governance and performance of microfinance institutions in central and eastern europe and the newly...

World Development Vol. 33, No. 10, pp. 1627–1643, 2005� 2005 Elsevier Ltd. All rights reserved

Printed in Great Britain
0305-750X/$ - see front matter
doi:10.1016/j.worlddev.2005.06.001www.elsevier.com/locate/worlddev

Governance and Performance of Microfinance

Institutions in Central and Eastern Europe and the

Newly Independent States

VALENTINA HARTARSKA *

Auburn University, AL, USA

Summary. — This paper presents the first evidence on the impact of governance on outreach andsustainability of microfinance institutions (MFIs) in Central and Eastern Europe and the NewlyIndependent States. The results indicate that performance-based compensation of managers isnot associated with better-performing MFIs; lower wages suggested for mission-driven organiza-tions worsen outreach, while managers’ experience improves performance. The results also identifytradeoffs between MFI outreach and sustainability depending on stakeholders’ representation onthe board and provide strong support for independent boards with limited employee participation.In the study region, external governance mechanisms seemed to play a limited role.
� 2005 Elsevier Ltd. All rights reserved.
Key words — microfinance, governance, Central and Eastern Europe and the Newly IndependentStates, board, rating, audit, regulation

* The author would like to thank the Microfinance

Center for Central and Eastern Europe and the Newly

Independent States and Justyna Pytkovska and Kasia

Pawlak, in particular, who provided most of the data

that made this study possible. Thanks to Tatyana

Lysenko and the Office of the Chief Economist at the

European Bank for Reconstruction and Development

who provided the latest institutional indexes. This paper

benefited from comments by participants at the 5th

Annual Microfinance Meeting in Budapest, 2002. Also,

thanks to members of the Rural Finance Team at The

Ohio State University who commented on the 2002

governance survey. Special thanks to Bob Nelson for his

comments and encouragement. Final revision accepted:June 8, 2005.

1. INTRODUCTION

Microfinance is the provision of loans andother financial services to the poor. The micro-finance institution (MFI) has evolved as a resultof the efforts of committed individuals andassistance agencies to reduce poverty by pro-moting self-employment and entrepreneurship.The MFI faces unique challenges because itmust achieve a double bottom line—providefinancial services to the poor (outreach) andcover its costs (sustainability). Microfinance isa significant and growing industry, yet fewstudies explore the link between governanceand performance. Previous studies focused onthe main strengths of MFIs, namely, howinnovative lending technologies enabledlending to the poor, and how microfinanceaffects borrowers’ welfare (Conning, 1999;Navajas, Schreiner, Meyer, Gonzalez-Vega, &Rodriguez-Meza, 2000).Microfinance practitioners have recognized

that good governance is critical for the successof MFIs (Campion, 1998; Rock, Otero, &Saltzman, 1998), but only few studies on regu-lation in microfinance have touched upongovernance issues (McGuire, 1999). Closer

162

examination of the role of various governancemechanisms is important because MFI manag-ers control significant resources. In Central andEastern Europe and the Newly IndependentStates (CEE & NIS), the asset base of theseorganizations is estimated to be 1.2 billion dol-lars (Foster, Green, & Pytkowska, 2003).While the impact of property rights and gov-

ernment behavior on institutional developmentin postcommunist countries has attracted

7

1628 WORLD DEVELOPMENT

attention, the performance of nongovernmentalorganizations (NGOs) and their contributionto institutional development is less understood.As the nonprofit sector in these countries growsand becomes important in microfinance andother areas, understanding what governancemechanisms strengthen these organizations be-comes important.The lack of studies on the effect of MFI gov-

ernance on performance is to a large extent dueto scarcity of data. Performance data are con-sidered proprietary and are hard to obtain.Although the majority of MFIs are funded withpublic funds channeled through large interna-tional development agencies, until recently thepractice was to withhold performance informa-tion from the general public. 1

In addition, the microfinance industry isquite diverse in terms of organizational types,with MFIs organized as NGOs, banks, creditcooperatives, or nonbank financial institutions.This diversity makes it difficult to choose anappropriate conceptual framework for theanalysis of MFI governance. However, a 1998survey of the microfinance industry shows thatthe objectives and performance of MFIs orga-nized under different legal forms do not differsubstantially (Campion, 1998). 2 Therefore, toidentify the impact of various governancemechanisms, the empirical analysis will use in-sights from the governance literature on for-profit firms, nonprofit firms, and banks thatare relevant to MFIs.This paper uses unique data from the recently

conducted surveys in CEE & NIS to study therelationships between MFI performance andgovernance. The empirical model explores thejoint and individual influence of managementcompensation, board effectiveness and diver-sity, and external governance mechanisms onboth MFI sustainability, and the depth andbreadth of outreach while controlling for indi-vidual MFI characteristics, as well as country-specific institutional and macroeconomicfactors. Results indicate that while perfor-mance-based compensation does not improveperformance, underpaying managers diminishesoutreach. More independent boards are moreeffective, but the differences in emphasis on out-reach and sustainability by various stakeholdersrepresented on the board seem to have differen-tial impact on performance (e.g., MFIs withmore donor representatives have better out-reach but worse sustainability). External gover-nance mechanisms such as auditing, rating, andregulation, have a limited impact.

The rest of the paper is organized as follows.Section 2 presents theoretical considerations,Section 3 describes the data and the empiricalmodel, Section 4 discusses the results, and Sec-tion 5 offers conclusions.

2. THEORETICAL CONSIDERATIONS

In microfinance, governance refers to themechanisms through which donors, equityinvestors, and other providers of funds ensurethemselves that their funds will be used accord-ing to the intended purposes. 3 Such controlmechanisms are necessary because managersand providers of funds may have divergingpreferences and objectives. For example, MFImanagers may work toward fulfilling the mis-sion of the MFI, but they may also have prefer-ences for nonpecuniary rewards. In thecorporate governance literature, this problemis known as the agency problem. This literaturerefers to the manager as an Agent, who unlike aPrincipal, does not own the resources of thefirm. The Principal bears the residual risk andis the residual claimant of the firm’s wealth(Jensen & Meckling, 1976). Costs associatedwith this problem are called agency costs andrepresent costs that residual claimants bear inorder to benefit from the professional servicesof managers. The goal of many governancemechanisms is to minimize agency costs byaligning the objectives of the owner-Principalwith the objectives of the manager-Agent.The key mechanisms of an effective gover-

nance framework are ownership structure(including institutional and managerial owner-ship), CEO (manager) and director (boardmember) remuneration, board structure (sizeand composition), auditing, information disclo-sure, and the market for corporate control(Keasey, Thompson, & Write, 1997). Typically,governance studies focus on the individual im-pact of one of these mechanisms (e.g., compen-sation, board size, independence and diversity,and external market forces, Shleifer & Vishny,1997). Some recent studies have recognized thatsome of these mechanisms may act to comple-ment each other and may be correlated witheach other (Hermalin & Weisbach, 2003). Thispaper explores the impact of all mechanisms ex-cept ownership, because the database does notcontain ownership data.MFIs have some unique characteristics that

complicate the study of their governance. Forexample, apart from financial sustainability,

GOVERNANCE AND PERFORMANCE OF MICROFINANCE INSTITUTIONS 1629

MFIs need to fulfill an outreach mission byserving poor clients, and many operate asNGOs, which makes them similar to nonprofitfirms. Many MFIs are similar to banks becausethey are regulated or supervised by a regulatorybody and/or because they collect deposits. Theorganizational diversity of MFIs makes theempirical study of their governance difficult.This challenge is addressed by formulatingand testing hypotheses based on insights fromthe literature on corporate governance, gover-nance in banks, and in nonprofit organizations,and by estimating the impact of the governancemechanisms on both sustainability and out-reach.A focus on both outreach and sustainability

is necessary because the debate on whether out-reach and sustainability are substitutes or com-plements is ongoing (Morduch, 2000; Navajaset al., 2000). There is no evidence that the twodimensions of performance in every MFI inthe region are substitutes or complements.Therefore, estimating the impact of governancemechanisms on both dimensions may provideinsights into possible tradeoffs between out-reach and sustainability.

(a) Managerial compensation as anincentive-aligning mechanism

According to the agency literature, compen-sation that includes both a performance-basedelement and a fixed element is the best mecha-nism to align the interests of managers withthat of equity holders and donors. The empiri-cal literature on corporations confirms thatthere is a nonlinear pay–performance link, butthe sensitivity is relatively small. 4

Since banking is a regulated industry, regula-tion may substitute for or complementincentive features in managerial contracts(John, Mehran, & Qian, 2004). High-poweredincentives (remuneration where the bonus partis very sensitive to firm performance) may alignthe interests of managers too much with thoseof equity holders, and induce managers to takehigher risks at the expense of depositors, whowould suffer most if the MFI fails; thus, lowpay–performance sensitivity is recommended(John & John, 1993). Indeed, pay–performancesensitivity in banking is lower than that in otherindustries (Adams & Mehran, 2003; Houston &James, 1995; John & Qian, 2003).In nonprofit firms, many forms of incentive

pay are illegal. In fact, it has been shown thatthe asymmetric information between clients

and managers (i.e., managers know more aboutthe product than clients) makes fixed manage-ment salaries the better choice for mission-dri-ven organizations (Easley & O’Hara, 1988).Specifically, since managers get fixed salaries,they are indifferent between telling the truthand lying, and thus will tell the truth. Clientsand donors will find the information providedby nonprofit managers more credible, and thiswill lead to better-funded and better-perform-ing firms.In spite of the argument against heavier reli-

ance on performance-based remuneration infinancial and NGO firms, the trend towardcommercialization of microfinance suggestswider use of bonuses. This recommendation isbased on positive experience with the impactof bonuses on loan officers’ productivity docu-mented, for example, in the microfinance banksestablished by the Internationale Projekt Con-sult (Holtmann, 2002). This type of compensa-tion, nevertheless, is less common for MFImanagers who usually receive a fixed salarypaid under a contract for technical assistance.Since 23% of the managers in the sample re-ceived performance-based bonuses during thestudy period, the recommendations of the the-ory can be tested against the recommendationsof commercialization proponents.

Hypothesis 1. Since MFIs are similar to banksand nonprofit firms, and less similar to com-mercial companies, performance-based com-pensation may not be effective. MFIs wheremanagers receive a fixed salary will not performworse than MFIs where managers receiveperformance-based remuneration.

Another part of the nonprofit literature sug-gests that, instead of offering performance-basedcompensation, nonprofit boards may be able torecruit managers by offering compensationpackages combining lower wages with some per-quisites that only individuals committed to theorganization’s mission will self-select to take(Handy & Katz, 1998). Additionally, the appealof a position of power in nonprofit firms may besufficient to attract good managers (James,1983). Theoretical studies also show that ifwagespaid to NGOmanagers are similar to those paidto for-profit managers, and if the NGO technol-ogy is superior to that of the for-profit firm, theNGOs will be the dominant organizational formin the industry (Scott & Hopkins, 1999). 5

Although many microfinance analysts areadvocating transformation of microfinance


NGOs into regulated financial institutions, themajority of MFIs in CEE & NIS continue tooperate asNGOs. Thus, it is important to under-stand how the level of compensation of manag-ers may affect MFI performance.

Hypothesis 2. MFIs where managers feel theyare paid wages lower than what they couldreceive in alternative employment are as effec-tive as MFIs whose managers feel they areadequately compensated.

(b) The MFI board

The board of directors is an internal gover-nance mechanism that helps resolve the agencyproblems between Principals and Agents.Board member incentives are aligned with thatof the Principals because of the provision thatthe board can be held legally responsible forfailing to perform effective monitoring. In addi-tion, board members offer their reputations ascollateral to the public and will try to minimizetheir own risk of losing their reputations(Handy, 1995). Although directors may haveconsiderable incentives to slack off or get alongwith managers, peer policing decreases theincidence of inappropriate behavior (Fama &Jensen, 1983a; Holmstrom, 1999). Board mem-bers no longer committed to the mission leave,and substitution is done by the remainingboard members based on mutually agreed uponcriteria (Fama & Jensen, 1983b).In the empirical corporate governance litera-

ture, the quality of monitoring by the board ismeasured by the proportion of independentdirectors (Bhagat & Jefferie, 2002). Indepen-dent directors (nonemployees, not related tothe company) are expected to act as bettermonitors and advisors. Empirical studies havefound both positive and negative relationshipsbetween the proportion of outside directorsand firm value (Hermalin & Weisbach, 2003).Banks have larger proportions of outside

directors and since MFIs are similar to banksthey may share this characteristic (Adams &Mehran, 2003). Nonprofit boards are typicallycomprised of outsiders, and the potential con-flict between insiders and outsiders is rarelyexplored. 6

The microfinance board consists of threemajor groups: (1) independent (unaffiliated)directors; (2) insiders who have a monetary

interest in the firm, such as employees; and(3) representatives of donors, investors, credi-tors, and clients, who do not receive compensa-tion but who represent the diverse interests ofmicrofinance stakeholders. 7 Since the thirdgroup of directors is different from the groupof paid employees and unaffiliated directors, itwill be separated in the analysis and the studywill evaluate the role of directors–employees(insiders), outsiders and, implicitly, the role of‘‘gray directors.’’ 8

Hypothesis 3. MFI performance will be af-fected negatively by the proportion of emplo-yees on the board and positively by theproportion of nonaffiliated outsiders.

Board efficacy can be influenced by boardsize, with larger boards being less effective thansmaller boards because when the board gets toobig, free riding by some directors may becomean issue (Jensen, 1993). This hypothesis is con-firmed by studies of both large corporateboards and boards of small firms (Eisenberg,Sundgren, & Wells, 1998; Yermack, 1996).Compared to other organizations, financialintermediaries have larger boards, although lar-ger bank boards are less effective monitors(Adams & Mehran, 2003).The impact of board size on the perfor-

mance of nonprofit firms is not clear. Osterand O’Regan (2003) put forward the hypothesisthat in these organizations board size may needto be larger due to the extra duties of boardmembers related to supervision of fundraising.However, these authors did not find empiricalevidence to support their hypothesis. Basedon these considerations, the following hypothe-sis is formulated:

Hypothesis 4. MFIs with smaller boardsachieve better results than MFIs with largerboards.

Board diversity in terms of women andminority representation is another issue thathas attracted attention. Board diversity is desir-able because it is fair (equity considerations),and because evidence suggests that diversityimproves firm performance and shareholderwealth (Carter, Simkins, & Simpson, 2003;Westphal & Milton, 2000).Traditionally, women have been under-repre-

sented on corporate boards, especially in bank-


ing. The impact of board composition on theperformance of financial firms is especiallyimportant because of the relatively limited re-search in this area (Macey & O’Hara, 2003).Nonprofit boards, in contrast, are very diverse.For this reason, better-performing nonprofitfirms do not necessarily have proportionallymore women and minorities on the boards,although evidence shows that women directorsspend more time on monitoring activities(Oster & O’Regan, 2003).The skills that board members bring to the

board may also matter. The occupation ofboard members in nonprofit firms has not beenfound to affect fundraising, but it does affect thetime spent on monitoring (Oster & O’Regan,2003). The unique nature of microfinance activ-ities requires that board members have financialand banking skills as well as social services expe-rience (Campion, 1998). In addition, sincemicrofinance boards include representatives ofdonors, equity investors, and creditors (who of-ten provide a significant amount of the fund-ing), the mix of board member skills mayaffect the ability of various groups to reach con-sensus in an efficient manner.An important debate in microfinance is

whether clients should be represented on theboard. Advocates argue that clients will helpprovide better information on the target clien-tele and thus improve performance while oppo-nents argue that clients on the board mayweaken the organization (Campion, 1998).The impact of stakeholder representation andof board member skills on MFI performancewill be tested via Hypothesis 5.

Hypothesis 5. The higher the proportion ofstakeholder representatives and the moresophisticated their skills (e.g., financial liter-acy), the better the performance in the areaimportant to these stakeholders (e.g., sustain-ability).

(c) External governance mechanisms

The manager of a corporation is disciplinedby market forces through the market for man-agers and through the market for takeovers.These market forces have a limited role inmicrofinance because the market for MFI man-agers is thin and because most MFIs do nothave true owners. As the microfinance industrygrows and matures, the competition for dona-

tions and customers has shifted the focustoward external governance mechanisms.Donors and creditors are increasingly relyingon information from audited financial state-ments and rating agencies, as well as on infor-mation disclosed under rules imposed byregulators. The main objective of such externalgovernance mechanisms is to reduce informa-tion asymmetries between the different stake-holders and the firm (Healy & Palepu, 2001).The impact of audit on firm performance has

not received much attention, perhaps becauseof widespread mandatory audits. Audits andaudit stringency are especially important inbanking because evidence shows that countrieswith greater audit stringency are less vulnerableto bank crises (Tadesse, 2003). In addition,empirical evidence suggests that firms’ demandfor quality auditors is driven by active stake-holders (Ashbaugh & Warfield, 2003). Inmicrofinance, auditing is not mandatory, andit can serve as an effective external governancemechanism by ensuring potential investorsand donors that an MFI complies with theaccounting practices and managers do not mis-represent financial information. Thus, the nexthypothesis is as follows:

Hypothesis 6. MFIs that have their financialstatements audited achieve better results thanMFIs that do not have their financial state-ments audited.

In the absence of developed equity and debtmarkets, donors and investors could bene-fit from independent evaluation of MFIs’ per-formance. Rating can help impose marketdiscipline by revealing new information andthus encourage better management. 9 In recentyears, rating agencies specializing in rating ofMFIs appeared in an attempt to fill in the needfor better information disclosure. Unlike typi-cal rating agencies that rate the riskiness of is-sued debt, microfinance rating agencies ratethe overall performance of the MFI in termsof outreach and sustainability. While ratingwas just emerging during the study period, theneed for and the impact of this industry is cur-rently being reconsidered. The debate will behelped by a better understanding of the impactof rating.

Hypothesis 7. Rated MFIs achieve better re-sults than MFIs whose performance is notrated.


Some MFIs previously operating as NGOshave raised funds, reorganized, and trans-formed into regulated financial intermediaries,and even more MFIs are considering such atransformation mainly because of access tocommercial funds (Campion & White, 1999).Regulatory involvement may impact MFI per-formance by changing the internal rules of theorganization. Prudential regulation is imposedon organizations that accept deposits becausedepositors own part of the resources and arethus Principals. Since small and disperseddepositors are unable to protect their interests,an external agency can restore (ex post) effi-ciency by creating rules that would guaranteethe safety of deposits (Dewatripont & Tirole,1994). Thus, deposit-taking institutions shouldbe regulated, those without deposits from thepublic should not, and MFIs who fall in be-tween should have some form of targeted regu-lation (Hardy, Holden, & Prokopenko, 2003;Van Greuning, Galardo, & Randhawa, 1999).Since regulation introduces the regulator as

an additional stakeholder in the governancestructure of the MFI, microfinance profession-als worry about the impact of the new stake-holder on the mission (Dichter, 1997).Regulatory involvement may lead to ‘‘mis-sion-drift’’ if demands to fulfill regulatoryrequirements divert attention away from serv-ing the poor, and may hold back innovationin lending technology that has been the drivingforce behind MFIs’ ability to serve even poorerborrowers. Although recent evidence does notsupport this hypothesis, it is important to deter-mine if regulation in the region influences MFIperformance (Hartarska, in press).

Hypothesis 8. Regulated MFIs may have bet-ter sustainability but may not achieve betteroutreach.

3. DATA AND EMPIRICAL MODEL

Data for this study came from three surveys.The first survey was conducted in 1998 by a re-gional network organization, the MicrofinanceCenter for Central Europe and the Newly Inde-pendent States (MFC). The survey collecteddata on MFI boards, their governance and per-formance. The second survey was conducted in2001 by the same regional network. In this sur-vey, MFIs reported their performance and

organizational characteristics for the period1998–2001. Since 2000, many MFIs have beensending annual reports to the MFC and theirprofiles were updated with 2002 data by thisorganization’s staff. The third survey focusedspecifically on governance and was conductedin 2002 by the author in cooperation with theMFC. The data on MFI performance, boardcharacteristics, and mechanisms of externalcontrol were used to develop the database. 10

Of the 140 organizations (including credit un-ions and cooperatives) that participated in the2001 survey, only 71 were independent organi-zations with their own board of directors. MFIswith boards were contacted in 2002 and askedto complete a second survey with detailed ques-tions on governance. The response rate wasnearly 50%, with 34 organizations completingthe survey. Credit unions and cooperatives (24organizations) were excluded from the databasebecause they have distinctively different gover-nance. 11 Lending technology data collectedby MFC were also added.Since the resulting data come from four dif-

ferent sources, and the survey data come fromsurveys with a different focus (e.g., the surveyconducted in 2002 collected only governanceand not performance data), sample selectionbias is minimized. Therefore, it seems reason-able to assume that the resulting data constituteseveral random samples of data on governanceand performance for MFIs in the region. TheMFIs included in the regressions represent be-tween a third and a quarter of the MFIs inthe region.The empirical model used to test the hypoth-

eses follows that of Molyneux, Lloyd-Williams,and Thornton (1992) and Barth, Nolle, Phu-miwasana, and Yago (2003) and is of the form

P ijt ¼ constant þ a0Sijt þ b0Mijt þ /0Bijt

þ c0EGijt�1 þ d0MIjt þ eijt;

where Pijt is a performance variable for MFI iin country j at time t; Sijt are MFI-specific vari-ables; Mijt are management specific variables;Bijt are board-specific variables, EGijt�1 areexternal governance mechanisms; and MIjt arethe country-specific macroeconomic and insti-tutional variables. A single-equation model isproposed because in spite of theoretical sugges-tions to the contrary, the hypothesis that vari-ous governance mechanisms are endogenouslydetermined is not always supported by empiri-cal evidence. 12 Random effects estimation is


used in order to accommodate the impact oftime invariant explanatory variables such asMFI type, regulatory status, and lending tech-nology.In microfinance, performance is measured by

accounting-based indicators of outreach andsustainability. 13 In general, accounting mea-sures are considered more appropriate forlong-term studies because although managersmay be able to manipulate financial statementsfor a year, their ability to manipulate state-ments for longer periods is limited (Bhagat &Jefferie, 2002).Variables used in the regression analysis are

defined in Table 1. MFI performance is mea-sured in terms of sustainability and outreach.Outreach, in turn, is measured in two dimen-sions—breadth and depth. 14 Breadth of out-reach is measured by the logarithm of thenumber of active borrowers. Depth of outreachis measured by the variable DEPTH, which isthe average loan size divided by the annualGDP per capita, all in $US. 15 The higher thevalue of DEPTH, the less poor clients are beingserved. Therefore, from a poverty-alleviationperspective, a smaller value of this variable ispreferred.Sustainability is measured by return on assets

(ROA), and by operational self-sufficiency(OSS). Operational self-sufficiency measureshow well the MFI can cover its costs throughoperating revenues. The surveys asked MFIsto provide financial data adjusted for inflation,cost of capital, and in-kind donations, using thetypical industry standards. Although an at-tempt was made to double check the data andinclude only adjusted data, data from the sur-veys remain self-reported, and some MFIsmay have misrepresented their financial infor-mation.Variables included in Bijt are Fixed_wage,

which is a dummy for pay not based on perfor-mance, Lower_wage, which is a dummy thattakes the value of one if the manager stated thathis salary is lower than what he could com-mand in an alternative employment, and Expe-rience is used to proxy for a manager’s qualityand is measured by the years of work experi-ence.MFI specific variables (Sijt) are MFI size

measured by the logarithm of total assets,MFI age measured in years since inception,and MFI type measured by three dummies—NGO, Nonbank Financial Institution, andBank—with the latter being the omitted dum-my in the regression analysis. It is important

to control for lending technologies becausestudies show that the type of lending technol-ogy used influences the success of these organi-zation (Navajas, Conning, & Gonzalez-Vega,2003). Thus, the analysis included a variableIndividual, which is a dummy that takes thevalue of one if the MFI used individual lendingtechnology.Variables included in Bijt are board size

(Board_size), measured by the number of boardmembers; Employees measured as the propor-tion of MFI employees who are voting boardmembers (usually the managers); Independentmeasured as the proportion of nonaffiliatedboard members who do not represent otherstakeholders’ interest from the 2002 survey.From the 2001 survey, Women measured asthe proportion of women on the board, Donorsis the proportion of donors, Financiers is theproportion of members with financial skills,and Clients is the proportion of clients. The cat-egories in the 2001 survey overlap; that is, thesum of their proportion is greater than 100%because members could be classified into sev-eral categories (e.g., women and donors).Among the external control variables

(EGijt�1), audit and rating are lagged one peri-od because their impact is likely to be delayed.This specification avoids cases where an MFImay have requested an audit only after a goodyear. 16 The variables included in EGijt�1 comefrom the 2001 survey and include Supervised,which is a dummy that takes the value of oneif the MFI was supervised by the Central Bankor other bank supervisory agency; 17 Rated is avariable that indicates whether the MFI wassubject to independent evaluation or rating byan outside organization; Audited is a dummythat takes the value of one if there was an au-dited financial statement in the year t � 1. 18

Differences in economic conditions acrosscountries are controlled for by the size of theeconomy (Economy Size), measured by the log-arithm of a country’s GDP, and by the averageinflation rate (Inflation), measured by the aver-age consumer price index. These variables comefrom the World Bank Development Indicators.Differences in institutional development acrosscountries are captured by three indexes createdby the European Bank for Reconstruction andDevelopment (EBRD Transition Report,2003). 19 These indexes approximate thelevel of banking sector reform (BSR), regula-tions that promote competition (CP), andinfrastructure reform (IR). The first index cap-tures the level of regulation of MFIs and their

Table 1. Definition of variables

Variable Definition

ROA Return on assets; measures how well the MFI uses its total assets to generate returns; since self-reported may not be adjusted forgrants and donations

OSS Operational self-sufficiency = operating revenue/(financial expense + loan loss provision + operating expense). Measures howwell the MFI can cover its costs through operating revenues

NAB Logarithm of the number of current borrowers, that is the number of individuals that currently have an outstanding loan balancewith the MFI or are responsible for repaying any portion of the gross loan portfolio

DEPTH Average outstanding loans size/GDP per capita in $US. Higher values mean that the MFI serves richer borrowersFixed_wage A dummy that equals one if the manager receives a fixed salary, zero otherwiseLower_wage A dummy that equals one if the manager estimated that he is paid less than what he could get at a similar jobExperience The number of years of experience of the managerBoard_size Number of board membersEmployees The proportion of voting board members that are also employees of the MFIIndependent The proportion of board members who do not have an affiliation with any of the stakeholders of the MFIWomen The proportion of women on the boardDonor The proportion of board members who represent donors or grant-giving organizationsFinanciers The proportion of board members with financial skillsLocal_businessmen The proportion of board members who are local businessmenClients The proportion of clients on the boardAudited A dummy that equals one if the financial statement of the MFI is audited and zero otherwiseRated A dummy that equals one if the MFI is rated by a specialized MFI rating agency and zero otherwiseSupervised A dummy that equals one if the MFI is regulated/supervised by a government regulatory agency and zero otherwiseIndividual A dummy that equals one if the MFI used individual lending technology and zero otherwiseMFI age Number of years since inceptionMFI size Logarithm of the total assets of the MFI. Total assets include all assets net of contra asset accounts such as the loan loss reserve

and accumulated depreciationNGO The MFI is an NGONBFI The MFI is a nonbank financial institutionInflation Average annualized consumer price indexEconomy size Logarithm of the total GDP (gross domestic product of the country) for year tIR Index of infrastructure reform; higher values indicate better infrastructure, varies from 1 to 6CP Index of competition policies: higher values indicate better policies, varies from 1 to 6BSR Index of the banking sector reform; varies from 1 to 6; higher values indicate higher level of development

1634WORLD

DEVELOPMENT

Table 2. Summary statistics

Variable Mean Std. dev.

ROA 3.038 29.29OSS 91.99 45.38NAB 7,268 64,943


competitors. The last two indexes affect clients.If clients operate in a repressive environmentwith poor infrastructure and under regulatoryconditions that stifle competition, then such cli-ents will be more difficult to serve in a profit-able manner, ceteris paribus.

DEPTH 425 134Lower wage 0.324 0.469Fixed wage 0.768 0.424Experience 14.338 7.718Board_size 6.09 2.258Women 0.232 0.230Donor 0.183 0.301Financiers 0.209 0.274Local_businessmen 0.127 0.177Clients 0.040 0.138Employees 0.115 0.187Independent 0.582 0.332Audited 0.745 0.437Rated 0.262 0.440Supervised 0.368 0.483MFI size 13.913 1.910MFI age 2.881 1.859Individual 0.808 0.395NGO 0.659 0.475NBFI 0.087 0.283Inflation 0.181 0.301Economy size 23.182 1.435IR 1.971 0.530BSR 2.206 0.602CP 1.865 0.623

4. DISCUSSION OF THE RESULTS

The objective of the empirical part is to esti-mate (1) using all variables for which data isavailable. Since the data come from severalsources, not all variables are available for allyears. Moreover, in some specifications correla-tion among explanatory variables caused multi-collinearity, and thus some variables had to bedropped (e.g., MFI type in Models 1 and 2 ofTable 3). In addition, the number of includedvariables needed to be controlled to preservedegrees of freedom.A second objective of the analysis was to use

as many observations as possible. For that pur-pose, (1) was re-estimated with expanded oralternative samples in order to revalidate the re-sults from regressions with more explanatoryvariables but with smaller samples. Summarystatistics of the variables used in the analysisare presented in Table 2. Tables 3 and 4 showresults from estimations of specifications withthe best fit, representative of consistent resultswithin that sample. These specifications alsohad to satisfy the Breusch–Pagan test for ran-dom effects.Table 3 shows results from estimation of the

impact on sustainability (on ROA and onOSS), while Table 4 shows the impact on out-reach (on NAB and on DEPTH). Results indi-cate that MFIs are different from private firms.Specifically, the results confirm Hypothesis 1that the remuneration regime (performance-based compensation or fixed salary) does notmatter since the coefficient on Fixed_wage isnot significant in any of the specifications.This suggests that MFI managers may nothave reacted to performance-based compensa-tion offered during the study period and are,therefore, dissimilar to managers of privatefirms.What seems to matter is that managers are

adequately compensated. Hypothesis 2, whichstates that managers may care about the mis-sion and may achieve good results even if of-fered lower salary, is not supported by theevidence. On the contrary, results indicate thatMFIs whose managers were paid what they

considered inadequate (lower) wages reachedfewer borrowers (two different samples—Mod-els 1 and 2 in Table 4, replicated the result),although these MFIs did not differ in terms oftheir sustainability.Managerial skills (quality) matter as indi-

cated by the positive and significant sign ofExperience in the OSS and NAB regressions,as well as the desirable negative and significantcoefficient in the DEPTH regression indicatingthat MFIs with more experienced managersreached poorer borrowers and produced bettersustainability.The strongest result of this study is the sup-

port for Hypothesis 3 that MFIs with a higherproportion of unaffiliated directors had bettersustainability (ROA) and reached poorer bor-rowers. A complementary finding is that a high-er proportion of employees on the board isassociated with worse performance (in termsof sustainability ROA, as well as breadth ofoutreach NAB). These results are consistentwith the empirical and theoretical literatureon for-profit boards, and indicate that MFIs

Table 3. Random effects estimates of sustainability

(1) (2) (3) (4)ROA ROA OSS OSS

Constant �14.158 �39.659 �1.503 0.675(0.17) (1.41) (0.73) (0.49)

Management

Fixed_wage 3.195 0.079(0.47) (0.35)

Lower_wage �5.113 �0.195(1.01) (1.06)

Experience 0.376 0.024*

(1.16) (1.65)

Boards

Employees �53.644** 0.694(2.24) (0.70)

Independent 60.660*** 0.341(3.44) (3.78) (0.53)

Board_size �0.936 4.314** �0.075** 0.024(0.46) (2.67) (2.01) (0.65)

Women 2.111 7.432 0.387 �0.093(0.17) (0.52) (1.21) (0.29)

Donor 32.577 29.502 0.605 �0.845*

(1.41) (1.30) (0.38) (1.82)Financiers 4.806 6.909 0.871* 0.572

(0.26) (0.41) (1.70) (1.51)Local_businessmen 6.499 24.923 �0.989* 0.298

(0.25) (1.21) (1.73) (0.87)Client 10.241 1.100*

(0.40) (1.93)

External governance

Audit (lagged) 1.884 9.345** 0.016(0.35) (2.05) (0.09)

Rating (lagged) �10.591 7.039 0.056(1.15) (0.84) (0.37)

Supervised �18.669*** 6.054 0.044(3.61) (0.70) (0.18)

MFI characteristics

Individual �25.843* �5.903 0.318 �0.244(1.67) (0.55) (1.10) (1.22)

MFI size �0.924 0.168(1.05) (0.16)

MFI age �0.972 �0.942 0.149**

(0.53) (0.55) (2.01)MFI age squared �0.015*

(1.76)MGO �25.046 �0.292

(1.40) (0.72)NBFI �22.606 �0.393

(1.31) (0.95)

Markets and institutions

Inflation 67.494 �2.154* �0.604(1.62) (1.84) (1.07)


Table 4. Random effect estimates of outreach

(1) (2) (3) (4) (5) (6)NAB NAB NAB DEPTH DEPTH DEPTH

Constant 5.142 3.820 3.258 0.917 3.526 16.599***

(0.65) (0.63) (0.83) (0.22) (0.85) (5.14)

Management

Fixed_wage 0.256 �0.035 �0.087 �0.006(0.40) (0.06) (0.22) (0.02)

Lower_wage �1.037* �1.426*** 0.294 0.504(1.90) (3.03) (0.79) (1.58)

Experience 0.035 0.058** �0.016 �0.035*

(0.98) (2.23) (0.80) (1.93)

Boards

Employees �3.509** �2.191* �0.788(2.02) (1.66) (0.72)

Independent �1.369 0.770 �1.323**

(1.04) (0.98) (2.30)Board size �0.005 0.041 0.180** �0.041 �0.045

(0.04) (0.38) (2.05) (0.54) (0.55)Women 1.113 �3.422***

(1.26) (4.13)Donor �1.127 �2.184**

(1.29) (2.33)Financiers 3.059*** 1.368

(2.99) (1.53)Local_businessmen 1.657** 0.956

(2.03) (1.18)Client 0.230 4.795***

(0.17) (4.35)(continued next page)

Table 3—continued

(1) (2) (3) (4)ROA ROA OSS OSS

Economy size �1.494 0.116 0.036(0.41) (0.92) (0.56)

IR 17.108** 13.520** 0.392** 0.387***

(2.49) (2.24) (2.05) (3.01)CP 9.203 0.515 �0.110 �0.186

(1.49) (0.09) (0.71) (1.40)BSR 0.217 �0.583 �0.393*

(0.04) (1.20) (1.92)

R2 0.85 0.33 0.75 0.44v2 143 25 118 40Breush–Pagan 5.26 3.94 7.39 6.9P-value 0.02 0.04 0.007 0.008Observations 46 78 59 110

Absolute value of t-statistics in parentheses.Dependent variables measuring sustainability are ROA (return on equity), and OSS (operational self-sustainability),defined as operating revenue/(financial expense + loan loss provision + operating expense), which measures howwell the MFI can cover its costs through operating revenues.* Significant at the 10% level; ** Significant at the 5% level; *** Significant at the 1% level.


Table 4—continued

(1) (2) (3) (4) (5) (6)NAB NAB NAB DEPTH DEPTH DEPTH

External governance

Audit (lagged) 0.330 �0.144(1.33) (0.91)

Rating (lagged) 0.944* �0.098(1.65) (0.27)

Supervised �0.808 0.449(1.23) (1.05)

MFI characteristics

Individual 0.885 0.471 0.406 0.485 0.507 �0.018(1.39) (0.83) (0.70) (1.06) (1.15) (0.03)

MFI_size �0.034 0.003 �0.052(1.22) (0.08) (1.15)

MFI_age 0.065 0.001 0.078 0.039 0.059 �0.050(0.97) (0.01) (1.32) (0.87) (1.19) (0.79)

NGO 0.859 �0.617 0.575 �2.433*** �1.383** �2.017***

(0.65) (0.71) (0.74) (4.34) (2.39) (2.97)NBFI 1.381 �0.255 0.221 �2.311*** �1.345** �0.588

(1.01) (0.28) (0.26) (3.63) (2.18) (0.78)

Markets and institutions

Inflation �8.715*** �11.029*** �4.819*** 0.144 1.841 �1.238(4.68) (5.46) (3.88) (0.12) (1.48) (0.95)

Economy_size 0.004 0.066 0.051 0.422** 0.341 �0.406***

(0.01) (0.22) (0.29) (1.97) (1.57) (2.75)IR 0.614* 1.261*** 1.324*** 0.318 �0.184 0.057

(1.83) (3.94) (4.52) (1.53) (0.91) (0.20)CP 0.098 �0.025 �0.289 0.060 0.023 0.861***

(0.27) (0.06) (0.97) (0.30) (0.09) (2.90)BSR �0.262 �0.560 �1.119*** �1.913*** �1.896*** �0.637

(0.41) (0.90) (2.66) (3.76) (3.42) (1.39)

R2 0.82 0.78 0.62 0.750 0.76 0.60v2 89 122 90 56 67 49Breush–Pagan 4.61 30.09 27.64 9.33 13.49 6.38P-value 0.03 0.0001 0.0001 0.002 0.0002 0.01Observations 68 100 144 57 71 94

Absolute value of z-statistics in parentheses.The dependent variable measuring the impact on breadth of outreach is NAB (measured as the log of the number ofactive borrowers) and the dependent variable measuring the impact on depth of outreach is DEPTH (measured as theaverage size of outstanding loans size/GDP per capita in $US; lower values mean that the MFI serves poorerborrowers).* Significant at the 10% level; ** Significant at the 5% level; *** Significant at the 1% level.


can also benefit from more independent boards.Thus, while the opportunity for additionalmonetary rewards for managers did not seemto improve performance, and lack of adequatemonetary reward hurt the mission, noncom-pensated unaffiliated directors seem to deliverboth higher sustainability and outreach.The evidence for Hypothesis 4, namely, that

smaller boards are more efficient, is contradic-

tory. Model 2 in Table 3 uses the largest possi-ble sample to estimate the impact of board sizeon ROA (while controlling for other boardcharacteristics and external factors) and in thisspecification the impact of board size is positiveand significant. Similarly, in Model 3 of Table 4(which uses the largest number of observationsfor NAB), MFIs with larger boards seem to dobetter. However, according to the more com-


plete Model 3 in Table 3, MFIs with smallerboards achieve better sustainability. Sinceboard size (board efficiency) may be correlatedwith managerial remuneration regimes, the re-sults obtained from alternative samples didnot produce consistent evidence on the impactof board size.The impacts of board diversity and skills are

not always very strong (Hypothesis 5). A jointtest of the hypothesis that these variables arejointly zero is rejected in all specifications andwith data from all observations. Regardingthe specific effect of each group of representa-tives, board diversity (higher proportion ofwomen) did not influence sustainability. Thisis consistent with studies on the impact ofboard diversity on the financial performanceof Community Development Financial Institu-tions in the United States, which find that inorganizations like these where women are ade-quately represented on the board, there is nomarginal benefit of additional board diversity(Hartarska, 2004). However, diversity may im-pact outreach, because MFIs with higher pro-portions of women on their board reachedpoorer borrowers (Model 6, Table 4).Donor representatives may have also focused

on depth of outreach as indicated by the nega-tive (desirable) and significant coefficient onDonors (Model 6, Table 4), at the expense offocus on sustainability (negative coefficient onOSS). In addition, donor representatives’ abil-ity to raise funds may have brought in easymoney and, thus, lower incentives for sustain-ability.As expected, board members with financial

skills improve sustainability. Surprisingly,MFIs with more financially sophisticatedboards and with higher proportions of localbusinessmen improve breadth of outreach,although the latter impacts sustainability(OSS) negatively. While financial literacy ofboard members seems to matter, the skills oflocal businessmen have less clear impact.Results concerning the role of clients on the

board are interesting. This category of boardmembers affects sustainability positively, butat the expense of depth, perhaps because clientsmay have engaged in rent-seeking behavior, bypromoting lending to wealthier borrowers.Since the categories of clients and donors inthe 2002 survey fall within the category of‘‘gray’’ directors, the contradictory results onthe impact of these groups seem to underscorethe importance of relying on independent direc-tors not affiliated with any stakeholder.

The results also suggest that external mecha-nisms of control play a limited role in improv-ing MFI performance. The joint hypothesisthat all external governance mechanisms influ-ence performance significantly is rejected forOSS and DEPTH, although not for ROA andNAB. The coefficient on the lagged Auditedvariables is positive and significant in Model 2of Table 3, where the (possible) requirementthat MFIs of a certain type have audited finan-cial statements is controlled for by includingthe variables NGO, NBFI as well as MFI ageandMFI size. This result is not replicated, how-ever, with the smaller sample where manage-ment compensation is controlled for (Model 1in Table 3).The evidence that rating improves perfor-

mance is scarce—rating is significant at the10% level only in the outreach regression(Model 1 in Table 3) but not in the regressionson sustainability variables. Although these re-sults can be idiosyncratic to the sample periodwhen rating in the region was not widely usedand usually was donor mandated, they are con-sistent with the recent developments in microfi-nance rating, where raters struggle to surviveperhaps because they failed to become effectiveexternal governance mechanism.MFIs supervised by Central Banks or bank

supervision authority do not reach more bor-rowers and, according to results from Model1, Table 3, have lower ROA, although the resultis not replicated with the larger sample. Never-theless, this result is consistent with previousempirical studies which found that govern-ments should not rely exclusively on direct reg-ulation of banking activities (both prudentialregulation and entry restrictions) to promotebank efficiency (Barth, Caprio, & Levine,2004).MFI specific characteristics such as size and

age are strongly correlated with governancecharacteristics and no conclusions regardingthe impact of learning and economies of scaleon performance can be made since, in somecases, these variables had to be dropped dueto multicollinearity. Results from regressionof OSS using the largest sample of observa-tions, where external governance and manage-rial compensation schemes were notcontrolled for (Model 4 in Table 3) reveals anonlinear impact of age, whereby OSS im-proves with age up to the fifth year anddecreases with age afterwards. Size was not sig-nificant in any of the regressions where it couldbe included without causing mulicollinearity.


Except for Model 1 in Table 3, where MFIusing individual lending had lower ROA (sig-nificant at the 10% level), results indicate thatlending technology did not influence perfor-mance, which is surprising because individuallending is both recommended and dominantin the region (Almendariaz de Aghion & Mor-duch, 2000).The results demonstrate the importance of

controlling for crosscountry differences. Thelevel of inflation, for example, negatively affectssustainability but MFIs in higher inflationaryenvironments seem to reach more borrowers.The size of the economy does not affect sustain-ability although its impact on depth of outreachis not consistent in alternative samples.Improvement in infrastructure is associated

positively with both sustainability (OSS) andoutreach (NAB) perhaps because it decreasesthe costs to MFIs and their clients. Improve-ment in competition policy only worsens thedepth of outreach, perhaps because it makeswealthier borrowers more attractive.Banking sector reform has an interesting ef-

fect on sustainability. Countries with a more re-formed banking sector seem to have MFIs thatregister lower OSS. However, MFIs in coun-tries that have achieved progress in banking re-form reach poorer borrowers, perhaps becausecompetition from other banks forces MFIs toserve the niche of poorer clients. These resultsare also consistent with the literature thatclaims that more developed banking regulations(on which this EBRD index is based) are notassociated with better-performing banks, butimprovements in market-based mechanisms ofbank control improve the performance of finan-cial institutions (Barth et al., 2004).

5. CONCLUSIONS

This paper studies how governance mecha-nisms affect performance of MFIs in Centraland Eastern Europe and the Newly Indepen-dent States. Using insights from the corporategovernance literature, and the literature onnonprofit and bank governance, the paperexamines the impact of management remunera-tion, board independence and diversity, andexternal mechanisms of control, holding institu-tional, macroeconomic, and MFI-specific fac-tors constant. Results indicate that not allknown governance mechanisms affect perfor-mance and, moreover, different factors have dif-ferential effects on outreach and sustainability.

This study finds that some mechanisms thatserve to align the interests of managers withthose of other stakeholders may have a limitedrole in microfinance. Since MFI managers mustconsider the welfare of clients as well as thefinancial success of the institution, managersperform several tasks. In multitask environ-ments, explicit and implicit incentive schemessuch as compensation, perks, etc., become lesspowerful, and less able to motivate managers(Dewatripont, Jewitt, & Tirole, 1999). Multipletasks in organizations lead to a costly lack offocus and standard mechanisms of control suchas performance-based compensation becomeless effective, while the board of directors be-comes an even more important governancemechanism.The most important implication of this study

is that the microfinance board is very impor-tant. Consistent with results from other studieson board independence, microfinance boardswith larger proportions of unaffiliated directorsachieve better results. Thus, independence ofthe microfinance board should be promoted.This is important because affiliated board

members have a differential impact on perfor-mance due to the different emphasis on out-reach and sustainability by various groups ofstakeholders. For example, results show thatdonor representatives improve depth of out-reach but worsen sustainability. Results alsoshow that MFIs with clients represented onthe board have better sustainability at the ex-pense of depth of outreach, suggesting rent-seeking behavior by these stakeholders. Thus,those who advocate including clients on theboard should be aware of the negative impactthat this group may have on depth of outreach.Audit, rating, and supervision by central

bank authorities play only a limited role. Theevidence suggests that the attempt to developan independent microfinance rating industrythat could help improve the channeling ofscarce donor funds may not have been success-ful. Recent observations by microfinance pro-fessionals that the rating industry is strugglingare consistent with the empirical findings in thispaper. In addition, new information disclosureefforts, such as the creation of the MIXMAR-KET information exchange platform, may havesuccessfully satisfied the needs of donors andinvestors and may have sidelined rating as tooexpensive an alternative.The results on the impact of external gover-

nance mechanisms and managerial compensa-tion should be interpreted with caution,


however, because they may be idiosyncratic tothe region and perhaps the sample. Studies inother regions show that transformation intocommercial banks is desirable. In Eastern Eur-ope, by 2002, only three MFIs transformed intobanks and only one of them is in the sample. Inaddition, of the ten microfinance banks createdby IPC, only one is included in the sample andthus the impact of prudential regulations maynot be well captured.This paper presents the first evidence on the

link between governance mechanisms and per-

formance in microfinance. Clearly, some tradi-tional governance mechanisms work formicrofinance organizations. More broadly,mechanisms such as independent boards maybe appropriate for other organizations in theregion, both for-profit and nongovernmental.As for the microfinance industry, more researchand, more importantly, better data will need tobe collected and analyzed to learn more aboutthe impact of governance in order to ensurethat strong organizations channel scarce re-sources to the entrepreneurial poor.

NOTES

1. The industry practice has been to publish perfor-mance information aggregated by region, such as thedata published by the Microbanking Bulletin. Individualdata are also published but only in terms of the numberof clients, for example, the data collected and publishedby the Microcredit Summit.

2. Local regulation, rather than donor’s objectives,may dictate the choice of legal form. For example, somecountries may not permit lending by nonprofit firms,while in others only licensed banks can provide loans butlending-only activity may be permitted in the nonprofitsector (e.g., foundations, etc.).

3. This definition is based on the definition by Shleiferand Vishny (1997) where corporate governance isdefined as the mechanism through which shareholders(providers of funds) ensure themselves that they willreceive maximum return on their investments.

4. In a widely cited study, Jensen and Murphy (1990)find that, for large corporations, pay–performancesensitivity is only $3.25 for every $1,000 increase inshareholder value. Recent papers show that this sensi-tivity has been increasing (Murphy, 1999).

5. In this model, in the first period, donors fund bothNGOs and for-profit firms (as in the case of microfi-nance) and only the efficient organizations in the second.As the industry matures, donors are increasingly con-cerned with efficiency and are willing to fund only theefficient MFIs, so the prediction that the lending/savingtechnology, and not staff wages, will determine survivalis an important insight. A caveat of this model suggeststhat wages could even be lower if personnel are verycommitted to the MFI mission.

6. Insiders are board members who receive pay, butbecause pay is atypical in nonprofit boards, this measureis not very useful (Callen & Falk, 1993; Oster &O’Regan, 2002).

7. In the corporate governance a third group (graydirectors) is also identified but these directors are usuallyadded to the group excluded from the analysis. Forexample, gray directors are considered insiders if themodel studies the impact of independent directors(Block, 1999).

8. The term gray board member is used to describeboard members who are not employees of the MFI butare involved in some of the MFI activity.

9. In the law literature, Manne (1999) proposes asimilar solution for NGO governance, namely, that anexternal, for-profit company (which is disciplined bymarket forces) serves as a monitoring mechanism.NGOs will contract with it to be monitored in termsof the charitable and financial aspects of their opera-tions. However, according to Manne, these privateorganizations should not just be raters; rather, theyshould have the right to sue NGOs to rectify violations.

10. For most categories, board categories do notoverlap, except for the case of clients and donorrepresentation. The survey on 2002 focused on classify-ing members as independent, employees, and graydirectors (connected to clients or donors), while the2001 survey focused on skills and diversity.

11. The most distinctive feature is the rule ‘‘oneperson–one vote,’’ which changes the decision makingprocess.

12. Although the properties of the Hausman test forendogeneity are not well understood in small samples,this test does not indicate that individual governancemechanism are endogenous (Wooldridge, 2002).

13. An alternative, market-based indicator of perfor-mance is Tobin’s q, measured as the current marketvalue of the company divided by the replacement cost of


the company’s assets, which is usually measured as thebook value of the company’s assets.

14. Navajas et al. (2000) provide a detailed descriptionof several outreach dimensions.

15. GDP per capita data were obtained from theEuropean Bank for Reconstruction and Development(EBRD).

16. Bijt are not lagged because during the study period,boards met four times a year or more. Moreover, duringthe study period, most board members’ composition (by

categories) remained virtually unchanged. The impactof Mij is also in the current year because current yearremuneration schemes are hypothesized to affect currentyear performance.

17. Unfortunately no information on whether theMFIs were subjected to prudential regulations or othertypes of regulations is available.

18. Rating is reported in the current year but in allcases was based on previous years’ financial statements.

19. Larger values of the indexes are preferred.

REFERENCES

Adams, R., & Mehran, H. (2003). Is bank holdingcompany governance different? Economic PolicyReview, 9(1), 123–142.

Almendariaz de Aghion, B., & Morduch, J. (2000).Microfinance beyond group lending. Economics ofTransition, 8(2), 401–420.

Ashbaugh, H., & Warfield, T. (2003). Audit as acorporate governance mechanism: evidence fromthe German market. Journal of InternationalAccounting Research, 2, 1–22.

Barth, J., Caprio, G., & Levine, R. (2004). Bankregulation and supervision: what works best? Journalof Financial Intermediation, 13(2), 205–248.

Barth, J., Nolle, D., Phumiwasana, T., & Yago, G.(2003). A cross-country analysis of the bank super-visory framework and bank performance. FinancialMarkets, Institutions & Instruments, 12(2), 67–120.

Bhagat, S., & Jefferie, R. (2002). The econometrics ofcorporate governance studies. Cambridge, MA: TheMIT Press.

Block, S. (1999). The role of nonaffiliated, outsidedirectors in monitoring the firm and the effect of onshareholder wealth. Journal of Financial and Strate-gic Decisions, 12, 1–8.

Callen, J., & Falk,H. (1993). Agency and efficiency in non-profit organizations. Accounting Review, 68, 48–65.

Campion, A. (1998). Current governance practices ofmicrofinance institutions. Occasional Paper No. 3,The Microfinance Network, Washington, DC.

Campion, A., & White, W. (1999). Institutional meta-morphosis: transformation of microfinance NGOsin regulated financial institutions. OccasionalPaper No. 3, The Microfinance Network, Washing-ton, DC.

Carter, D. A., Simkins, B. J., & Simpson, W. G. (2003).Corporate governance, board diversity, and firmvalue. The Financial Review, 38(1), 33–63.

Conning, J. (1999). Outreach, sustainability, and lever-age in monitored and peer-monitored lending. Jour-nal of Development Economics, 60, 51–77.

Dewatripont, M., Jewitt, I., & Tirole, J. (1999). Theeconomics of career concerns, Part I: comparinginformation structures. Review of Economic Studies,66(1), 183–198.

Dewatripont, M., & Tirole, J. (1994). A theory of debtand equity, diversity of securities and manager-shareholder congruence. Quarterly Journal of Eco-nomics, 109, 1027–1054.

Dichter, T. (1997). Appeasing the Gods of sustainability:the future of international NGOs in microfinance. InD. Hulme, & E. Michael (Eds.), NGOs, states anddonors: Too close for comfort? International politicaleconomy series (pp. 128–139). New York: St. Mar-tin’s Press in association with Save the Children.

Easley, D., & O’Hara, M. (1988). Optimal non-profitfirms. In Rose-Ackerman (Ed.), Economics of non-profit institutions. New York: Oxford UniversityPress.

EBRD Transition Report (2003). European Bank forReconstruction and Development, London.

Eisenberg, T., Sundgren, S., & Wells, M. (1998). Largerboard size and decreasing firm value in small firms.Journal of Financial Economics, 48, 35–54.

Fama, E., & Jensen, M. (1983a). Agency problems andresidual claimants. Journal of Law and Economics,26, 327–349.

Fama, E., & Jensen, M. (1983b). Ownership and control.Journal of Law and Economics, 26, 349–371.

Foster, S., Green, S., & Pytkowska, J. (2003). The stateof microfinance in CEE & NIS. Washington, DC:Consultative Group to Assist the Poor.

Handy, F. (1995). Reputation as collateral: an economicanalysis of the role of trustees of nonprofits.Nonprofit and Voluntary Sector Quarterly, 24(4),293–305.

Handy, F., & Katz, E. (1998). The wage differentialbetween non-profit institutions and corporations:getting more by paying less? Journal of ComparativeEconomics, 26, 246–261.

Hardy, D., Holden, P., & Prokopenko, V. (2003).Microfinance institutions and public policy. Journalof Policy Reform, 6, 147–158.

Hartarska, V. (in press). Community developmentfinancial institutions in the southeast: board sizeand diversity as governance mechanisms. SouthernBusiness and Economics Journal.

Hartarska, V. (2004). Community Development Finan-cial Institutions in the Southeast: Board Size and


Diversity as Governance Mechanisms. The South-ern Business and Economics Journal, 27(1&2),102–110.

Healy, P., & Palepu, K. (2001). Information asymmetry,corporate disclosure, and the capital markets: areview of the empirical disclosure literature. Journalof Accounting & Economics, 31, 405–440.

Hermalin, B. E., & Weisbach, M. S. (2003). Board ofdirectors as an endogenously determined institution:a survey of the economic literature. Economic PolicyReview, 9(1), 7–20.

Holmstrom, B. (1999). Managerial incentive problems: adynamic perspective. Review of Economic Studies,66(1), 169–182.

Holtmann, M. (2002). Principles for designing staff incen-tive schemes. Available from www.MicroSave.org.

Houston, J., & James, C. (1995). CEO compensationand bank risk: is compensation in banking structuredto promote risk taking? Journal of Monetary Eco-nomics, 36, 405–431.

James, E. (1983). How non-profits grow: a model.Journal of Policy Analysis and Management, 2,350–365.

Jensen, M. (1993). The modern industrial revolution,exit, and the failure of internal control system.Journal of Finance, 48(3), 831–880.

Jensen, M., & Meckling, W. (1976). Theory of the firm:managerial behavior, agency costs and ownershipstructure. The Journal of Financial Economics, 3,305–360.

Jensen, M., & Murphy, K. (1990). Performance pay andtop-management incentives. Journal of PoliticalEconomy, 98, 225–264.

John, K., Mehran, H., & Qian, Y. (2004). Regulation,subordinated debt and incentive features of CEOcompensation in the banking industry. Availablefrom http://ssrn.com/abstract=497922.

John, K., & Qian, Y. (2003). Incentive features in CEOcompensation in the banking industry. EconomicPolicy Review—Federal Reserve Bank of New York,9(1), 109–121.

John, T. A., & John, K. (1993). Top-managementcompensation and capital structure. Journal ofFinance, 48(3), 949–974.

Keasey, K., Thompson, S., & Write, M. (1997). Corpo-rate governance: Economic and financial issues.Oxford: Oxford University Press.

Macey, J., & O’Hara, M. (2003). The corporate gover-nance of banks. Economic Policy Review, 9(1),91–107.

Manne, G. A. (1999). Agency costs and the oversightof charitable organizations. Wisconsin Law Review,227.

McGuire, P. B. (1999). Policy and regulation forsustainable microfinance: country experiences inAsia. Journal of International Development, 11(5),717–729.

Molyneux, P., Lloyd-Williams, D., & Thornton, J.(1992). European Banking—an analysis of compet-itive condition. In J. Revell (Ed.), The changing faceof Europe banks and securities markets. New York:St. Martin’s Press.

Morduch, J. (2000). The microfinance shism. WorldDevelopment, 26, 783–790.

Murphy, K. J. (1999). Executive compensation. In O.Ashenfelter, & D. Card (Eds.). Handbook of laboreconomics (Vol. 3). Amsterdam: North-Holland.

Navajas, S., Conning, J., & Gonzalez-Vega, C. (2003).Lending technologies, competition and consolidationin the market for microfinance in Bolivia. Journal ofInternational Development, 15(6), 747–770.

Navajas, S., Schreiner, M., Meyer, R., Gonzalez-Vega,C., & Rodriguez-Meza, J. (2000). Microcredit andthe poorest of the poor: theory and evidence fromBolivia. World Development, 28(2), 333–346.

Oster, S., & O’Regan, K. (2002). Does governmentfunding alter non-profit governance. Journal ofPolicy Analysis and Management, 21, 359–379.

Oster, S., & O’Regan, K. (2003). Does the structure andcomposition of the board matter? The case of non-profit organizations. Working paper, Yale School ofManagement.

Rock, R., Otero, M., & Saltzman, S. (1998). Principlesand practices of microfinance governance. Washing-ton, DC: ACCION International.

Scott, C., & Hopkins, R. (1999). The economics ofNGO’s. LSE, DEDPS No. 15, London.

Shleifer, A., & Vishny, R. (1997). A survey of corporategovernance. Journal of Finance, 52(2), 737–783.

Tadesse, S. (2003). Bank fragility and disclosure: inter-national evidence. Mimeo, Moore School of Busi-ness, University of South Carolina.

Van Greuning, H., Galardo, J., & Randhawa, B. (1999).A framework for regulating microfinance institu-tions. The World Bank Policy ResearchWorking Paper No. 2061, The World Bank, Wash-ington, DC.

Westphal, J., & Milton, L. (2000). How experience andnetwork ties affect the demographic minorities oncorporate boards. Administrative Science Quarterly,45, 366–398.

Wooldridge, J. (2002). Econometric analysis of crosssection and panel data. Cambridge, MA: MIT Press.

Yermack, D. (1996). Higher valuation of companieswith a small board of directors. Journal of FinancialEconomics, 40, 185–212.

http://www.MicroSave.org

http://ssrn.com/abstract=497922

governance and performance of microfinance institutions in central and eastern europe and the newly...

Documents