actions reduces the need for social reputationsand hence the group loses the ability to punishdelinquent borrowers. Second, the U.S. poor arediverse and hence it is difficult to find other poorpotential entrepreneurs to guarantee a group loan.In U.S. markets, there is also a limit to the poten-tial number of small-business ideas. In developingcountries, a group of borrowers may all enter thebasket-making market with success because of themuch larger local economy. The group guaranteesthe loan but also offers advice to help succeed inthe market. In the United States, the demand formicro-businesses is much smaller and diversegroups of people must start diverse business ven-tures. There is little value to the group outside ofa loan guarantee because group members don’tshare the same risk to their businesses. Third,defaults are often not enforced in group settings,as found by Hung (2003). Finally, groups oftenbreak down in the United States because the poorhave access to other forms of credit. This creditmay be more attractive because it doesn’t requirethe transaction costs of dealing with a group.

For the United States, pure Grameen-stylegroup lending schemes have failed to deliver sub-stantial results, but that is not to say they havenot benefited the poor. Rather, microfinance oper-ations in the United States have often switchedto individual lending operations that requireborrowers to attend mandatory small businesstraining programs or offer loans to attend spe-cialized schooling for particular professions. Afundamental difference is that microfinance inthe United States helps place the poor into exist-ing wage-earning jobs rather than create new jobs.The additional training substantially raises coststo the point that many U.S. MFIs are not self-sustaining, instead relying on grants and subsi-dies. Edgcomb, Klein, and Clark (1996) foundthat the average cost to make and service a loanwas $1.47 per dollar lent, with a range of costsfrom $0.67 to $2.95. Without charging usuriousinterest rates, it can be difficult to earn such asimilarly high return, particularly with the smaller

microfinance market. Taub (1998) reports thatfrom 1989 to 1992, the Good Faith Fund averagedonly 18 new loan customers per year.19 In thefollowing years, the average number of new loancustomers rose into the mid 20s, before a changein management and change in focus substantiallyreduced those numbers. With small loans, aver-aging just $1,600 per year for the first four years,it became impossible for the Good Faith Fund toeven come close to matching the combined staffsalaries of $450,000.

Due in part to these high-cost structures,Bhatt, Tang, and Painter (2002) found direct evi-dence that nearly a third of MFIs started inCalifornia in 1996 had ceased to exist by 1998.Instead of focusing on becoming self-sufficient,Schreiner (2002, p. 82) argues for more quantita-tive evaluation of MFIs. He claims that “the dirtysecret in micro-enterprise is that few evaluationsare really tests…[E]valuations were funded andconducted by people who already believed thatmicro-enterprise was worthwhile.” Schreinerthus concludes that a main goal in helping alle-viate poverty should be to evaluate the efficiencyof MFIs and, if need be, reallocate resources toother training programs that specialize in povertyalleviation, not economic development.

THE FUTURE OF MICROFINANCEThe number of MFIs has been growing

steadily, and the top 100 MFIs are increasingtheir client base at a rate of 26 percent per year.20

To fund this spectacular growth, MFIs have turnedto a variety of sources, many of which rely onfunding from local sources to guard against for-eign currency risk. MFIs are currently movinginto the international market and confrontingchallenges such as developing standard ratingmethods; guarding against foreign currency riskand country risk; and meeting the large volume



19 At the time of Taub’s study, population density in Bangladeshwas 814 per square kilometer, while the population densities ofArkansas counties served by the Good Faith Fund were only 36,9, 8, 9.1, and 10.33 per square kilometer (Jefferson, Lincoln,Desha, Chicot, and Ashley counties, respectively).

20 MIX Market analysis of top 100 MFIs; www.mixmarket.org.

requirements for an international offering. But,according to Reddy (2007) of Accion International,“Many believe that savings mobilized from localdepositors will ultimately be the largest sourceof capital for microfinance. Foreign capital pro-vides 22 percent of funding for the ‘Top 100’ MFIs,but savings is the first source of capital, represent-ing 41 percent of all assets in 2005.”21 ManyMFIshave a mandatory or suggested savings rate; and,for larger loans, MFIs will often require borrowersto deposit 5 percent of the loan back into a savingsaccount. Some, but not all, have restrictions onwhen and how that money can be accessed.

Although not the main source of funding,foreign capital still represents a significant por-tion of current funding for the top 100 MFIs. AsElizabeth Littlefield of CGAP found, U.S. invest-ment in foreign microfinance in 2006 was $4 bil-lion, which is more than double the 2004 total of$1.6 billion. This funding comes from two mainsources: international financial institutions andmicrofinance investment vehicles. To access thisforeign investment, MFIs are beginning to use newvehicles of debt-structured finance, includingcollateralized debt obligations (CDOs) andsecuritization.

To date, one of the most well-known interna-tional debt issues was structured by Blue OrchardFinance in 2004. This deal, worth $40 million,linked 90 investors with nine MFIs in LatinAmerica, Eastern Europe, and Southeast Asia.The main innovation of the Blue Orchard dealwas the introduction of a tiering system (of fivetranches) that allowed for different risk appetitesamong investors. Microfinance is also beginningto raise money in the equity market, throughorganizations such as Accion Investments, whichhas invested $12.4 million in five institutions(Reddy and Rhyne, 2006).

In 2006, the first securitized microfinancereceivables went on the market from theBangladesh Rural Advancement Committee(BRAC). BRAC is an NGO that lends money tothe extremely poor, focusing mainly on offeringwomen credit to develop their own income-

generating activities. The transaction was struc-tured by RSA Capital, CitiGroup, the NetherlandsFinancing Company, and KfW Bank of Germanyand has securitized $180 million in receivablesover a period of six years.

According to CitiGroup, 65 percent of theloans are to the extremely poor, who borrow from$50 to $100. BRAC offers three loans, based pri-marily on the land holdings of the borrower. Forthose with less than one acre of land, borrowerscan obtain from $50 to $500 at a flat 15 percentrate, payable over one year through 46 weeklyinstallments. The marginally poor, those whoown more than one acre of land and are involvedin agricultural enterprise, can qualify for loansbetween $166 and $833 with a flat 15 percentinterest rate. This product must be repaid in equalmonthly installments, with a 12- or 18-monthhorizon. Finally, BRAC offers larger loans to entre-preneurs to start their own business. These loansare monthly products (12, 18, or 24 months) witha 15 percent interest rate.22 BRAC employs adynamic lending scheme, wherein timely repay-ments guarantee future access to credit. Thismechanism is similar to a joint lending liability,except in this case borrowers are liable to theirfuture selves.

International Financing Review Asia honoredthe BRAC deal with the title of best securitiza-tion in Asia Pacific for 2006 because “one of themost impressive aspects of the transaction is theway that it deals with the sheer complexity of adynamic pool that will contain about 3.3 millionshort tenor loans for which the average outstand-ing principal is around US$95.”23 The securitywas given an AAA rating from the localBangladesh markets, with CitiGroup andNetherlands Financing Company each purchasingone-third of the certificates. The remaining one-third was split among CitiGroup Bangladesh andtwo local Bangladeshi banks.

This deal differs from the collateralizeddebt obligations that Blue Orchard Loans for



21 Data taken from MIX Market analysis of the top 100 MFIs;www.mixmarket.org.

22 See BRAC’s economic development and microfinance informationat www.brac.net/microfinance.htm.

23 CitiGroup: “Innovative BRAC Microcredit Securitization honoredin Bangladesh,” accessed 1/16/07; www.citigroup.com/citigroup/press/2007/070116b.htm.

Development issued in April 2006, in whichfunding for 21 MFIs from 12 countries was pack-aged into a $99.1 million commercial investment.The main difference between a CDO and securi-tization is that a CDO relies on the ability of theMFI to repay the loan, unlike a securitized loanthat relies on the underlying borrowers to repay.A CDO is another vehicle to bring mainstreaminvestors to microfinance, but is still limited bythe ability to rate the creditworthiness of differ-ing MFIs. To help with this issue, S&P released arating methodology for microfinance in June 2007.By applying a common methodology, S&P willbe able to send a stronger signal to potentialinvestors about the quality of MFI investments.It is unclear yet whether the 2007 subprime mort-gage meltdown in the United States will have aneffect on investors’ risk appetites for more collat-eralized securities and whether microfinancesecurities will be viewed as “subprime” loans.

Walter and Krauss (2006) argue that theopposite should be true—namely, that microfi-nance can reduce portfolio volatility—and theirempirical tests show that microfinance institu-tions have a low correlation to general marketmovements. They suggest that this phenomenonis brought on by the continuous and diverse fund-ing through international donor agencies andbecause micro-entrepreneurs may be less inte-grated into the formal economy. When marketsenter a downturn, micro-entrepreneurs may expe-rience a countercyclical effect, as consumersshift their consumption downward to cheapergoods.

Outside of international credit markets,microfinance has continued to receive grassrootssupport and popular media coverage. Organiza-tions such as Kiva.org serve as intermediariesand connect individual donors with micro-entrepreneurs. Kiva.org allows individuals tochoose a business, originate their ownmicro-loan,and in return receive electronic journal updatesand payments from their borrower. Most loans aresmall, between $50 and $100 and have repaymentterms from six months to a year, but the lenderdoes not receive any interest on their loan. Rather,journal updates and progress reports serve asinterest, letting lenders know that their money

has been put to good use. At the end of the year,providers can start the cycle anew or withdraw.To date, 128,547 individuals have lent over $12million with a self-reported repayment rate greaterthan 99 percent. Popular media outlets such astheWall Street Journal (September 23, 2007,August 21, 2007, October 21, 2006), New YorkTimes (March 27, 2007, December 10, 2006),National Public Radio (September 7, 2007,June 19, 2007, April 6, 2007), and others havegiven Kiva.org frequent and broad exposure, mak-ing the microfinance movement as accessible tolenders as the Grameen Bank made microcreditaccessible to borrowers.

CONCLUSIONWith the recognition of the Nobel Peace Prize

in 2006, Muhammad Yunus’s vision of extendingcredit to the poor has reached a global level.Microfinance is not a panacea for poverty allevi-ation; but, with committed practitioners, a wealthof theoretical work, and a surging demand for bothinternational and individual investment, micro-finance is a poverty-alleviation tool that hasproven to be both effective and adaptable. Throughinnovations in group lending and dynamic incen-tives, MFIs have been able to successfully lend tothose traditionally ignored by commercial banks,because of their lack of collateral and credit scores.The poor have responded in kind, by repayingtheir loans with significant repayment rates. AsMFIs have grown and reached new clients, theyhave continued to innovate by offering individualloans, savings options, and life insurance andseeking new forms of capital in domestic andinternational markets. Microfinance has spreadto five continents and hundreds of countries, yetits success in U.S. markets has been ill-defined,as lenders struggle with higher transaction costsof offering loans and starting micro-enterprises.As more and more MFIs become self-sufficientand continue to expand their client base, it willbe the duty of all parties concerned with povertyrelief to look for other ways to innovate. For now,microfinance remains a viable solution to eco-nomic development and poverty alleviation, both



in Bangladesh and around the world. With moretransparency from institutions and better ratingstandards, the influx of investment capital frominternational markets will continue to drivemicrofinance toward Yunus’s goal of a poverty-free world.

REFERENCESAhlin, Christian and Townsend, Robert M. “UsingRepayment Data to Test Across Models of JointLiability Lending.” Economic Journal, February2007, 117(517), pp. F11-51.

Aleem, Irfan. “Imperfect Information, Screening, andthe Costs of Informal Lending: A Study of a RuralCredit Market in Pakistan.” World Bank EconomicReview, September 1990, 4(3), pp. 329-49.

Armendáriz de Aghion, Beatriz and Morduch,Jonathon. “Microfinance: Beyond Group Lending.”Economics of Transition, July 2000, 8(2), pp. 401-20.

Armendáriz de Aghion, Beatriz and Morduch,Jonathon. The Economics of Microfinance.Cambridge, MA: MIT Press, 2005.

Bhaduri, Amit. “A Study in Agricultural BackwardnessUnder Semi-feudalism.” Economic Journal, March1973, 83(329), pp. 120-37.

Bhaduri, Amit. “On the Formation of UsuriousInterest Rates in Backward Agriculture.”Cambridge Journal of Economics, December 1977,1(4), pp. 341-52.

Bhatt, Nitan, Tang, Shui Yan and Painter, Gary.“Microcredit Programs in the United States: TheChallenges of Outreach and Sustainability,” inJ. Carr and Z.–Y. Tong, eds., ReplicatingMicrofinance in the United States. Washington, DC:Woodrow Wilson Center Press, 2002.

Bottomley, Anthony. “Interest Rate Determination inUnderdeveloped Rural Areas.” American Journalof Agricultural Economics, May 1975, 57(2),pp. 279-91.

Braverman, Avishay and Guasch, J. Luis. “Rural



Credit Markets and Institutions in DevelopingCountries: Lessons for Policy Analysis fromPractice and Modern Theory.” World Development,October/ November 1986, 14(10/11), pp. 1253-67.

Dehejia, Rajeev; Montgomery, Heather and Morduch,Jonathan. “Do Interest Rates Matter? Credit Demandin the Dhaka Slums.” Unpublished manuscript,March 2005.

Edgcomb, Elaine; Klein, Joyce and Clark, Peggy. “ThePractice of Microenterprise in the US: Strategies,Costs, and Effectiveness.” Washington, DC: AspenInstitute, 1996.

Emran, M. Shahe; Morshed, A.K.M. Mahbub andStiglitz, Joseph E. “Microfinance and MissingMarkets.” Unpublished manuscript, October 2006.

Foster, J.; Greer, J. and Thorbecke E. “A Class ofDecomposable Poverty Measures.” Econometrica,1984, 52, pp. 761-66.

Ghatak, Maitreesh. “Screening by the Company YouKeep: Joint Liability Lending and the Peer SelectionEffect.” Economic Journal, July 2000, 110(465),pp. 601-31.

Ghatak, Maitreesh and Guinnane, Timothy. “TheEconomics of Lending with Joint Liability: Theoryand Practice.” Journal of Development Economics,October 1999, 60(1), pp. 195-228.

Goetz, Anne Marie and Gupta, Rina Sen. “Who Takesthe Credit? Gender, Power, and Control Over LoanUse in Rural Credit Programs in Bangladesh.” WorldDevelopment, January 1996, 24(1), pp. 45-63.

Hung, Chi-kan Richard. “Loan Performance of Group-Based Microcredit Programs in the United States.”Economic Development Quarterly, November 2003,17(4), pp. 382-95.

Khandker, Shahidur R. “Microfinance and Poverty:Evidence Using Panel Data from Bangladesh.”World Bank Economic Review, September 2005,19(2), pp. 263-86.

MicroBanking Bulletin. “Focus on Savings,” July2003, No. 9, pp. 72-76; www.mixmbb.org.

Microcredit Summit Campaign Report 2000, 2000;www.microcreditsummit.org/campaigns/report00html#overview.

Morduch, Jonathan. “The Microfinance Promise.”Journal of Economic Literature, December 1999,37(4), pp. 1569-614.

Pitt, Mark M. and Khandker, Shahidur R. “TheImpact of Group-Based Credit Programs on PoorHouseholds in Bangladesh: Does the Gender ofParticipants Matter?” Journal of Political Economy,October 1998, 106(5), pp. 958-96.

Reddy, Rekha M. “Microfinance Cracking the CapitalMarkets II.” InSight, May 2007, 22, pp. 1-17.

Reddy, Rekha M. and Rhyne, Elisabeth. “Who WillBuy Our Paper: Microfinance Cracking the CapitalMarkets?” Insight, April 2006, 18, pp. 1-19.

Rhyne, Elisabeth. Mainstreaming Microfinance: HowLending to the Poor Began, Grew, and Came of Agein Bolivia. Bloomfield, CT: Kumarian Press, 2001.

Schreiner, Mark. “Lessons for MicroenterprisePrograms from a Fresh Look at the UnemploymentInsurance Self-Employment Demonstration.”Evaluation Review, October 1999, 23(5), pp. 503-26.

Schreiner, Mark. “Evaluation and MicroenterprisePrograms in the United States.” Journal ofMicrofinance, Fall 2002, 4(2), pp. 67-91.

Schreiner, Mark and Woller, Gary. “MicroenterpriseDevelopment Programs in the United States and inthe Developing World.” World Development,September 2003, 31(9), pp. 1567-80.



Smith, Stephen C. “Village Banking and Maternaland Child Health: Evidence from Ecuador andHonduras.” World Development, April 2002, 30(4),pp. 707-23.

Stiglitz, Joseph and Weiss, Andrew. “Credit Rationingin Markets with Imperfect Information.” AmericanEconomic Review, June 1981, 71(3), pp. 393-410.

Taub, Richard P. “Making the Adaptation AcrossCultures and Societies: A Report on an Attempt toClone the Grameen Bank in Southern Arkansas.”Journal of Developmental Entrepreneurship,Summer 1998, 3(1), pp. 353-69.

Walter, Ingo and Krauss, Nicolas A. “Can MicrofinanceReduce Portfolio Volatility?” Working paper,November 9, 2006; http://ssrn.com/abstract=943786.

Yunus, Muhammad. “Grameen Bank II: Designed toOpen New Possibilities.” Grameen Bank, October2002; www.grameen-info.org/bank/bank2.html.

Yunus, Muhammad. Banker to the Poor: Micro-Lending and the Battle Against World Poverty.New York: Public Affairs, 2003.

Yunus, Muhammad. “What Is Microcredit?”Grameen Bank, September 2007;www.grameen-info.org/bank/WhatIsMicrocredit.htm.


A Primer on the Mortgage Marketand Mortgage Finance


This article is a primer on mortgage finance. It discusses the basics of the mortgage market andmortgage finance. In so doing, it provides useful information that can aid individuals in makingbetter mortgage finance decisions. The discussion and the tools are presented within the contextof mortgage finance; however, these same principles and tools can be applied to a wide range offinancial decisions. (JEL G0, G1)


rates on mortgage-backed securities rose, whilerates on risk-free Treasury bills declineddramatically.2

Against this backdrop, this article serves asa primer on mortgage finance. It discusses thebasics of the mortgage market and mortgagefinance, providing useful information that canaid individuals in making better mortgage financedecisions. Although the discussion and the toolsare presented within the context of mortgagefinance, these same principles and tools can beapplied to a wide range of financial decisions.

ETYMOLOGYThe term mortgage comes from the Old

French, and literally means “death vow.” Thisrefers not to the death of the borrower, but to the“death” of the loan. This is because mortgages,like many other types of loans, have a fixed termto maturity—that is, a date at which the loan isto be fully repaid. Today, mortgages are paid in

The United States was in the midst of aresidential real estate boom from 1996to 2005, and the U.S. Census Bureaureports for that period show that home-

ownership—the percentage of home-owninghouseholds—increased from 65.4 percent to 68.9percent. During this decade, the Standard &Poor/Case-Shiller Home Price Index rose at acompounded annual rate of 8.5 percent per year,more than four times faster than the rate of infla-tion. Growth in home prices was particularlystrong during the period 2000-05, when homeprices rose at an annual rate of 11.4 percent.However, since the first quarter of 2006, houseprice growth has slowed dramatically; and, inthe first quarter of 2007, prices fell for the firsttime since 1991. These price declines, combinedwith higher interest rates, have led to increasedmortgage delinquency, especially in the subprimemortgage market. Federal Reserve ChairmanBernanke reported recently that the “rate ofserious delinquencies for subprime mortgageswith adjustable interest rates…has risen to about12 percent, roughly double the recent low seenin mid-2005.”1 On news that the subprime woesmay spill over to borrowers with good credit,

1 Bernanke (2007).

2 For a discussion of the development of the subprime mortgagemarket, see Chomsisengphet and Pennington-Cross (2006).

Daniel J. McDonald is a research analyst and Daniel L. Thornton is a vice president and economic adviser at the Federal Reserve Bank ofSt. Louis.


installments (most often, monthly), so that theloan is repaid over time rather than as a lump sumon the maturity date. The word for this repaymentis amortization, which derives from the MiddleEnglish for “kill.” It refers not to the borrower’smurder, but to “killing off” the mortgage by pay-ing it down over time. The morbid etymology ofthese real estate terms must have some subliminalimpact on potential borrowers, as many continueto find the process of getting a mortgage unnerv-ing; however, a mortgage is nothing to be afraidof, as we hope to demonstrate in the remainderof this article.

MORTGAGE BASICS“Mortgage” is nothing more than the name

given to a particular type of loan; in this case, areal estate loan.3 Like any other loan, it is really anIOU—that is, a promise to repay a sum of moneyreceived today at some future time. Although thenames of loans change for a variety of reasons,they all have the same basic characteristics: theloan amount, the loan term, the schedule forrepayment, and the contract interest rate.

The amount of a loan is just that—a sum ofmoney that the borrower receives upon signingthe loan agreement. The term (or maturity) of theloan is the length of time over which the loanamount is to be repaid. The schedule for repay-ment simply states how the loan is to be repaid.Loans can be repaid in installments over the termof the mortgage, in a lump sum at the terminaldate of the contract, or in some combination ofinstallments and a final lump sum payment. Inthe case of mortgages, auto loans, and other con-sumer loans, the convention is that the loan isrepaid in fixed periodic payments, typicallymonthly. The contract interest rate is the interestrate that the borrower pays the lender in exchangefor having the money today.

There are two risks associated with lending.The first, called default risk, is the possibility

that the borrower fails to repay the loan. The sec-ond, calledmarket risk, arises when interest rateschange over time. If market interest rates riseafter the lender has offered a mortgage contract,not only will the lender earn less interest thanhe would have had he waited and lent at thehigher interest rate, but the market value of theinvestment will decline. Of course, the reverse isalso true: If market interest rates fall, the lenderwill earn more interest than if he waited and themarket value of his investment will increase.The risk is due to the fact that it is very difficultto predict whether interest rates will rise or fall.The lender also risks losing the higher interesthe would earn if the individual decides to refi-nance the loan at a lower rate.

The prospect of default has led societies todevelop laws and mechanisms to protect thelender. One of these is collateral—an asset ownedby the borrower that becomes the lender’s in theevent that the borrower fails to repay the loan. Inthe case of mortgages, the collateral is nearlyalways the property being purchased. Loan agree-ments may also contain a variety of restrictions.Some of these are intended to protect the lender,while others protect the borrower. For example,in the past, many mortgages were “assumable,”meaning that if the borrower sold the house, themortgage could be assumed or transferred to thenew owner. This hurt lenders when interest ratesrose because the new owner could get a “below-market interest rate” by assuming the previousmortgage. Today, mortgages are typically notassumable. There was also a time when manymortgages (and other consumer loan contracts)had a prepayment penalty. That is, the lendercould assess a fee if the borrower repaid the loanbefore the terminal date of the contract. Present-day mortgage contracts typically stipulate thatthere is no penalty for paying the loan off beforeits maturity date.

Types of Mortgages

There are a number of different types of mort-gages, but the most common are the fixed-ratemortgage and the adjustable-rate mortgage (orARM). Other types tend to be combinations of

McDonald and Thornton


3 Legally speaking, the loan takes the form of a note and the mort-gage per se is the agreement that secures the note by pledgingthe real estate as collateral. It is commonplace to refer to both thenote and mortgage agreement that secures the note as the“mortgage.”

these two. Fixed-rate mortgages are by far themost common type of mortgage, accounting forabout 70 percent of the total mortgage market.Figure 1 shows the percentage of the total mort-gage market accounted for by 15- and 30-yearfixed-rate mortgages since 1990 as well as theaverage contract interest rate. One would expectthat lower contract interest rates would lead to ahigher percentage of fixed-rate mortgages, as bor-rowers try to lock in low rates. This relationshipseems to hold true over most of the period, butbreaks down after 2002. The benefits of a fixed-rate mortgage are as follows: (i) the monthly pay-ment (interest and principal) is constant for theterm of the mortgage and (ii), regardless of thebehavior of market interest rates, the interest ratepaid by the borrower is the same for the life ofthe loan.

ARMs, however, have interest rates that varyover the term of the loan in step with some index.The two most common indices are the EleventhFederal Home Loan Bank Board District Cost ofFunds Index (COFI) and the National Cost ofFunds Index. ARMs have various features depend-ing on the mortgage broker. Most often, an intro-ductory rate is fixed for a period of time rangingfrom 2 to 5 years. Following this period, the

interest rate will rise or fall with the index (plusa fixed markup called themargin) at some speci-fied time interval, generally every six months.Typically, the amount that the interest rate canrise or fall in a particular interval is limited andupper and lower bounds for the interest rateover the life of the loan are set.

Rates on ARMs are lower than on otherwiseequivalent fixed-rate mortgages. The reason isthat the borrower is bearing some of the marketrisk. Market risk arises because of the inverse (ornegative) relationship between interest rates andbond prices. Specifically, if the market interestrate rises, the value of the bond (mortgage) fallsand vice versa. For example, consider the effectof an increase in the market interest rate on themarket value of a 30-year, $200,000, 5 percentfixed-rate mortgage. The price of the 30-year mort-gage decreases by $20,925.31 (from $200,000 to$179,074.69) if the market interest rate rises from5 percent to 6 percent. If the holder of the mort-gage were to sell it, they would suffer what isreferred to as a capital loss. Moreover, the priceof a longer-term mortgage falls by more than theprice of a shorter-term mortgage for a givenincrease in market interest rates. For example, the



0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

Jan90

Jan91

Jan92

Jan93

Jan94

Jan95

Jan96

Jan97

Jan98

Jan99

Jan00

Jan01

Jan02

Jan03

Jan04

Jan05

Jan06

Jan07

0.00

2.00

4.00

6.00

8.00

10.00

12.00

Fixed-Rate Share (left scale)

Contract Rate (right scale)

Figure 1

Market Share of Fixed-Rate Mortgages and Contract Interest Rate

price of a 5-year mortgage would have decreasedby just $4,774.97 (from $200,000 to $195,225.03)with the same increase in the interest rate (from5 percent to 6 percent). Because mortgages havematurities that are relatively long—up to 30 years,they have a relatively high degree of market risk.

Of course, the reverse is also true. If the marketinterest rate were to fall, the value of the mortgagewould rise and the holder of the mortgage wouldrealize a capital gain. The problem is that interestrates are extremely difficult to predict. If the mar-kets were populated by investors who are indif-ferent to whether they sustain a capital loss or acapital gain (i.e., indifferent to risk), the fact thatbond prices and interest rates are inversely relatedwould not be an issue. Interest rates would beinvariant to the maturity of the asset. However,financial markets are populated by risk-adverselenders (i.e., those more concerned with sufferinga capital loss than a getting a capital gain). Conse-quently, there is a risk premium on bonds (includ-ing mortgages) that increases as the term of theloan increases. The risk premium is tiny—essen-tially zero—for loans of only a few months. Therisk premium for 30-year loans can be fairlylarge, depending on market circumstances.

Because the interest rates on ARMs adjustover the term of the loan, ARMs have less marketrisk than the corresponding fixed-rate loan withthe same maturity. Consequently, with an ARM,some of the market risk associated with mortgage

lending is assumed by the borrower. As notedearlier, like anything else, risk is priced. Hence,ARMs have an initial rate that is lower than therate on an otherwise equivalent-maturity fixed-rate loan. Table 1 shows the annual average dif-ference between the initial rates on conformingfixed-rate mortgages and ARMs from 1997 to2004. The differences vary from year to year, butrange from about 50 to about 100 basis points.4

Because ARMs have a lower initial interest rate,they are particularly good for individuals whoplan either to sell their house or pay off the loanafter a short period of time.

THE MORTGAGE MARKETThemortgage market is a phrase that describes

a vast array of institutions and individuals whoare involved with mortgage finance in one wayor another. This market is broken down into twoseparate yet connected entities: the primary mort-gage market and the secondary mortgage market.The primary mortgage market is a market wherenew mortgages are originated. The secondarymortgage market is a market where existing mort-gages are bought and sold. Historically, the sec-ondary mortgage market was small and relativelyinactive. Two entities, the Federal National



Table 1Comparing Effective Interest Rates on Fixed- and Adjustable-Rate Mortgages (assuming an LTVratio between 0.8 and 0.9)

Fixed-rate ARM Difference

1997 7.91 6.95 0.96

1998 7.21 6.69 0.52

1999 7.47 6.93 0.54

2000 8.3 7.5 0.8

2001 7.19 6.72 0.47

2002 6.84 6.13 0.71

2003 6.05 5.2 0.85

SOURCE: Federal Housing Finance Board, historical summary tables, by loan to price ratio; www.fhfb.gov/Default.aspx?Page=53.

4 One basis point is one one-hundredth of a percentage point.

Mortgage Association (Fannie Mae) and theFederal Home Loan Mortgage Corporation(Freddie Mac), have changed that.5 These firmswere chartered by Congress to create a secondarymarket in residential mortgages. They are privatecompanies and not part of the U.S. government;however, they are called government-sponsoredenterprises (GSEs) because the government placeslooser restrictions on them relative to fully privatecompanies. Specifically, Fannie Mae and FreddieMac are exempt from state and local taxes (exceptproperty taxes) and have conditional access to a$2.25 billion line of credit from the U.S. Treasury.

Fannie Mae and Freddie Mac issue debt anduse the proceeds from the sale of their debt topurchase mortgages in the secondary market.Although the debt that they issue is not backedby the full faith and credit of the United Statesgovernment—i.e., is not explicitly government

debt—GSE debt typically trades at interest ratesonly a few basis points more than that of other-wise equivalent government debt. This suggeststhat investors believe that the United States gov-ernment would honor GSE debt in the event of acrisis.

Because of Fannie Mae and Freddie Mac andthe increased sophistication of U.S. financialmarkets more generally, the secondary market inresidential mortgages expanded rapidly in the1990s and now plays a major role in residentialfinance. Figure 2 shows the growth of the second-ary mortgage market since 1989 on the left axis.The right axis displays the percentage of second-ary market value created by GSEs. Although theGSEs account for much of the secondary mortgagemarket growth in the late 20th century, theirinfluence has decreased sharply in recent yearsas more and more private firms have entered themarket. Before the growth of the secondary mort-gage market, banks and savings and loan associa-tionsmademost of the residential real estate loans.



5 For a more detailed discussion of the evolution of the secondarymortgage market, see Gerardi, Rosen, and Willen (2007), Frameand White (2005), and Green and Wachter (2005).

0

500

1,000

1,500

2,000

2,500

3,000

1989 1991 1993 1995 1997 1999 2001 2003 2005

Billions of Dollars

0

10

20

30

40

50

60

70

80

90

100

Percentage of Market Value Held by GSEs

GSE (left scale)

Private (left scale)

Percent GSE (right scale)

Figure 2

Secondary Market Activity of Total Mortgage Loans

Most often, they originated the loan, serviced theloan contract, and actually lent the money. Thegrowth of the secondary market has resulted inincreased specialization in mortgage finance. Itis now frequently the case that the originator ofthe loan does not hold it until maturity. They takeapplications and do all of the necessary creditchecks and paper work until the time that theloan is closed (i.e., the loan agreement is signed).In many cases the mortgage originator initiallymakes the loan; however, their intention is to sellthe loan quickly. Such firms generate earningsfrom the fees they charge. The individual or entitythat purchases the mortgage is actually makingthe loan. It is also the case that the entity thatmakes the loan does not necessarily service theloan contract—that is, collect the periodic inter-est and principal payments, notify the borrowerof overdue payments, keep records, and makeproperty tax and homeowner’s insurance pay-ments. Instead, other firms charge a fee for pro-viding these services. In some cases, loans aresold individually, while in other cases they arepackaged together and sold as a single asset.The practice of consolidating loans or other debtinstruments into single assets or securities iscalled securitization. Securitization is now com-mon in the mortgage market. Mortgage-backedsecurities, as these assets are called, are boughtand sold in financial markets much like stocksor IOUs from private companies or the govern-ment: for example, corporate bonds, governmentTreasury bills and bonds, commercial paper, andnegotiable certificates of deposit.

To limit the risk of default, Fannie Mae andFreddie Mac place restrictions on the mortgagedebt that they will purchase. Factors that play animportant role in assessing the risk of a particularloan are as follows: the payment-to-income ratio,the debt-to-income ratio, the loan-to-value ratio,and the size of the loan. The payment-to-incomeratio is the monthly loan payment including realestate taxes divided by the borrower’s monthlyincome. The debt-to-income ratio is the ratio of allmonthly debt expenses to monthly gross income.The loan-to-value, or LTV, ratio is the loan amountdivided by the estimated (or appraised) value ofthe property where the difference between the

estimated property value and the loan amount isthe down payment.

There are no hard and fast rules about limitsto these ratios because other factors, such as anindividual’s credit history, enter in to the deter-mination of an individual’s creditworthiness;however, there are some guidelines. Traditionally,a payment-to-income ratio much larger than 25percent or a debt-to-income ratio of more thanabout 36 percent is considered cause for concern.A loan is considered “conventional” or “conform-ing” if the LTV ratio is 80 percent or smaller. Asa general rule, the higher these ratios are, thegreater is the risk of default. Loans made to bor-rowers that have ratios significantly larger thanthose stated above or other impairments, such aslow credit scores, are considered subprime.6

Fannie Mae and Freddie Mac do not purchaseloans that exceed a certain amount. The maximumloan amount changes yearly based on the resultsof a survey by the Federal Housing Finance Board.For a one-family home in the lower forty-eightstates in 2007, the maximum loan amount is$417,000. Loans larger than this amount arereferred to as jumbo loans. Taken together, theseguidelines and requirements give lenders an ideaof the level of risk that the secondary market iswilling to bear.

Like anything else, risk has a price. Lenderscompensate for making higher-risk loans by charg-ing a higher interest rate. There are a number ofways this can be done. The most obvious is thatthe lender merely charges a higher interest rateon more-risky mortgage loans—the greater therisk of default, the higher the rate. Hence, it is notsurprising that on average subprime loans have ahigher stated interest rate than conventionalloans. There are other ways to charge a highereffective rate, however. For example, in the caseof an LTV ratio that is greater than 80 percent,the lender often requires the borrower to purchaseprivate mortgage insurance (PMI), whereby a thirdparty bears the risk of default. The borrower mayprefer this option to paying a higher mortgage ratebecause once the LTV ratio reaches 80 percent(either by an appreciation of the property value

6 For more details, see Chomsisengphet and Pennington-Cross (2006).



or a reduction in the loan balance over time), thePMI can be discontinued.

The lender is also protected if the borrowerachieves an LTV ratio of 80 percent by taking asecond mortgage to make up the difference. Forexample, the borrower may have an 80-10-10mortgage, indicating that 80 percent of the loanis financed by the first mortgage, 10 percent isfinanced by a second mortgage, and 10 percent isa down payment. Even smaller down payments,including no down payment at all, are possible.Such loans are frequently, but not always, sub-prime. Because the second mortgage is subordi-nate to the first—meaning that in the case ofdefault, it is repaid only after the first mortgageis repaid—the holder of the second mortgagebears most of the default risk. Consequently, theinterest rate on the secondmortgage is higher thanthat on the first. Borrowers may benefit fromusing this method, however, because the secondmortgage typically has a shorter maturity thanthe first. Hence, once the second mortgage is paidoff, the borrower has only the lower-interest firstmortgage. In any event, borrowers with LTV ratiosgreater than 80 percent can expect to pay moreeither by paying a higher rate on the first mort-gage, by taking out PMI, or by having a higher-interest second mortgage. Mortgage borrowerswith LTV ratios less than 80 percent do not, how-ever, typically receive significantly lower interestrates. The reason is that the default risk is verysmall when the LTV ratio is 80 percent. Lendersknow that with this LTV ratio, it is very likelythat they will be able to recover all or nearly allof the loan balance in the event of a default. Con-sequently, a smaller LTV ratio provides essentiallyno reduction in default risk; hence, there is noreason for the lender to compensate the borrowerby giving the borrower a lower interest rate.

The existence of a secondary mortgage marketis beneficial to both the borrower and the lender.For the borrower, robust mortgage trading allowsfor more intense competition; 20 or 30 years ago,local financial institutions were the only optionfor some borrowers. Today, borrowers have accessto national (and even international) sources ofmortgage finance. Additionally, the Internet hasprovided an outlet to quickly compare mortgage

rates. Investors also benefit by having a widerrange of investments that they can use to diver-sify their portfolio. Moreover, a well-functioningsecondary mortgage market allows investors torealign their portfolios as circumstances change.

MORTGAGE FINANCENow that we have discussed some facts about

mortgages and the mortgage market, it is time todiscuss the nuts and bolts of mortgage paymentschedules and the real effective interest rate thatone pays when taking out a mortgage. We beginour discussion by showing how the fixed, monthlypayment on a fixed-rate mortgage is determined.To make the discussion as concrete as possible,we consider a borrower who wants a $200,000,30-year, fixed-rate mortgage with a contract inter-est rate of 6 percent annually. The question ishow much will this borrower have to pay eachmonth to pay off the loan in 30 years? The answer,$1,199.10, is obtained from the formula

(1)

whereMP is the monthly mortgage payment,MB0 is the initial mortgage balance—the amountborrowed—n is the number of months over whichthe loan is amortized, and r is the monthly interestrate (annual interest rate divided by twelve).Consequently, the monthly payment is

This formula may seem complicated, but ithas an intuitive explanation. The first part of theformula,MB0�1 + r�n, is just the total outstandingbalance if someone borrowed $200,000 and madeno payments for 30 years. It demonstrates theeffect of what is called compound interest—thatis, accumulating interest on both the principaland the interest in the previous month everymonth for 30 years. To illustrate, assume that nopayment is made during the first month. The out-standing balance at the end of the first month,

MP =

+( )+( )

$ , ..

.200 000 1 0 06 12

0 06 12

1 0 06 12

360//

/ 3360 11199 10

−

= $ , . .

MP MB rr

rn

n= +( )+( ) −

0 11 1

,



MB1, would be $201,000 = $200,000 + $200,000�0.005�, i.e.,MB1 =MB0 + rMB0. Note that thisexpression can be rewritten more compactly asMB1 =MB0�1 + r�, i.e., $200,000�1.005� = $201,000.If no payment were made the next month, by theend of the second month the total amount owed,MB2, would be $201,000 + $201,000�0.005� =$201,000�1.005� = $202,005—the amount of theinitial loan, plus $1,000 in interest for the firstmonth and $1,005 in interest the second month.Note that the additional $5 for the second monthis interest paid on the $1,000 in interest owed atthe end of the first month—earning interest oninterest, i.e., “compound interest.” Also, note thatthe amount owed at the end of the second monthcould be written more compactly as $200,000�1.005�2 = $202,005, i.e.,MB2 =MB1�1 + r� =MB0�1 + r��1 + r� =MB0�1 + r�2. This process gen-eralizes to any number of months so that

(2)

Equation (2) is the equation for compound inter-est. In the case of our 30-year mortgage example,if no payments were made for the life of theloan, the balance at the end of 30 years would be$200,000�1.005�360 = $1,204,515.04.

The second part of the monthly paymentequation, r/[�1 + r�n – 1], aggregates the effects ofmonthly interest and principal payments. Itreflects the fact that rather than letting the interestaccumulate over time, the fixed monthly paymentcovers all of the interest accrued during the monthand pays off part of the principal. Instead of owing$201,000 at the end of the first month of the mort-gage if no monthly payment were made, the indi-vidual who makes monthly payments would owe$199,800.90 = $201,000 – $1,199.10. Each succes-sive month, more of the fixed monthly paymentgoes to principal and less goes to interest as theprincipal balance declines. An amortizationschedule for our hypothetical loan is presentedin Table 2. Notice that it takes a long time to repaythe principal. After 10 years of the 30-year mort-gage, only about 16 percent of the principal hasbeen repaid. It takes 21 years before half of theprincipal has been repaid.

MB MB rmm= +( )0 1 .

Annual Percentage Rate

The analysis above is based on the contractrate on the mortgage. The effective rate on themortgage can be higher—in some cases, consider-ably higher. The purpose of this section is to dis-cuss the factors that affect the effective rate thatis paid on a mortgage. To help borrowers comparethe cost of borrowing, the Truth in Lending Actrequires that lenders disclose the annual percent-age rate, or APR. The Federal Truth in LendingAct was contained in the Consumer CreditProtection Act of 1968. This act is implementedby the Board of Governors of the Federal ReserveSystem with Regulation Z. Among other things,Regulation Z requires that all lenders disclose theAPR on credit to potential borrowers. The purposeof the APR is to make the interest costs of loanswith different structures, fees, etc., comparable.However, because loans can differ in many ways,the stated APR may not reflect the actual interestrate paid by the borrower. We begin by discussingthe rationale for the APR and its calculation. Wethen discuss reasons and situations where thestated APR will not reflect the true interest ratepaid by the borrower.

Calculating the APR. To understand thecalculation of the APR, it is necessary to showhow to determine the current price of any asset.Basically, the value of any asset is equal to thepresent value of the income it generates overtime. The idea of present value is closely relatedto the idea of compound interest covered herepreviously. Compound interest answers the ques-tion: If I invest a sum of money (say $200,000)today, how much will I have at some future date(say 30 years from now) if the annual interestrate is r percent (say 6 percent)? In our mortgageexample, the question was fundamentally thesame—if I borrow $200,000 today at an interestrate of 6 percent, how much will I owe in 30years if I make no monthly payments? Ouranswer was $1,204,515.04.

Present value asks the reverse question: If Iam to get a sum of money (say $1,204,515.04) atsome future date (say 30 years from now), howmuch would it be worth to me today if the annualinterest rate is 6 percent? Now of course theanswer is $200,000. Hence, the present value





Table 2Partial Amortization Table for a 6 Percent Fixed-Rate Mortgage

Beginning Monthly Interest Principal EndingMonth mortgage balance payment for month repayment mortgage balance

1 $200,000.00 $1,199.10 $1,000.00 $199.10 $199,800.902 199,800.90 1,199.10 999.00 200.10 199,600.803 199,600.80 1,199.10 998.00 201.10 199,399.714 199,399.71 1,199.10 997.00 202.10 199,197.605 199,197.60 1,199.10 995.99 203.11 198,994.496 198,994.49 1,199.10 994.97 204.13 198,790.367 198,790.36 1,199.10 993.95 205.15 198,585.218 198,585.21 1,199.10 992.93 206.17 198,379.049 198,379.04 1,199.10 991.90 207.21 198,171.8310 198,171.83 1,199.10 990.86 208.24 197,963.5911 197,963.59 1,199.10 989.82 209.28 197,754.3112 197,754.31 1,199.10 988.77 210.33 197,543.98

35 192,641.11 1,199.10 963.21 235.90 192,405.2236 192,405.22 1,199.10 962.03 237.07 192,168.14

59 186,641.83 1,199.10 933.21 265.89 186,375.9460 186,375.94 1,199.10 931.88 267.22 186,108.71

118 168,447.40 1,199.10 842.24 356.86 168,090.54119 168,090.54 1,199.10 840.45 358.65 167,731.89120 167,731.89 1,199.10 838.66 360.44 167,371.45121 167,371.45 1,199.10 836.86 362.24 167,009.21122 167,009.21 1,199.10 835.05 364.06 166,645.15

238 109,964.76 1,199.10 549.82 649.28 109,315.48239 109,315.48 1,199.10 546.58 652.52 108,662.96240 108,662.96 1,199.10 543.31 655.79 108,007.17241 108,007.17 1,199.10 540.04 659.07 107,348.11242 107,348.11 1,199.10 536.74 662.36 106,685.75

251 101,266.24 1,199.10 506.33 692.77 100,573.47252 100,573.47 1,199.10 502.87 696.23 99,877.23253 99,877.23 1,199.10 499.39 699.71 99,177.52254 99,177.52 1,199.10 495.89 703.21 98,474.30

355 7,070.36 1,199.10 35.35 1,163.75 5,906.61356 5,906.61 1,199.10 29.53 1,169.57 4,737.04357 4,737.04 1,199.10 23.69 1,175.42 3,561.63358 3,561.63 1,199.10 17.81 1,181.29 2,380.33359 2,380.33 1,199.10 11.90 1,187.20 1,193.14360 1,193.14 1,199.10 5.97 1,193.14 0.00

formula is just the inverse of the compoundinterest formula, i.e.,

(3)

In the case of mortgages, and most invest-ments, all of the money is not paid on the matu-rity date. Rather, income is received periodicallyover time. The present value of the entire streamof income to be received over time is just the sumof the present value of each of the future payments.In the case of our mortgage example, the presentvalue of the mortgage loan is given by

(4)

whereMPi denotes the monthly payment to bereceived imonths in the future. In the case of afixed-rate loan, the monthly payments are thesame—that is,MPi =MPj for all i and j—andequation (4) can be written more compactly as

(5)

For our hypothetical mortgage, the present valueof receiving $1,199.10 per month for 30 years is

This shows that themortgage lender is, in essence,purchasing an investment that pays $1,199.10per month for each of the next 360 months.

In this example, we knewMP and r, so wecould solve the equation for the present value,MP0. Although it is more complicated to solve, ifwe knewMP andMP0 we could have solved theequation for r. The question is, If I were to pay$200,000 today for the right to receive $1,199.10each month for the next 360 months, what wouldbe the effective annual interest rate? Of course,we know the answer is 6 percent (0.005 times 12).

Equation (5) can be modified slightly to deter-mine the APR. The modification stems from thefact that there are expenses associated withfinancing the purchase of a home rather than

$ $1199 101 005 1

0 005 1 005

360

360..

. .

( ) −( )

= 2200 000, .

MB MPr

r r

m

m01 1

1=

+( ) −+( )

.

MBMP

r

MP

r

MP

r0

11

22

3603601 1 1

=+( )

++( )

+ ++( )

L ,

MB MB rmm

0 1= +( )/ .

paying cash for it. These additional expenses areconsidered pre-paid interest. For example, if youborrow $200,000 to buy a home but, in doing so,incur $3,000 in expenses solely because you arefinancing the purchase, you are in effect onlyborrowing $197,000. The calculation of the APRaccounts for this fact by making an adjustmentfor these expenses, which are referred to as fees.Hence, the APR is the interest rate that solves

(6)

So, applying this formula to our hypotheticalexample, solving the equation

for r, yields a monthly interest rate of 0.512 per-cent or an annual APR of 6.142 percent.

Obviously, the larger are the fees, the smalleris the effective loan and the higher is the APR.Hence, when considering a mortgage, one mustconsider both the stated mortgage rate and thefees that are required to get this rate. Indeed, it isoften possible to get a lower mortgage rate by pay-ing higher fees. When considering such options,the APR can be very useful for deciding whichmortgage option is best.

It is important to note that the APR is notalways calculated the same way by all financialinstitutions; different fees may or may not beincluded. According to the Federal Reserve Board,fees that are considered part of the finance chargeare as follows: interest, service charges, buyer’spoints, assumption fees, and insurance chargesrequired by the lender (with a few exceptions).Fees that are not part of the finance charge areapplication fees (if charged to all applicants), latefees, bounced check fees, seller’s points, titlingfees, appraisals, credit report fees, taxes, notaryfees, and fees for opening an escrow account.7

$ , $ , $ , .200 000 3 000 1199 101 1

1

360

36− =+( ) −

+( )r

r r 00

MB MPr

r r

m

m01 1

1− =

+( ) −+( )

Fees .

7 The general rule is that if the fee is charged solely because thepurchase is being financed, it should be included. Excludingcredit report fees would appear to violate this rule because theyare included solely because the purchase is being financed.Congress wrote this exclusion into the Truth in Lending Act.



Table 3 displays the fees that are included andexcluded from the APR. The third column dis-plays fees which may or may not be included,depending on the lender and the size of the fee.

It is also important to note that the lender hassome leeway in terms of the accuracy of the APRthat he reports. The actual finance charge can beunderreported by as much as $100. Also, accord-ing to Regulation Z, the reported rate is consideredaccurate if it is within one-eighth of 1 percent ofthe true rate. If one bank quotes a rate of 6.125percent while another bank quotes a rate of 6.25percent, it is hard to determine which rate is reallylower because of the allowed margin of error.

Value of the APR. The APR is very useful,but it has limitations. Important among these isthe fact that the APR assumes that you will havethe mortgage for its entire term. Although mostmortgages have a term of 30 years, only a smallportion of mortgages last their full term. Mostmortgages are paid off early, because the borrowerprepays the loan, sells the property, refinancesthe mortgage, or defaults. According to DouglasDuncan, chief economist of the Mortgage BankersAssociation, the average term of a mortgage is 3to 5 years. The APR for our previous hypothetical$200,000, 30-year mortgage—assuming closingcosts of $3,000—is 6.142 percent. This APR isbased on the assumption that this mortgage willrun to term (i.e., 30 years). But if the house is

sold or the mortgage refinanced after 3 years,the effective APR would be 6.577 percent. If itis sold or refinanced after 5 years, the effectiveAPR would be 6.367 percent. A modification toour original formula is necessary to calculate theAPR of a loan that is paid off before maturity.The modification comes from the fact that ratherthan paying off the entire loan over the term ofthe mortgage, the borrower must pay off theremainder of the mortgage balance, RBm, whenthe loan is repaid. The modification takes thepresent value of this payment into considerationin calculating the APR. Specifically, the modifiedAPR formula is

(7)

The remaining balance on our mortgage can beread off the corresponding row of our amortiza-tion table (Table 2). After 5 years, the remainingmortgage balance is $186,375.94 (the balance atthe end of 59 months or the beginning of 60months). Applying the formula to our example,

$ , $ ,

$ , .

200 000 3 000

1199 101 1

1

59

59

− =

+( ) −+( )

r

r r

+

+( )$ , .

;186 375 94

1 59r

MB MPr

r r

RB

r

m

mm

m01 1

1 1− =

+( ) −+( )

+

+( )Fees .



Table 3Fees and the Annual Percentage Rate

Included Excluded Sometimes included

Interest Late payment fees Appraisals (excluded if required of all applicants)

Service or carrying charges Returned check fees Home inspections and pest inspections(excluded if required of all applicants)

Broker fees Title fees Voluntary insurance

Private mortgage insurance Taxes Application fees (excluded if required of allapplicants, otherwise included)

Assumption fees License fees

Points Appraisal fees

Fees for establishing an Credit report feesescrow account

solving for r gives a monthly interest rate of 0.531percent, or an annual rate of 6.367 percent. Hence,the quoted APR understates the true effectiveinterest rate if the borrower plans to prepay theloan before its maturity date.

The APR is also less useful for comparingARMs. The quoted APR on an ARM is not onlybased on the full term of the loan, but alsoassumes that the index to which future rateadjustments are linked will remain constant forthe life of the loan. It neither accounts for thevolatility of the index nor allows borrowers tocompare the different indices that may be avail-able. It also ignores the maximum rates allowedunder a particular adjustable rate structure.

Refinancing

There are three reasons that someone mightwant to refinance a mortgage: to obtain a lowerinterest rate, to consolidate interest paymentsthat are not tax deductible into mortgage interestpayments which are tax deductible, or to obtaincash for some other purpose. Refinancing to lowerinterest payments is often a good idea if interestrates have fallen since the original mortgageclosed or if a person currently has an ARM andwants to avoid the uncertainty of future interestrate adjustments. There are two important factsto keep in mind when considering refinancingsolely to obtain a lower interest rate. The first isthe term of the loan. If the new mortgage has aterm that is shorter than the term remaining onthe existing mortgage, the only issue is whetherthe effective interest rate is lower than that on thecurrent fixed-rate mortgage. If the term on the newmortgage is longer than the remaining term of theexiting mortgage, the decision is more compli-cated. For example, if one refinances a 30-yearmortgage with a remaining term of 20 years witha new 30-year fixed-rate mortgage, at a lowerinterest rate, the interest rate savings may be off-set by the fact that interest will be paid over 30years instead of 20. Of course, if the loan has noprepayment penalty, the effective term of anymortgage can be set anywhere the borrowerdesires simply by adjusting the payment to thatof the desired term. For example, assume thatafter 10 years we want to refinance our current

$200,000, 30-year mortgage that we took out wheninterest rates were 6 percent with a new 30-yearfixed-rate mortgage with a 5 percent rate. Theamortization table (Table 2) shows that the remain-ing balance on the loan is $167,371.45. Usingequation (1) we calculate that our monthly pay-ment for borrowing $167,371.45 for 30 years is$898.49, which is $300.62 less than the currentmonthly payment of $1,199.10. While the interestrate is lower, the total interest cost over the lifeof the loan is $156,083.56, compared with$120,412.80 for a 20-year fixed-rate mortgage at6 percent (the current mortgage). The differenceis due to the fact that interest is being paid over30 years with the newmortgage and only 20 yearswith the old. Hence, while the annual interestrate is lower, the total interest cost over the lifeof the loan is higher.

Because there are no prepayment penalties,the borrower can effectively determine the termof the mortgage simply by adjusting the monthlypayment. For example, using equation (1) wefind that the monthly payment on a 20-year,fixed-rate loan with an annual rate of 5 percentis $1,104.58. Hence, with a monthly payment of$1,104.58, the 30-year loan would be paid off atthe same time that the existing loan would havebeen paid off (20 years), with a total interest costof $97,727.15. Alternatively, one could maintainthe monthly payment at the level of the oldmortgage, $1,199.10. In this case, the loan wouldbe paid off in about 17 years, 6 months, with atotal interest cost of about $84,000.8

Some financial institutions offer a no-costrefinance. This means that there are no costs tothe loan. In this case, the stated rate and the APRare identical. In effect, the costs are covered in theinterest rate: That is, the costs have been financed,resulting in a contract interest rate that is higherthan the rate for loans that have finance costs.

Comparisons such as the above are particu-larly important when considering consolidatingnon-tax-deductible debt (e.g., credit card debtand auto loans) into a mortgage. The mortgagehas two advantages: the interest rate will likely

8 Note that ifMB0,MP, and r are known, it is possible to solveequation (5) form.



be lower and the interest is deductible for taxpurposes. However, if one anticipated paying offthe consolidated loan before the term of the newmortgage, the interest costs could be higherbecause the loan is being repaid over a muchlonger period.

Home Equity Loans

The equity in a home is the differencebetween the current market value of the homeand the remaining balance on all of its mortgages.Of course, the true market value of the home isnot known until the house is actually sold; con-sequently, the home’s equity is estimated bysubtracting the principal remaining on existingmortgages from an estimate of the property’smarket value. A home equity loan is simplymoneyborrowed using the equity in the home as collat-eral. Home equity loans have two advantages:First, because the loan is collateralized by thehome, the interest rate is lower than what couldbe obtained on an otherwise identical unsecuredloan. Second, with some exceptions, the interestpaid on home equity loans is deductible for taxpurposes. Hence, home equity loans (or homeequity lines of credit) are low-cost methods offinance for many homeowners. For many people,the equity in their home is their major source ofwealth. Hence, using home equity loans to financecurrent consumption may put their wealth at risk.

A reverse mortgage can be thought of as aparticular type of home equity loan, because inthis case the individual is borrowing money usingthe equity in the home as collateral. Instead ofmaking payments, the homeowner receives pay-ments. The homeowner can select to have a fixedmonthly payment, a line of credit, or both. Theamount owed increases with the payments ordraws on the line of credit, and interest cost isbased on the outstanding loan balance.

From the point of view of the lender, reversemortgages are investments. Instead of receivingmonthly payments to cover interest, fees, andprincipal, all of the money lent, interest pay-ments, and incurred fees are received in a singlepayment when the house is sold.

Reverse mortgage loans are available only toindividuals who are 62 years or older. The loan

payments are not taxable and generally do notaffect Social Security or Medicare benefits. Likeother mortgages, lenders charge origination feesand other closing costs; consequently, the effectiveinterest rate will be higher than the contract rate.As is the case for regular mortgages, this meansthat the effective interest rate may be considerablyhigher for individuals who stay in their homesfor only a short time after taking out a reversemortgage. Lenders may also charge servicing feesduring the life of the mortgage. As with regularmortgages, the interest rate can either be fixed orvariable, with the variable rate tied to a specificindex that fluctuates with market rates. Reversemortgages may be useful for people with homeequity but relatively low periodic income. Becausethe loan is repaid when the home is sold, thedanger is that the borrower will use up all of theequity in the home, having nothing to leave totheir heirs. Most reverse mortgages have a “non-recourse” clause, which prevents the borrower,or their estate, from owing more than the valueof the home when it is sold. This protects theborrower, but it also means that the lender willbe conservative in determining how much theyare willing to lend. There are basically two typesof reverse mortgages: (i) federally insured reversemortgages known as home equity conversionmortgages (HECMs), which are backed by theU.S. Department of Housing and Urban Develop-ment (HUD), and (ii) proprietary reserve mort-gages, which are privately funded.

As with any mortgage, care must be exercisedwhen considering the costs and benefits of areverse mortgage. To better understand reversemortgages, it is useful to consider a hypotheticalexample of how a reverse mortgage works. Assumethe homeowner would like to receive a monthlypayment of $1,000 and that they can obtain areverse mortgage at an annual interest rate of 6percent. At the end of the first month, the home-owner would owe $1,005, the $1,000 paymentreceived at the beginning of the month plus $5interest for the month. Letting LBi denote the loanbalance at the end of the ith month, the amountowed at the end of the first month would beLBi =MP�1 + r�. The balance at the end of thesecond month would be the $1,005 balance at



the end of the first month, plus the interest onthis amount for the month—that is, $1,005�1.005�—plus the $1,000 payment at the beginning ofthe second month plus interest—that is, $1,000�1.005�. This can be expressed as LB2 =MP�1 + r�2

+ MP�1 + r�. The amount at the end of themthmonth is given by

which can be written more compactly as

(8)

Note the similarity between this equation and theright half of equation (1). Now ask the question,What would be the outstanding balance at theend of 10 years if an individual drew $1,000 permonth and the annual interest rate charged was6 percent? The answer is $163,879.35—that is,

This means that if a homeowner had equity of$165,000, they could draw $1,000 per month for10 years before the total amount of the loans plusinterest essentially consumed all of the home’sequity.

Of course, the question that individuals con-sidering a reverse mortgage are most concernedabout is, How much will I be able to receive eachmonth given the value of my home? The answer isobtained by solving equation (8) forMP—that is,

(9)

where HE replaces LBm and denotes the home-owner’s equity—the maximum amount that thelender will lend on a reverse mortgage. Again,using our example, if the home equity is $165,000and the annual interest rate is 6 percent, equation(9) indicates that the individual could receive$1,006.84 per month for 10 years.

Equation (9) considers only the interest costs.

MP HEr

r m=+( ) −

1 1,

$ ,.

.$ , . .1 000

1 005 10 005

163 879 35120( ) −

=

LB MPrrm

m

=+( ) −

1 1.

LB

MP r MP r MP r

m

m m

=

+( ) + +( ) + + +( )−1 1 11 L ,

It ignores loan origination fees and other closingcosts, as well as servicing fees that the lendermay charge. These costs and fees are treated asloans. Origination fees and closing costs areincurred at the time of the loan, whereas servicingfees may be charged in each period. Such costsreduce the equity available to make monthlypayments. For example, assume that the closingcosts are $1,000. We know from our compoundinterest formula, equation (2), that in 10 yearsthe total amount owed on this $1,000 loan plusinterest will be $1,819.40. This means that only$163,180.60 of the home’s equity will be avail-able for monthly payments. In our example, thismeans that monthly payment would be reducedfrom $1,006.84 to $995.74.

An important factor in determining the sizeof the monthly payment is the period of time overwhich payments are expected to be made. Forexample, assume the individual is 65 and expectsto live in the home until age 85. Hence, they wouldlike to receive monthly payments for 20 years.Following up on our example, if we assume thereare no closing costs, the monthly payment thatwould exhaust the $165,000 in home equity in20 years would be $357.11. If we assume there is$1,000 in closing costs, this amount is reducedto just $349.95.

Generally speaking, the older you are whentaking out the reverse mortgage, the more you willbe able to borrow. This is due to the fact that theperiod over which you receive payments is likelyto be shorter. Also, the higher the value of yourhome and the larger the equity, the more you canborrow. Your monthly payments will also behigher the lower the interest rate. Because theinvestor must project the home’s future value,which is often difficult to do, reverse mortgagesare relatively risky investments. Consequently,the interest rates on reverse mortgages are typi-cally higher than those on otherwise equivalentmortgages.

CONCLUSIONSThis paper addresses a number of significant

issues facing the prospective home buyer. For



most people, buying a home is the largest pur-chase they will ever make, and a thorough under-standing of the terminology and structure of theresidential finance market can mean the differ-ence between an agonizing experience and arewarding one. Although the mortgage industryis too sophisticated to describe completely in thisshort paper, hopefully the concepts elucidatedhere will reduce the anxiety for those trying tofinance the American dream.

REFERENCESBernanke, Ben S. “The Housing Market and SubprimeLending.” Speech to the 2007 InternationalMonetary Conference, Cape Town, South Africa,June 5, 2007; www.federalreserve.gov/boarddocs/speeches/2007/20070605/default.htm.

Chomsisengphet, Souphala and Pennington-Cross,Anthony. “The Evolution of the Subprime MortgageMarket.” Federal Reserve Bank of St. Louis Review,January/February 2006, 88(1), pp. 31-56.

Frame, W. Scott and White, Lawrence J. “Fussingand Fuming over Fannie and Freddie: How MuchSmoke, How Much Fire?” Journal of EconomicPerspectives, Spring 2005, 19(2), pp. 159-84.

Gerardi, Kristopher; Rosen, Harvey S. and Willen,Paul. “Do Households Benefit from FinancialDeregulation and Innovation? The Case of theMortgage Market.” CEPS Working Paper No. 141,Center for Economic Policy Studies, March 2007.

Green, Richard K. and Wachter, Susan M. “TheAmerican Mortgage in Historical and InternationalContext.” Journal of Economic Perspectives, Fall2005, 19(4), pp. 93-114.



http://research.stlouisfed.org/publications/review/06/01/ChomPennCross.pdf

http://research.stlouisfed.org/publications/review/06/01/ChomPennCross.pdf



Changing Trends in the Labor Force: A Survey

Riccardo DiCecio, Kristie M. Engemann, Michael T. Owyang, and Christopher H. Wheeler

The composition of the American workforce has changed dramatically over the past half centuryas a result of both the emergence of married women as a substantial component of the labor forceand an increase in the number of minority workers. The aging of the population has contributedto this change as well. In this paper, the authors review the evidence of changing labor force par-ticipation rates, estimate the trends in labor force participation over the past 50 years, and find thataggregate participation has stabilized after a period of persistent increases. Moreover, they examinethe disparate labor force participation experiences of different demographic groups. Finally, theysurvey some of the studies that have provided explanations for these differences. (JEL J21, E32)


labor market relative to the total number of resi-dents in the country. Such a trend has likelycontributed to the rise in U.S. living standards(e.g., income per capita) over the postwar period.

In spite of this long-run rise in the LFPR,there has been a modest drop in the overall par-ticipation rate within the past six years, whichhas generated some concern among economists.If this decrease represents a change in the trendLFPR, the U.S. economy may be faced with fewerwork-oriented individuals per resident in thecoming decades.

What accounts for the changes in the LFPRin the United States over the past several decades?Numerous studies have documented changes invarious U.S. demographics, including the age andethnic composition of the population, that havesignificantly affected the nature of the labor force.In 1960, prime-age white males—from 25 to 54years of age—comprised, by far, the largest laborforce component: nearly 40 percent. Althoughthis group still represented 31 percent of the work-

One of the primary indicators of thestate of the U.S. labor market is thelabor force participation rate (LFPR).It is measured each month by the

Bureau of Labor Statistics (BLS) as the fractionof the civilian, non-institutional population 16years or older who are either working or activelyseeking work. The LFPR is a useful complementto other indicators, such as employment and theunemployment rate, in assessing labor marketconditions. For example, a low unemploymentrate is a much stronger indication of a tight labormarket when accompanied by a high participa-tion rate.

Although the LFPR is constantly changingover the business cycle, the most noticeable fea-ture is its dramatic increase over the post-WorldWar II period. Between 1948 and 2006, the U.S.LFPR rose by more than 7 percentage points, withthe majority of the rise taking place between theearly 1960s and 2000. This increase implies that,compared with several decades ago, there aremore individuals currently participating in the

Riccardo DiCecio is an economist, Kristie M. Engemann is a senior research associate, Michael T. Owyang is a research officer, andChristopher H. Wheeler is a research officer at the Federal Reserve Bank of St. Louis. The authors benefited from conversations withAlessandra Fogli, Leora Friedberg, and Natalia Kolesnikova.


force in 2005, the gap with other demographicgroups appears to be closing. In particular,increases in both minority and women workershave dramatically altered the composition ofthe workforce. Moreover, the aging of the babyboomers has changed the age profile of theAmerican population.

In this paper, we review the trends in laborforce participation over the past half century,including a look at both the long-run movementsin the LFPR as well as its short-run fluctuations.We then examine the components of the LFPR—disaggregating by gender, age, and race—to deter-mine the extent of and possible explanations forthe dispersion in labor force participation acrossdemographic groups. Finally, we consider thefuture of the LFPR in the United States.

TRENDS IN AGGREGATE LABORFORCE PARTICIPATION

The BLS maintains a monthly history oflabor force participation statistics dating back to1948. These figures are derived from the CurrentPopulation Survey, which reports informationon approximately 60,000 households.1 The toppanel of Figure 1 shows the historical path ofaggregate labor force participation in the UnitedStates; the overall increase in the participationrate since 1948 is evident. In January 1948, theoverall rate of labor force participation in theUnited States was roughly 59 percent. This rateheld fairly steady until the early 1960s, when it

DiCecio, Engemann, Owyang, Wheeler


Level and Trend

1950 1960 1970 1980 1990 2000

60

62

64

66

Cyclical ComponentPercent

1950 1960 1970 1980 1990 2000

−0.4

−0.2

0

0.2

0.4

0.6

Percent

Figure 1

Labor Force Participation Rate

NOTE: Top panel: level (gray) and trend (blue) extracted with an LP96 filter. Bottom panel: cyclical component extracted with BP18,96filters. The shaded areas denote National Bureau of Economic Research (NBER) recessions.

1 Unless otherwise noted, the source for data used in the figures isthe BLS.

began to rise; during the first quarter of 2000,the LFPR achieved its highest level—67.3 per-cent of the working-age population. Over thepast few years, however, the LFPR fell from its2000 level to 65.8 percent in January 2005. Ithas since rebounded to 66.4 percent as recentlyas December 2006.

Although the LFPR unquestionably trendedupward over the latter half of the twentieth cen-tury, the nature of its recent decline has sparkedsome debate. Some economists argue that thedecrease reflects structural changes in the labormarket (i.e., a change in the trend of the LFPR),whereas others view it primarily as a cyclicaldeviation from the trend. Aaronson et al. (2006),for example, use a cohort-based model to showthat the decline from 2000 until 2002 occurredas a result of the weak labor market conditionsstemming from the 2001 recession. Althoughthe initial drop in the LFPR was due to cyclicalfactors, their estimates indicate that the trendLFPR began to decline in 2003. On the other hand,Bradbury (2005) argues that the decline since2001 is a reflection of slack in the economy andis purely cyclical, although she does note that,relative to previous economic recoveries, theperiod following the 2001 recession was charac-terized by unusually low participation rates amongteenagers and women.

To gauge the degree of labor market tightness,it is important to determine how changes in thetrend and in the cyclical component contributeto movements in the LFPR. If structural factorscause most of the decline in participation, thena low unemployment rate indicates a tight labormarket. On the other hand, if business cyclemovements cause most of the decline, a substan-tial part of it should be reversed in a relativelyshort period of time. People who temporarilydropped out of the labor force will start lookingfor jobs again and thus will be recorded as unem-ployed. In the latter case, a low unemploymentrate overstates labor market tightness.

A cursory examination of Figure 1 suggeststhat there may have been at least three differentregimes describing the LFPR over the past sixdecades: zero growth before 1960, a constanttrend growth between 1960 and 2000, and a

declining trend subsequent to the turn of the cen-tury. To estimate more formally how the trend inthe LFPR has evolved over time, we follow a stan-dard technique in which we use a low-pass filterto remove high-frequency fluctuations from theraw data.2 In particular, we remove cycles with aperiod less than 96 months. The resulting trendis shown in the top panel of Figure 1. Consistentwith the business cycle tradition, we then identifythe business cycle component of the LFPR datawith cycles of periods between 1.5 and 8 years—i.e., 18 and 96 months—which we extract byapplying a band-pass filter.3 The bottom panel ofFigure 1 shows the business cycle componentsof the LFPR.4

Based on these calculations, we find that thetrend component peaked in October 1998 at 67.2percent, declined afterward to a minimum of66.1 percent in January 2005, and increased by0.2 percentage points by the end of 2006.5 Thecyclical component of the LFPR increased slightlyafter the 2001 recession, declined until August2004, and recovered afterward.

To demonstrate some of the short-run cyclicalproperties of the LFPR series, we compare thebusiness cycle component of the LFPR with thoseof a common indicator of aggregate activity, indus-trial production.6 The correlation between thetwo and their relative standard deviations arereported in Table 1. Based on the correlation, wesee that the LFPR is moderately procyclical (i.e.,it rises during expansions and falls during con-tractions). This finding is consistent with the ideathat during economic upturns, potential workersare lured into the labor force because they per-ceive their job prospects to be strong. Duringrecessions, on the other hand, workers not only



2 For an overview of terminology, see the appendix.

3 See Baxter and King (1999) and Christiano and Fitzgerald (2003).

4 To consider the decline in the LFPR from early 2000 to the end of2005 and its slight recovery afterward as purely cyclical phe-nomena, fluctuations with a period of up to 36 years would needto be removed in the definition of the trend.

5 Clark and Nakata (2006) estimate the trend growth rate in theLFPR to be 0.3 percent from 1957-81 and 0.2 percent from 1981-2005. They attribute the decline in the trend to the decelerationof women’s LFPR.

6 Industrial production data come from the Federal Reserve Board.

lose jobs—thereby increasing unemployment—but also exit the labor force altogether because thenumber of employment opportunities becomesrelatively scarce.

These fluctuations in the LFPR, however, aresmall compared with those in industrial produc-tion. Indeed, we find that the LFPR is one-tenthas volatile as industrial production at businesscycle frequencies. This property of the LFPR mayreflect a high degree of inflexibility in the averageindividual’s labor force participation decisionover time. Because individuals need an incometo support their consumption, many decide towork (or at least seek work) regardless of whetherthe economy is expanding or contracting.

The vast majority of the movement of theLFPR, however, is associated with its trend, notits cyclical components. In the next sections, weexplore the long-run evolution of the LFPR bylooking at its disaggregate components—specifi-cally, its gender, age, and racial components.

GENDER AND THE LFPRA substantial portion of the rise in the aggre-

gate LFPR beginning in the 1960s can be attrib-uted to the rise in the labor force participationof women. In 1950, approximately one in threewomen 16 years of age or older participated inthe labor force. Figure 2 illustrates the rise infemale labor force participation over the latterhalf of the twentieth century: The LFPR for allwomen is depicted by the solid blue line, whilethe solid black and dotted lines show the LFPR

for married and single women, respectively.7 By1999, the overall female LFPR rose to its peak of 60percent. As Figure 2 shows, much of the increasein women’s participation can be attributed tomarried women, whose LFPR rose by more than30 percentage points between 1955 and 2005. TheLFPR of single women has also increased overthe past several decades, but much more mod-estly. Since 1999, the overall women’s LFPR hasremained fairly steady: between 59 and 60 percent.

Figure 3 highlights the differences in the LFPRacross genders. In particular, Figure 3 reveals apersistent decline in men’s LFPR since 1950, thesame period over which women’s LFPR saw itsmost significant increase. During that period,the male LFPR fell by 13 percentage points to its2006 rate of less than 75 percent.

The cycle decomposition of the LFPR bygender bears some similarity to that of the aggre-gate (Table 1); however, two important differencesemerge. First, men’s participation tends to besomewhat more procyclical than women’s par-ticipation. The correlation between industrialproduction and men’s LFPR at business cyclefrequencies is 0.41, whereas the same correlationfor women’s LFPR is 0.28. This result may, inpart, reflect the added-worker effect, in whichwomen enter the labor force to compensate for aspouse’s loss of a job. That is, as men becomeunemployed during an economic downturn, somewomenmay choose to enter the labor force to off-

7 A fourth category, not shown in the figure, includes widowed,divorced, and separated women. Their LFPR held fairly steadyaround 40 percent between the mid-1950s and the mid-1970s andincreased by less than 10 percentage points by the early 2000s.



Table 1Second Moments of the Business Cycle Components of LFPR, Total and by Gender

Total Men Women

Corr(x,ip) 0.35 0.41 0.28(0.13, 0.53) (0.24, 0.56) (0.04, 0.46)

Std(x)/Std(ip) 0.09 0.07 0.19(0.07, 0.11) (0.05, 0.08) (0.14, 0.24)

NOTE: ip denotes industrial production. Block-bootstrapped 95 percent confidence intervals are in parentheses.

set the loss in household income. This phenom-enon may temper the procyclicality of the femaleparticipation rate somewhat. Second, the LFPRfor women is nearly three times more volatile thanthe LFPR for men, which may reflect the ideathat society has viewed women as the primarychild-rearer and the secondary earner.8 Thus,while these days more women may be workingat any given time, they may remain more likelythan men to move in and out of the labor force.

The dramatic increase in married women’slabor force participation has been the subject of

many studies, too numerous to detail here. Wehighlight only a few of the myriad possible expla-nations. The improvement in labor-saving house-hold technologies has simplifiedmany daily tasks,such as cooking and cleaning, thereby givingwomen greater time to pursue work outside ofthe home (Greenwood, Seshadri, and Yorukoglu,2005). This hypothesis is supported by evidenceon the differences in married women’s labor forceparticipation decisions across cities. In a recentstudy, Black, Kolesnikova, and Taylor (2007) findthat married women’s LFPR was substantiallylower in cities with more traffic congestion, prox-ied by longer average commuting time. Control-ling for other factors such as the woman’s age,education, non-labor income, number of childrenby age group, andMSA unemployment rate amongwhite men, a small increase in a city’s averagecommuting time significantly reduced married



8 Compared with the first part of the sample, the volatilities of thebusiness cycle components were 1.5 and 3 times smaller for menand women, respectively, after 1984. Stock and Watson (2003)demonstrate the decline in volatility of many macroeconomicvariables, but they do not consider the labor force. However, theydo show that the conditional variance for civilian employmenthas declined since the mid-1980s (or since the mid-1970s when atrend is included).

Percent

1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

30

35

40

45

50

55

60

65

70

All

Married, Spouse Present

Single, Never Married

Figure 2

LFPR by Marital Status (Women)

NOTE: The data used in this figure are unadjusted annual percentages. The shaded areas denote NBER recessions.

SOURCE: 1955-75, U.S. Census Bureau, Statistical Abstract of the United States, 2003, www.census.gov/statab/hist/HS-30.pdf; 1976-2005,Bureau of Labor Statistics.

women’s labor force participation.9 If cities withgreater congestion involve greater amounts oftime required to run errands (part of householdproduction), women might participate in thelabor force less in those cities.

Alternatively, medical advances, such as thebirth control pill, have allowed women to delaymarriage and pregnancy, thus providing moreopportunity to invest in a career early in life(Goldin and Katz, 2002). Changes in societal atti-tudes have also made it more acceptable for

married women and women with young children(under the age of 6) to work (Aaronson et al.,2006). Fernández, Fogli, and Olivetti (2004) findthat men whose mothers worked when they wereyoung children seemingly had a preference forwives who also worked. These social changeshave enabled more women to pursue careers inprofessional fields such as business, law, andmedicine, which has in turn led to higher returnsto experience, both in absolute terms and relativeto those of men (Goldin, 2006).

Fogli and Veldkamp (2007) consider that“learning” is the underlying force behind thesharp rise in participation rates for marriedwomenwith young children. When deciding whether tojoin the labor force, women try to understand

9 For women with children under the age of 5, the effects werelargest. In particular, an increase of one minute in the average MSAcommuting time led to a 0.53-percentage-point decrease in theLFPR of women with a high school education. For women with acollege education, their LFPR decreased by 0.22 percentage points.



Levels and TrendsMenPercent

74

76

78

80

82

84

86

Women

1950 1960 1970 1980 1990 2000

35

40

45

50

55

60

Cyclical ComponentsMen

−0.5

0

0.5

Women

1950 1960 1970 1980 1990 2000

−0.5

0

0.5

Percent

Percent Percent

1950 1960 1970 1980 1990 2000 1950 1960 1970 1980 1990 2000

Figure 3

LFPR by Gender

NOTE: Left column: levels (gray) and trends (blue) extracted with an LP96 filter. Right column: cyclical components extracted with BP18,96filters. The shaded areas denote NBER recessions.

how important stay-at-home child rearing is indetermining the future labor market outcomes oftheir offspring; stated another way, they observethe outcomes of children of working women toassess the importance of nature versus nurture.As more women join the labor force, learninghappens at a faster pace, which reinforces theincrease in participation and generates the S-shaped participation rate observed in the data.Fogli and Veldkamp’s model is consistent withsurvey data that indicate an increasing positiveattitude toward mothers who participate in thelabor force. The model is also consistent withthe continuous increase in women’s wages overthe past two decades, despite a flattening par-ticipation rate. In a related paper, Fogli, Marcassa,and Veldkamp (2007) argue that the rise inwomen’s LFPR over the second half of the previ-ous century can be partly explained by a spatialcomponent. In other words, a county whoseneighbors have high female LFPRs is likely toalso have a high female LFPR. This is because,over time, learning occurs and cultural effects—including women’s increased participation inthe labor force—spread to nearby counties.

Figure 2 shows that the increase in women’sLFPR began to slow in the 1990s. Blau and Kahn(2007) argue that the responsiveness of women’slabor supply to changes in their wages decreasedby about half between 1980 and 2000. Moreover,changes in their husbands’ wages had less of animpact on married women’s labor supply duringthis period. One explanation that the authorsprovide is that, as more women entered the laborforce, they became more attached to working andthus less responsive to changes in wages. Also,higher rates of labor force participation meantthat fewer women were on the margin, taking await-and-see approach to entering the labor force.The end result was slower growth of women’sLFPR during the previous decade.10

As with the rise in women’s LFPR, many fac-tors have likely caused the decline in men’s LFPR.Hotchkiss (2005) cites several reasons that have

led to earlier retirement, such as the creation ofSocial Security in 1935 and firms’ increased pro-vision of private pensions following the RevenueAct of 1942. She also notes that the expansionof Social Security to include disability insurancegave workers more incentive to leave the laborforce due to disability. Juhn (1992) argues thatthe decline in real wages of less-skilled workersbetween 1967 and 1987 causedmost of the declinein employment of prime-age men over the sam-ple’s last 15 years. Similarly, Welch (1997) findsthat a shift in labor supply caused the LFPR ofprime-age men to decrease in the late 1960s andearly 1970s, but the decline in the subsequenttwo decades was caused by a change in relativewages (i.e., lower wages for less-educated com-pared with college-educated men). One mightthink that the rise in women’s participation rateswould be a contributing factor to the decline inmen’s participation rates; primarily, if a husbandhas a working wife, he has less incentive to be inthe labor force. However, Juhn andMurphy (1997)show that the evidence does not support thisclaim. Despite an increase in employment amongwives of low-wage men between 1969 and 1989,the change was much less than the increase inemployment among wives of middle- and high-wage men.

AGE AND THE LFPROne of the most important demographic

changes affecting the U.S. LFPR is the evolutionof the population’s age distribution. Most notice-ably, the approximately 78 million individualsbelonging to the baby-boom generation—thoseborn between 1946 and 1964—have been reachingthe latter stages of their working lives. With sucha large fraction of the U.S. population growingolder, the recent decline in the overall LFPR isunderstandable.

To get a sense of the influence the boomershave exerted on the LFPR, consider first thechange in the median age of the U.S. labor force.As the baby-boom cohort (representing roughlyone-third of the potential workforce) has grownolder, the median age of the U.S. labor force has



10 Cohany and Sok (2007) discuss changes in the LFPR of marriedwomenwith children of various ages. They show that of all marriedmothers, those with infants have experienced the largest declinein their LFPR since the late 1990s.

risen from less than 35 in 1980 to almost 41 in2005. As demonstrated by Figure 4, which showshow the median age of the labor force has changedacross decades going back to 1970 as well as pro-jections through 2050, the aging of the labor forceis expected to continue at least until 2020.

The importance of the baby-boom generationin explaining the recent trends in the LFPR canbe inferred from Figure 5, which highlights thedifferences in labor force participation across agegroups. In 2000, the baby boomers were 36 to 54years old, putting them in the prime-age workinggroup. Not surprisingly, this group had relativelyhigh participation rates that year: 91.6 percent ofmen and 76.7 percent of women had or activelysought employment (Toossi, 2005). However, after2000, baby boomers began moving into age cate-gories with typically lower LFPRs. In 2005, theage group 55 to 59 was composed entirely of babyboomers, and only 78 percent of these men and66 percent of these womenwere in the labor force.

Aaronson et al. (2006) estimate that about 95percent of the total decline in the LFPR between1995 and 2005—which was 0.44 percentagepoints—can be attributed to changing populationshares of the different age groups. The decline inthe population shares of those aged 25 to 34 and35 to 44 caused the LFPR to decrease by 0.57 and0.35 percentage points, respectively. The increasein the share of those aged 45 to 54, which wasmade up entirely of baby boomers, caused a rise inthe LFPR of 0.41 percentage points. The increasein those aged 55 to 64 put downward pressureon the LFPR, causing a 0.1 percent decline. Asbaby boomers begin to approach retirement, how-ever, further downward pressure will be exertedon the overall rate of labor force participation.According to Aaronson et al. (2006), the LFPRwill fall by 0.87 percentage points between 2005and 2010 as a result of the population being moreheavily concentrated among older age groups.They expect the increase in the shares aged 55 to



1970 1980 1990 2000 2010 2020 2030 2040 205034

35

36

37

38

39

40

41

42

Median Age

Figure 4

Median Age of the Labor Force

NOTE: The 1970 and 1980 data were obtained from Toossi (2002), and the remaining data were obtained from Toossi (2006).

64 and 65 and older to cause a total decline inthe LFPR of about half a percentage point.

There is, however, one especially interesting,countervailing trend that has partially offset thenatural decrease in the LFPR among an agingpopulation. Within the past two decades, therehas been a steady rise in the participation rateamong individuals 55 years of age and older (seeFigure 5). Just among those aged 55 to 64, theLFPR has increased by approximately 10 percent-age points over the past two decades.

A number of reasons may help explain theincrease in participation rates among olderworkers.11 First, the ability to draw full benefitsfrom Social Security depends on a person’s year

of birth; later generations must work longer toreceive full benefits. For example, full retirementoccurs at age 65 for individuals born in 1937 orearlier, age 66 for those born between 1943 and1954, and age 67 for individuals born in 1960 orlater. Furthermore, delaying retirement until age70 allows workers to be eligible for even higherbenefits. These features of the Social Securityprogram should push back the age at which someworkers exit the labor force (Social Security



11 Friedberg (2007) gives an overview of some possible explanationsfor the recent increase in delayed retirement (e.g., changes inSocial Security, Medicaid, and Medicare benefits and changes inpreferences).

Levels and TrendsPercent

16−

19

40

60

80

20−2

4

40

60

80

25−3

4

40

60

80

35−4

4

40

60

80

45−5

4

40

60

80

55−6

4

40

60

80

65an

dO

ver

1960 1980 200010203040

Cyclical ComponentsPercent

−202

−202

−202

−202

−202

−202

1960 1980 2000

−202

Population SharesPercent

10

20

10

20

10

20

10

20

10

20

10

20

1960 1980 2000

10

20

Figure 5

LFPR by Age Group

NOTE: Left column: levels (gray) and trends (blue) extracted with an LP96 filter. Middle column: cyclical components extracted withBP18,96 filters. Right column: population shares. The shaded areas denote NBER recessions.

Administration, 2006). Second, Social Securitybenefits have grown at a substantially slowerpace since the mid-1980s. Data from the SocialSecurity Administration show that real averagemonthly benefits rose by 88 percent between 1965and 1985 but by only 23 percent over the follow-ing 20 years. For 65 percent of the beneficiaries,Social Security benefits represent over half oftheir total income (Social Security Administration,2006). Hence, this decreased growth in benefitscould force some retirees back into the labor forceto help finance their retirement years. Third,Americans are now living longer than in previousdecades. For 65-year-old men, life expectancy hasrisen by nearly four years since 1970; for women,it has risen by three years. With greater numbersof both productive years in which they can workand “retirement” years that they must finance,individuals may decide to work longer. Fourth,older workers may choose to remain employedlonger to maintain health insurance coverage.Recent surveys (Kaiser Family Foundation andHealth Research and Educational Trust, 2006;Kaiser Family Foundation and Hewitt Associates,2006) have suggested that the fraction of firmsoffering their active workers the benefit of healthinsurance after they retire decreased by one-halfbetween 1988 and 2005 (Burtless, 2006). Becauseworkers, in general, do not qualify for Medicareuntil age 65, this development may also encour-age workers to delay retirement.

Along with the baby-boom generations,teenagers have also contributed to the recentdecline in labor force participation. Althoughthe teen LFPR has been trending downwardsince the 1970s, it experienced a sharper-than-

usual decline beginning in 2000. Over half of thedecline in the overall LFPR since then can beattributed to changes among those aged 16 to 19.Between 2000 and 2003, their LFPR dropped by7.5 percentage points—a much larger declinethan the 0.6-percentage-point drop in the overallLFPR. Since that time, teen participation rateshave yet to recover, and they remain around 44percent (Aaronson, Park, and Sullivan, 2006).

Once again, economists studying this down-ward trend have identified a number of possibleexplanations. Because teen workers have a weakattachment to the labor market, they are particu-larly sensitive to economic downturns. Conse-quently, when the U.S. economy entered its mostrecent recession, teen participation rates declinedsignificantly. However, Aaronson, Park, andSullivan (2006) argue that a weakened demandfor teen labor is unlikely to be the main source ofthe recent downturn, especially because there wasno simultaneous increase in the rate at whichteenagers reported that they sought employment.Instead, they argue that the failure of the teenLFPR to rebound within the first five years afterrecovery means that the decline is caused bysupply-side factors—namely, the decision toacquire more education.

The fraction of 16- to 19-year-olds who arecurrently enrolled in school has risen over thepast 20 years: from 61 percent in 1987 to 68 per-cent in 1997, and further to 73 percent by 2005.A large part of this rise can be linked to theincrease in the economic return to education,especially a college degree, since the late 1970s.There is also some evidence that the expansionof educational opportunities, particularly in the



Table 2Second Moments of the Business Cycle Components of LFPR by Age Group

16-19 20-24 25-34 35-44 45-54 55-64 65+

Corr(x,ip) 0.55 0.30 0.18 0.10 0.03 –0.14 0.10(0.42,0.66) (0.07,0.50) (–0.03,0.40) (–0.11,0.28) (–0.21,0.28) (–0.31,0.03) (–0.07,0.25)

Std(x)/Std(ip) 0.57 0.22 0.09 0.06 0.08 0.15 0.55(0.48,0.69) (0.18,0.27) (0.08,0.12) (0.05,0.08) (0.07,0.10) (0.12,0.20) (0.43,0.71)

NOTE: ip denotes industrial production. Block-bootstrapped 95 percent confidence intervals are in parentheses.

form of increased financial aid, has led to anincrease in college enrollment. A possible expla-nation for the recent larger-than-normal declinein the teen LFPR could stem from teenagers plac-ing even higher value on education than in thepast (Aaronson, Park, and Sullivan, 2006).

In addition to its influence on the long-runLFPR trends, the age distribution of the Americanworkforce can also influence the short-run fluc-tuations exhibited by the aggregate participationrate because there are substantial differences inthe cyclical properties of the LFPRs of variousage groups. In particular, the business cycle com-ponents for persons older than 20 are moderatelyprocyclical/acyclical (Table 2), while teen partici-pation is strongly procyclical. The volatilitiesare low for those between 25 and 54 years of age,but much higher for young and elderly workers.12

Thus, changes in the labor force’s age distributionmay lead to variations in how the LFPR respondsto business cycle conditions.

RACE, ETHNICITY, AND THE LFPRA third demographic feature influencing the

evolution of the aggregate LFPR is the increasein the racial and ethnic diversity of the U.S. pop-ulation over the past several decades. Whetherbecause of social, economic, or political factors,participation rates appear to vary across racialgroups. Figure 6 plots the LFPRs for white, black,and Hispanic men since the 1970s: Clearly,Hispanic men tend to have higher participationrates than either white men or black men. Overthe sample time frames, the average LFPR forHispanic men was 80.5 percent. White men andblack men averaged 76.8 percent and 70.1 percent,respectively.13 Similarly, the average yearly



12 The volatilities declined after 1984 by a factor of 1.6 to 1.8 forages 20 to 54 and by a factor of 1.3 for teenagers. For workers 55and over, the volatilities increased slightly in the post-1984 period.

13 The sample spans from January 1972 to November 2006 for whitesand blacks and June 1976 to November 2006 for Hispanics.

Percent

1975 1980 1985 1990 1995 2000 2005

60

65

70

75

80

85

90

White

Black

Hispanic

Figure 6

LFPR by Race/Ethnicity: Men

NOTE: The shaded areas denote NBER recessions.

growth rate among Hispanic men was higherthan for the other two groups. The LFPR of whitemen declined by 0.2 percent each year and thatof black men declined by 0.3 percent each yearon average, while the LFPR of Hispanic men hadzero growth on average. From the early 1970s to2000, the Hispanic share of the total populationincreased by more than 7 percentage points (to11.3 percent); it is therefore not surprising thatthe aggregate U.S. LFPR rose in the decades priorto 2000.

Among women, however, these three groupsshow the reverse ordering: Hispanics tend tohave the lowest participation rates while blackstend to have the highest. The average over theentire sample of the LFPR for white, black, andHispanic women was 54.7 percent, 57.0 percent,and 52.2 percent, respectively. The rise in thefraction of Hispanic women in the population,therefore, very likely had the opposite effect thatthe rise in the fraction of Hispanic men had: It

decreased the average LFPR. Still, participationrates among women of all racial groups showedgeneral increases between 1980 and 2000, andthese increases were similar across the threegroups (Fullerton and Toossi, 2001). Figure 7 plotsthe evolution of women’s labor force participa-tion broken down by race/ethnicity. In this case,white women experienced the highest averageyearly growth in their LFPR. It increased by 0.9percent each year on average, while Hispanicwomen saw their LFPR increase by 0.8 percentper year and black women saw theirs increasethe least, by 0.7 percent per year.

Once again, a number of explanations existfor differences in rates of labor force participationacross races.14 One of the most likely causes forthe higher LFPR of Hispanic men is that theytend to be younger than the general population

14 For a more complete overview of the black-white gaps in variouslabor force statistics, see Bradbury (2000).



Percent

1975 1980 1985 1990 1995 2000 2005

40

45

50

55

60

65

70

White

Black

Hispanic

Figure 7

LFPR by Race/Ethnicity: Women

NOTE: The shaded areas denote NBER recessions.

and in age groups that have higher LFPRs. Severalstudies cite the increased demand for skilledlabor beginning in the 1980s as a reason for somemen—especially less-skilled black men—to dropout of the labor force. Chandra (2000) shows thatin 1940, employment rates for white and blackmen were similar across education groups. How-ever, in 1990 the less-educated black men weremuch less likely to be employed than their whitecounterparts. Similarly, Bound and Holzer (1993)show that although industrial shifts frommanufac-turing to other sectors hurt wages for both whiteand black men, black employment (especiallyamong less-educated young blacks) declined themost during the 1970s, which also carried overinto the 1980s.

Although all women saw an increase in theirLFPR over this time period, black women saw amuch larger increase during the 1990s than theother groups. Juhn and Potter (2006) argue thatblack women were affected the most by changesin welfare and tax policy during that time, whichled to a rise in the LFPR of single mothers.

CONCLUSION: THE FUTURE OFLABOR FORCE PARTICIPATION

During the past half century, the U.S. LFPRhas seen dramatic changes, which have beendriven by the rise of women’s participation, anaging of the baby-boom generation, and growingethnic diversity within the general population.What does the future hold for U.S. labor forceparticipation? According to a report publishedby the Bureau of Labor Statistics, the overall LFPRis projected to decrease slightly to 65.6 percent in2014 (Toossi, 2005). Twomain factors are expectedto continue to exert downward pressure on theparticipation rate: the continued decline in theteen LFPR—which is projected to decline from43.9 percent in 2004 to 39.3 percent in 2014—and the aging of the baby-boom generation. Thissecond factor, however, is likely to lower aggregateparticipation rates for the next several decades.

As mentioned earlier, the baby boomers havealready begun entering into the 55-and-older age

category. In her BLS report, Toossi (2005) pro-jected that the fraction of Americans in this agegroup will rise from 28.4 percent of the adultpopulation today to 33.7 percent by 2014; theCensus Bureau projects this figure to be 39 per-cent by 2030. In contrast, the fraction of thepopulation in the prime-age working group isprojected to fall from 55.3 percent today to 51.1percent by 2014 and 47 percent by 2030.

As baby boomers enter successive age groups,their LFPR should fall dramatically. For instance,the 55 to 59 age group had an LFPR of 72 percentin 2006, and the 60 to 64 age group had an LFPRof approximately 53 percent. Among those 65and older, the LFPR was just over 15 percent.These numbers, coupled with the increasingproportion of the U.S. population beyond theirprime working age over the coming years, suggestthat successive generations will be unable to com-pensate for the baby boomers’ exit from the laborforce and U.S. labor supply will decline.

To be sure, participation rates for groups 55and older are expected to increase, which willpartially offset the downward pull that oldergroups have on the overall LFPR. In fact, there isalready some evidence of this following the 2001recession, when this age group had larger-than-normal increases in the LFPR (Bradbury, 2005).In 2014, approximately 41 percent of the groupis expected to be in the labor force, up from 38percent in 2006.

Still, most studies estimate that the rate oflabor force growth in the United States willdecrease over the next decade, if not longer (e.g.,Aaronson et al., 2006). In the event of such adrop-off, it may become increasingly difficult tomaintain growth in our standard of living becausethere will be fewer workers generating goods,services, and income for each resident in thecountry. The principal challenge in the presenceof a declining LFPR, therefore, will be to findwaysto enhance the productivity of the individualsthat do choose to work. Investing in education,physical capital accumulation, and research anddevelopment may be three avenues to such an end.



REFERENCESAaronson, Daniel; Park, Kyung-Hong and Sullivan,Daniel. “The Decline in Teen Labor ForceParticipation.” Federal Reserve Bank of ChicagoEconomic Perspectives, First Quarter 2006, 30(1),pp. 2-18.

Aaronson, Stephanie; Fallick, Bruce; Figura, Andrew;Pingle, Jonathan and Wascher, William. “The RecentDecline in the Labor Force Participation Rate andIts Implications for Potential Labor Supply.”Brookings Papers on Economic Activity, 2006, 0(1),pp. 69-134.

Baxter, Marianne and King, Robert G. “MeasuringBusiness Cycles: Approximate Band-Pass Filtersfor Economic Time Series.” Review of Economicsand Statistics, November 1999, 81(4), pp. 575-93.

Black, Dan; Kolesnikova, Natalia and Taylor, Lowell.“The Labor Supply of Married Women: Why DoesIt Differ Across U.S. Cities?” Federal Reserve Bankof St. Louis Working Paper 2007-043B, November2007.

Blau, Francine D. and Kahn, Lawrence M. “Changesin the Labor Supply Behavior of Married Women:1980-2000.” Journal of Labor Economics, July 2007,25(3), pp. 393-438.

Bound, John and Holzer, Harry J. “Industrial Shifts,Skills Levels, and the Labor Market for White andBlack Males.” Review of Economics and Statistics,August 1993, 75(3), pp. 387-96.

Bradbury, Katharine L. “Rising Tide in the LaborMarket: To What Degree Do Expansions Benefit theDisadvantaged?” Federal Reserve Bank of BostonNew England Economic Review, May/June 2000,pp. 3-33.

Bradbury, Katharine. “Additional Slack in theEconomy: The Poor Recovery in Labor ForceParticipation During This Business Cycle.” PublicPolicy Briefs No. 05-2, Federal Reserve Bank ofBoston, July 2005.

Burtless, Gary. “The Recent Decline in the LaborForce Participation Rate and Its Implications forPotential Labor Supply: Comment.” BrookingsPapers on Economic Activity, 2006, 0(1), pp. 135-43.

Chandra, Amitabh. “Labor-Market Dropouts and theRacial Wage Gap: 1940-1990.” AEA Papers andProceedings, May 2000, 90(2), pp. 333-38.

Christiano, Lawrence J. and Fitzgerald, Terry J. “TheBand Pass Filter.” International Economic Review,May 2003, 44(2), pp. 435-65.

Clark, Todd E. and Nakata, Taisuke. “The TrendGrowth Rate of Employment: Past, Present, andFuture.” Federal Reserve Bank of Kansas CityEconomic Review, First Quarter 2006, 91(1),pp. 43-85.

Cohany, Sharon R. and Sok, Emy. “Trends in LaborForce Participation of Married Mothers of Infants.”Monthly Labor Review, February 2007, 130(2),pp. 9-16.

Fernández, Raquel; Fogli, Alessandra and Olivetti,Claudia. “Mothers and Sons: Preference Formationand Female Labor Force Dynamics.” QuarterlyJournal of Economics, November 2004, 119(4),pp. 1249-99.

Fogli, Alessandra; Marcassa, Stefania and Veldkamp,Laura. “The Spatial Diffusion of Female LaborForce Participation: Evidence from US Counties.”Unpublished manuscript, New York University,2007.

Fogli, Alessandra and Veldkamp, Laura. “Nature orNurture? Learning and Female Labor ForceDynamics.” Federal Reserve Bank of Minneapolis,Staff Report 386, February 2007.

Friedberg, Leora. “The Recent Trend Towards LaterRetirement.” Center for Retirement Research, WorkOpportunities for Older Americans, March 2007, 9,pp. 1-7.

Fullerton, Jr., Howard N. and Toossi, Mitra. “LaborForce Projections to 2010: Steady Growth andChanging Composition.” Monthly Labor Review,November 2001, 124(11), pp. 21-38.

Goldin, Claudia. “The Quiet Revolution ThatTransformed Women’s Employment, Education,and Family.” American Economic Review, May2006, 96(2), pp. 1-21.



Goldin, Claudia and Katz, Lawrence F. “The Powerof the Pill: Oral Contraceptives and Women’s Careerand Marriage Decisions.” Journal of PoliticalEconomy, August 2002, 110(4), pp. 730-70.

Greenwood, Jeremy; Seshadri, Ananth and Yorukoglu,Mehmet. “Engines of Liberation.” Review ofEconomic Studies, January 2005, 72(1), pp. 109-33.

Hotchkiss, Julie L. “Employment Growth and LaborForce Participation: How Many Jobs Are Enough?”Federal Reserve Bank of Atlanta Economic Review,First Quarter 2005, 90(1), pp. 1-13.

Juhn, Chinhui. “Decline of Male Labor MarketParticipation: The Role of Declining MarketOpportunities.” Quarterly Journal of Economics,February 1992, 107(1), pp. 79-121.

Juhn, Chinhui and Murphy, Kevin M. “WageInequality and Family Labor Supply.” Journal ofLabor Economics, January 1997, 15(1), Part 1,pp. 72-97.

Juhn, Chinhui and Potter, Simon. “Changes in LaborForce Participation in the United States.” Journalof Economic Perspectives, Summer 2006, 20(3),pp. 27-46.

Kaiser Family Foundation and Health Research andEducational Trust. Employer Health Benefits: 2006Annual Survey. Publication No. 7527, Kaiser FamilyFoundation, September 2006.

Kaiser Family Foundation and Hewitt Associates.Retiree Health Benefits Examined. Publication No.7587, Kaiser Family Foundation, December 2006.

Social Security Administration. Annual StatisticalSupplement to the Social Security Bulletin, 2005.SSA Publication No. 13-11700, February 2006.

Stock, James H. and Watson, Mark W. “Has theBusiness Cycle Changed and Why?” in MarkGertler and Kenneth Rogoff, eds., NBERMacroeconomics Annual 2002. Boston, MA: MITPress, 2003.

Toossi, Mitra. “A Century of Change: The U.S. LaborForce, 1950-2050.” Monthly Labor Review, May2002, 125(5), pp. 15-28.

Toossi, Mitra. “Labor Force Projections to 2014:Retiring Boomers.” Monthly Labor Review,November 2005, 128(11), pp. 25-44.

Toossi, Mitra. “A New Look at Long-Term LaborForce Projections to 2050.” Monthly Labor Review,November 2006, 129(11), pp. 19-39.

Welch, Finis. “Wages and Participation.” Journal ofLabor Economics, January 1997, 15(1), Part 2,pp. S77-S103.



APPENDIX

Trend and Business Cycle Components

Any time series can be decomposed into cyclical components of different frequencies. The fre-quency of a cycle is inversely related to its period. The period of a cycle is simply the time betweensubsequent peaks. We consider three components:

Trend component: The trend is obtained by removing fluctuations with periods higher than 8 years(i.e., 96 months) with a low-pass filter.

Business cycle component15: We extract the business cycle component of a time series using a band-pass filter, which removes the trend component (period higher than 8 years) and the high-frequencycomponent (period less than 1.5 years).

High-frequency component: This corresponds to fluctuations with periods less than 1.5 years.

Figure A1 illustrates this decomposition for industrial production. We use the business cyclecomponent of industrial production to determine the correlation of participation rates for variousdemographic groups with aggregate economic activity over the business cycle (see Tables 1 and 2).



15 We sometimes refer to the business cycle component as the cyclical component with a slight abuse of terminology.

Level

1950 1960 1970 1980 1990 2000

20

40

60

80

100

Trend

1950 1960 1970 1980 1990 2000

20

40

60

80

100

Business Cycle Component

1950 1960 1970 1980 1990 2000

−4

−2

0

2

4High Frequency Component

1950 1960 1970 1980 1990 2000

−1.5

−1

−0.5

0

0.5

1

1.5

Figure A1

Industrial Production

NOTE: Industrial production data are from the Federal Reserve Board. The shaded areas denote NBER recessions.

RE

VIE

WJANUARY/FE

BRUARY2008

•VOLU

ME90,

NUMBER1

Federal Reserve Bank of St. LouisP.O. Box 442St. Louis, MO 63166-0442

Jan Feb 2008 Review

Documents

armendriz de aghion

ua ry

bank rakyat indonesia

kaiser family foundation

federal reserve bank

federal reserve board

prior written permission

good faith fund