burger 2010

World Development Vol. 38, No. 9, pp. 1263–1277, 2010Crown Copyright � 2010 Published by Elsevier Ltd. All rights reserved

0305-750X/$ - see front matter

www.elsevier.com/locate/worlddevdoi:10.1016/j.worlddev.2009.12.018

Promoting Transparency in the NGO Sector: Examining

the Availability and Reliability of Self-Reported Data

RONELLE BURGERUniversity of Stellenbosch, South Africa

and

TRUDY OWENS *

University of Nottingham, United Kingdom

Summary. — Amid calls for NGOs to become more accountable, this work examines discrepancies between what NGOs say and do.Using a unique dataset of NGOs in Uganda it investigates the inaccuracies in reported financial transparency and community partici-pation. We find that the threat of being caught reduces the likelihood of financial misrepresentation, while a desire to maintain a goodreputation leads to misrepresentation of community consultation. Analysis provides indications that: NGOs with antagonistic relationswith government may be more likely to hide information; and that unrealistic donor demands may be an obstacle to transparency. Find-ings caution against an overly naıve view of NGOs and a reliance on self-reported information.Crown Copyright � 2010 Published by Elsevier Ltd. All rights reserved.

Key words — NGO, Africa, accountability, transparency, participation

* We would like to thank Richard Disney, Oliver Morrissey, and Daniel

Seidmann for helpful comments on early drafts. We would also like to

thank participants at the Centre for the Study of African Economies

Annual Conference, University of Oxford, March 2008; and seminar pa-

rticipants at the Southern Africa Labour and Development Research Unit,

University of Cape Town, April 2008. Final revision accepted: December17, 2009.

1. INTRODUCTION

The NGO sector has recently been described as undergoinga “crisis of accountability and transparency” (McGann &Johnstone, 2006, p. 66), which has damaged its credibility. De-spite rising pressure for more openness and transparency, itappears that most NGOs remain reluctant to share informa-tion, and many refuse to recognize the need for accountability(Elkington, 2003). 1 This is concerning as transparency is gen-erally regarded as a key requirement for successful feedbacksystems and good governance. Similar to corporations andgovernments, NGOs are not infallible, and the timely avail-ability of reliable information is essential for the effective reg-ulation and monitoring of NGOs by beneficiaries, donors, andgovernments.

While the lack of transparency in the sector has often beenmentioned in passing, there have been few attempts to under-take an in-depth empirical study of the phenomenon. Using a2002 representative survey of approximately 300 UgandanNGOs, we examine the availability and reliability of self-re-ported information in this sector.

The data suggest that while most NGOs claim to be trans-parent, it appears that a large number of NGOs are not pre-pared to face the risks and sacrifices associated with beingtruly transparent. According to the survey, 85% of NGOs thatproduced annual reports said that they provided these to thepublic upon request, and 73% of NGOs with balance sheetssaid that they shared their accounts. But many organizationsthat claimed to be transparent either failed to provide infor-mation or provided inaccurate information.

One quarter of the NGOs that said they compiled accountsthat were publicly available did not provide information whenrequested to do so. There were also instances of unreliable orinaccurate information provided. In 39% of cases, whereNGOs reported that they asked the community about theirneeds before initiating a project, community members main-

1263

tained that the NGO in question did not. Similarly, in 38%of cases, where NGOs claimed to have solicited feedback fromcommunities, the community representatives said that they re-ceived no requests for feedback.

Why are these NGOs choosing to withhold information orprovide incorrect information? This paper examines the dis-crepancy between what NGOs say they do and what they actu-ally do, in an attempt to understand why NGOs may beprompted to opt against transparency. It is plausible thatmany NGOs may feel compelled to withhold the truth fromthe public eye in order to keep afloat in a competitive fundingmarket dictated by unrealistic donor expectations and pres-sures. Alternatively, could it be that the confidentiality alibiand the prevalence of misrepresentation are simply symptomsof a sector-wide disillusionment with what they may perceiveas restrictive, but ineffective rules and requirements for theirsector? Are NGOs choosing to not answer questions openlyand truthfully to protect themselves from an antagonistic gov-ernment? Or are these merely efforts to hide behavior that isfraudulent, inappropriate, or inefficient?

The paper proceeds as follows: Section 2 describes the infor-mation asymmetry problem in the NGO sector. Section 3 out-lines the data set and Section 4 presents a number ofdescriptive statistics exploring reported and actual transpar-ency. Sections 5 and 6 provide the basic contours of a frame-work of hypotheses for understanding the observed lack oftransparency. Sections 7 and 8 translate the framework into

http://dx.doi.org/10.1016/j.worlddev.2009.12.018

1264 WORLD DEVELOPMENT

an econometric model, while Section 9 reports the results ofthe econometric analysis. Section 10 concludes the study.

2. BACKGROUND: WHY DOES NGOTRANSPARENCY MATTER?

Stiglitz (1999, p. 23) noted, “To pretend that any institutionis infallible, or that there is perfect confidence in the actionsbeing undertaken, is to fly in the face of reality.” In the samespeech Stiglitz then proceeds to argue that the case for trans-parency is largely based on an acknowledgment and recogni-tion of human fallibility. Society has checks and balances sothat the blunders, oversights, and missteps of individualsand organizations can be identified and rectified. Transpar-ency can be viewed as a precondition for the effective function-ing of such checks and balances. Without timely and accurateinformation, stakeholders are unable to hold decision makersaccountable.

Transparency is a key issue in the NGO sector because thereare private information and hidden actions in the NGO-bene-ficiary–benefactor nexus. NGOs’ relationships with theirstakeholders are fraught with information asymmetries, whichmake it more difficult for donors, government, and beneficia-ries to observe potential problems and to hold NGOs account-able. For instance, when donors are based in North Americaor Europe with no local presence, they are physically removedfrom the site of NGO activities, and the distances involvedmay make it prohibitively expensive for the donors to visitproject sites regularly to observe conditions, outputs, and out-comes. In such cases, donors may be heavily reliant on second-hand information, which may often be reported by the NGOitself. If the donor is not familiar with local circumstanceand culture and has no benchmark for cost comparisons, theinterpretation of accounts and reports may be a further obsta-cle to assessing performance and detecting ineffectiveness orfraud.

Beneficiary oversight has been proposed as a remedy for thedeficiencies in the information obtained by donors. It is arguedthat it may be more difficult for NGOs to keep informationand actions hidden from beneficiaries due to their close phys-ical and cultural proximity and direct involvement in develop-ment projects. However, beneficiaries cannot always observeall the relevant dimensions of a good or a service producedby an NGO (Mansuri & Rao, 2004; Tendler, 2000). For in-stance, children whose parents have never been to schoolmay not be able to assess the quality of teaching their childrenare receiving. Other studies have indicated that beneficiarieswho contribute little or nothing toward the goods or servicesthey are receiving may make inferior monitors because theydo not have an incentive to scrutinize NGO outcomes (Plat-teau & Gaspart, 2005; Veron, Williams, Corbridge, & Srivast-ava, 2006). If beneficiaries contribute little, they may often beeasy to please because they are likely to compare the goods orservices supplied by the NGO to receiving no goods or servicesat all, instead of comparing them to alternative goods and ser-vices that the project budget can buy. 2 Furthermore, recentstudies have shown that community monitoring may be vul-nerable to free-riding problems (Olken, 2005) and that projectmanagers know how to hide information from communitymembers (Olken, 2006). Additionally, case studies show thatthe members of poor communities sometimes lack the self-es-teem to act assertively (Burger, 2005; Kihato & Kabemba,2002).

The arguments outlined provide an analytical foundationfor concerns about opportunities for deception in the NGO

sector. However, some theorists have attempted to circumventthese concerns with theory models that aim to show that thedecision makers in nonprofit organizations have little incentiveto abuse the aforementioned information asymmetries.According to the contract failure theory (Fama & Jensen,1983; Hansmann, 1980; Thompson, 1980), NGOs are expectedto behave more honestly than for-profit firms. It is arguedthat, while there are incentives for for-profit organizations tolower their costs by reducing the quality of their goods or ser-vices, there are no such incentives for NGOs because they aresubject to the nondistribution constraint and are thus notmotivated to maximize profit. According to this perspective,the nondistribution constraint prevents the distribution ofresidual earnings to those who exercise control over the non-profit organization, and therefore organizations with a profitmotive will not choose this organizational form. The theoryis contentious because it is heavily reliant on the nondistribu-tion constraint; however, in practice it is unclear how effectivethis constraint is in averting profit-seeking or self-enrichmentby members of these organizations. As the classification of asurplus as profit is largely a matter of accounting, privategains can easily be concealed as inflated salaries, unrealisticbonuses, or undeserved benefits. 3 The presumption that thenonprofit sector will attract more honest and altruisticemployees is far from watertight. Due to the sector’s notoriouslack of transparency (Callamard, 2006; Edwards & Hulme,1996), its poor track record relating to effective monitoringand regulation (Callamard, 2006; Desai and Yetman,1995), 4 the lure of tax exemption, and the often intangibleand subjective nature of the services and goods provided inthis sector (Hansmann, 1987), there is a case to argue thatthe opposite may be true.

There are at least two alternative arguments worth notingwith regard to the assumption of altruism. The first deals withthe motivations of nonprofit funders. Atkinson (1989, p. 73)notes that while “. . . we cannot . . . be sure what their subjec-tive motives are, much less whether they are wholly selfless,”it is clear that “they are not going to receive a market returnon their investment, which [at least] manifests weak altruism.”However, even if one finds this argument convincing, there islittle basis for extending this assumed altruism to include anonprofit organization’s manager and staff. The funder sub-contracts the responsibility to deliver the services or producethe goods to a nonprofit organization’s employees and, with-out more information about the details of the contract andthe incentives offered, it would be imprudent to make anyassumptions about their personalities or their motivation. Infact, Rose-Ackerman (1996) argues that managerial shirkingmay be more prevalent in nonprofit organizations becausethere is no take-over threat in this sector.

The second argument pertains to attempts in the literatureto argue that nonprofit organizations may successfully attractworkers who are systematically more altruistic because theyoften undercompensate their workers—also known as thedonative labor theory. The view is that only individuals whoderive utility from helping others will accept jobs in a sectorthat pays below-market rates (e.g., Rowat & Seabright,2006). Such a model, however, requires strong assumptionsabout the labor market, which has little empirical basis. 5

Weisbrod’s (1998) overview of the empirical research onmotivations concludes that there is little evidence to justifythe bold assumption that NGOs are motivated by altruismand nonmonetary preferences. If the assumption of goodintentions and pro-social motivation is discarded, the allureof abusing an information advantage may be just as temptingand feasible for nonprofit organizations as for for-profit

PROMOTING TRANSPARENCY IN THE NGO SECTOR 1265

organizations. Due to the particular difficulties relating toobserving and verifying NGO behavior described above, it isanalytically conceivable that the problem could be worse fornonprofit organizations.

Additionally, the nonprofit sector has traditionally beensubjected to less stringent government regulation than eitherthe for-profit or the public realm—this despite the deficit insectoral disciplining mechanisms, due to the comparativelyweak, informal, and arbitrary nature of the sanctioning powerof consumers and beneficiaries. In the for-profit sector, con-sumer choice helps to provide feedback. In cases where thequality of services and products is observable, companies lesseffective or efficient than their competitors are expected tostruggle to retain customers and may eventually have to close.In the public realm, a politician’s poor judgment and ineffec-tive management can be punished by voting for one of hisopponents. Anticipating such behavior can help keep politi-cians in check. In the nonprofit sector, there are no analogousmechanisms to motivate NGOs to care about the plight oftheir consumers and beneficiaries. NGO beneficiaries can nei-ther vote against exploitative or corrupt NGOs nor punishunreliable or misbehaving organizations by ceasing to do busi-ness with them.

These information problems are frequently exacerbated indeveloping countries for a number of reasons. The increasedphysical distance between the funder (usually from a devel-oped country) and the beneficiary (usually from a developingcountry) raises the cost of sending and gathering information.Due to the high rate of unemployment and underemploy-ment 6 and frequently also, a dearth of other lucrative entre-preneurial activities with low set-up costs, it may beimprudent to assume that employees and entrepreneurs in thissector are attracted by altruistic motivations—as some devel-oped country theory models would predict. Additionally, theinstitutions in developing countries are often not as matureor robust, which weakens the impact of any legal deterrence.

These arguments suggest that due to the information asym-metry and the moral hazard dilemma in the relationships be-tween NGOs, benefactors, and beneficiaries there are ampleopportunities for deceit and ineffectiveness among NGOs.Furthermore, the assumption that NGO staff is altruistic isviewed as imprudent due to the lack of a robust empiricalfoundation; and the analytical arguments supporting thisassumption are weak and not applicable in developing coun-tries. If we cannot assume that NGO staff members are altru-istic, this leaves the opportunity for selfish motives that couldencourage the exploitation of weaknesses in the informationflows and the feedback system. It is vital to study the exploita-tion of such weaknesses in the NGO sector: NGOs play a piv-otal role in poverty alleviation 7 and these weaknesses canimpede the contribution of the sector significantly.

3. DATA: UGANDAN NGO SURVEY, 2002

The study uses a 2002 representative survey of the UgandanNGO sector, which incorporates two modules: (i) an NGOquestionnaire to collect information on the organization’sstructure, finances, and activities; and (ii) an interview with acommunity focus group to explore how the organizationsare perceived by community members. By capturing boththe community’s perceptions and the organization’s character-istics, the survey enables researchers to postulate links betweencommunity perceptions, such as what value has been added bythe organization, and self-reported organizational featuressuch as size and numbers of skilled workers.

The sampling frame for the first survey module (NGO ques-tionnaire) was constructed via a mini-census of the NGO pop-ulation of 14 districts. The survey sample (298 observations)was drawn from this sampling frame. With 255 questions,the questionnaire is extensive and includes information onfunding, ownership, expenditure, assets, and governance.The questionnaire was administered by Ugandan enumera-tors, usually to the manager at the office of the NGO. Themost contentious data collected were that on their accounts.To ease the burden, the account information asked for washighly aggregated and should have been readily available froman annual report. If the respondent was unable to answer thequestions either a colleague who could answer was called uponor the relevant sheets of the questionnaire were left with theunderstanding that the enumerator would return to collecton a designated day that suited the NGO. The second surveymodule comprised 268 community focus group interviews.These interviews aimed at extracting information on how ben-eficiaries perceived the NGOs surveyed in the first module.

It was crucial to ensure that the focus group evaluating theNGO was aware of their work. To do this, parishes to visitwere selected from a list of six communities which the NGOreported to be working in. The communities were quitesmall—the 2002 Census reports that the average parish con-sisted of 4,625 individuals—so it was plausible that the focusgroup participants would know about an NGO working intheir community. A series of questions were developed to filterout focus groups that did not know the NGO in question.First, the focus group leader would ask them about their prob-lems, then to describe the NGOs who helped them with theseproblems. If they did not list the NGO in question during thisdiscussion, the focus group leader would prompt them. 8 Ineach community visited, 6–10 focus group participants wererecruited via a community leader. This strategy was optedfor ensuring comparability and consistency of the communityperception variables across NGOs—a relatively simple ap-proach that was easily replicable.

The first module of the survey (NGO questionnaire) can bematched to 205 of the 268 observations from the second mod-ule (community focus groups). There were also cases wheresome NGOs were linked to more than one community. Toavoid problems with error terms, 19 duplicates were elimi-nated randomly, reducing the sample to 186 observations. 9

Barr, Fafchamps, and Owens (2005), Barr and Fafchamps(2006) and Barr, Fafchamps, and Owens (2003) provide moreinformation regarding the survey questionnaires and focusgroup interviews.

4. HOW TRANSPARENT ARE UGANDAN NGOS?

There does not appear to be much momentum toward great-er transparency in the sector (Christensen, 2004; Civicus CivilSociety Index Team, 2006). Zadek and Gatward (1996) con-tend that the reluctance to be transparent is clear from the lackof resources devoted to such activities. According to Stiglitz(1999), this reluctance is not unexpected, as there are manyincentives to conceal information. If organizations can choosewhen to disclose and when to conceal information, they canmanage information to ensure that they are praised for theiraccomplishments without being criticized for their failures,mistakes, and shortcomings. Secrecy can also help to hide cor-ruption and ineffectiveness and enable employees to avoid theburden of public scrutiny. Significantly, Stiglitz points out thatit is not difficult for an organization to mimic compliance withtransparency requirements by using codes, creeds, and


commitments, without the organization undergoing any realtransformation toward becoming more open and transparent.Mordaunt (2006) makes a similar point, pertaining specificallyto NGOs, claiming that there is often a considerable gap be-tween “espoused theories” and “theory in use.”

According to the 2002 survey of the Ugandan NGO sector,reported levels of transparency and accountability were rea-sonably high. Eighty-five percent of NGOs that produced an-nual reports asserted that these reports were available to thepublic upon request. Given that annual reports are often usedas marketing tools, it may be more interesting to consider thepercentage of organizations that were willing to share their ac-counts with the public. Seventy-three percent of the NGOsthat kept accounts said that they would share these accountswith a member of the public, should any member of the publicrequest it.

However, as the evidence in the next two sections highlights,these claims regarding transparency may be exaggerated. Aconsiderable proportion of NGOs report high levels of trans-parency, but were unwilling to provide information whenasked, or chose to dispense seemingly inaccurate information.This suggests that while most NGOs are aware that they areexpected to be transparent, many may merely pay lip serviceto the idea.

(a) Inaccurate information

Analysis of the survey shows that organizations claiming tobe transparent may frequently provide inaccurate information,presumably, in most cases, to project a more favorable view oftheir organization. For instance, it is clear that NGO descrip-tions of their efforts to involve the community in projects arefrequently not aligned with the accounts of the communities.

In 91% of cases, NGOs report that they asked the commu-nities about their needs, and 96% of NGOs reported that theyasked the communities for feedback during or after a project.This contrasted strongly with the much lower prevalence ofcommunity involvement cited by the community members’ fo-cus groups. Just 54% of the communities said that the partic-ular NGO concerned asked community members about theirneeds, while only 59% reported that they were requested togive feedback.

In 39% of cases (i.e., for 73 of the 186 matched NGOs in thesample with data available for such a comparison), 10 NGOsreported that they asked the community about their needs be-fore initiating projects, while the community maintained thatthe NGO did not. Similarly, we found that 38% of NGOs(71 of 186 NGOs) claimed to have solicited feedback fromcommunities while the community representatives said thatthey received no requests for feedback. There is a remarkableextent of overlap between the cases of misrepresentation withregards to community needs assessments and soliciting com-munity feedback: 69% of the NGOs that claimed that theyasked communities about their needs before initiating develop-ment projects and appeared not to have done so, also reportedthat they asked the community for feedback, but ostensiblydid not. According to Pearson’s chi-squared test, the hypoth-esis that the two indicators are independent cannot be ac-cepted (Pr < 0.001).

Given the high degree of overlap between misrepresentationin the two dimensions of community involvement and theplausible motive for misrepresentation, the discrepancies be-tween what the NGOs claimed and what the communities re-ported are interpreted as evidence that a large share of theseNGOs misrepresented their actions. 11 The community focusgroup’s answers with regards to community participation are

treated as reliable because we could find little analytical reasonto suspect an incentive for misrepresentation among commu-nity members, and because misrepresentation among membersof a group has a higher cost since it requires coordination andcooperation. Conversely, NGOs may have an incentive to paylip service to community involvement. They may be ashamedof their failure to consult the community, seeing that commu-nity participation is widely viewed as being “good, by defini-tion” (Botchway, 2001, p. 135) and is often depicted as the“magical missing ingredient” (Botchway, 2001, p. 149) indevelopment projects. This interpretation is further strength-ened by the observed asymmetry in the discrepancies—we findonly a negligible number of cases where the community re-ported participation and the NGO did not. 12

The interpretation of these discrepancies as misrepresenta-tion is supported by other cases of apparent misrepresentationin the survey. In their analysis of the same data set, Barr et al.(2005) conclude that the actual prevalence of external auditingwas considerably lower than the high levels reported in thesurvey. Many of the allegedly externally audited accounts con-tained numbers that were either implausible or did not tallywith other figures given. 104 NGOs in this matched sampleindicated that their reports were audited. However, for 50NGOs (or 48% of the 104 organizations that claim to havebeen audited), the accounts contained several careless mistakesor obvious inconsistencies, which one would expect to havebeen identified and resolved during an audit.

Additionally, during the community focus group sessions itwas observed that in 11% of cases communities did not knowor recognize NGOs that claimed to work in their parish. Whileit is conceivable that there may be cases where an NGO can belegitimate and effective without being widely known in the par-ish, it is likely that a high proportion of these “unknown”NGOs may be ineffective or dishonest—especially given thatparishes are quite small and that each of the parishes wereidentified by the NGO itself as a community where theyworked. It must be borne in mind that this 11% is on top ofa large percentage—varying between 70% and 85%, dependingon the district—of registered NGOs that were eliminated fromthe sample group because they could not be located. Based onthe low share of registered NGOs that could be located as wellas anecdotal evidence, Barr et al. (2003) assert that a consider-able proportion of Ugandan NGOs appear to be briefcaseNGOs that have a shadowy existence when they are notreceiving grant income.

(b) Unwillingness to provide information

The survey also shows that there may be a large discrepancybetween organizations’ reported and actual willingness to pro-vide information. Although 75% of the organizations reportedthat their accounts and annual reports were available to thepublic, the survey enumerators found that NGOs were oftenreluctant or unwilling to provide them when asked. For in-stance, 25% of organizations claiming to have accounts andreporting that these were available to the public upon request,failed to provide enumerators with basic expenditure and rev-enue information for either 2000 or 2001. Similarly, 9% of theNGOs that claimed to have compiled balance sheets andmaintained that these were available to the public upon re-quest, failed to provide enumerators with the estimated valueof the NGO’s stocks of cars, equipment, or inventory.

In cases where information is not provided, it may often bedifficult to distinguish between an unwillingness and an inabil-ity to provide information. Also, it may often be imprudent totake explanations in this regard at face value, given that an


organization may plausibly be motivated to save face byclaiming that it is unwilling to provide information when, infact, they are unable to do so. Similarly, it is conceivable thatan organization may try to avoid further queries, visits, andcalls by claiming to not have the necessary skills to generatethe information requested, when they are, in fact, simplyunwilling to provide the information.

5. HYPOTHESES ON THE ECONOMICS OFMISREPRESENTATION

To explore why NGOs would provide not comply withtransparency and accountability norms and provide inaccu-rate information about their compliance, we propose a frame-work reflecting decision-making at two consecutive decisionnodes. Due to the lack of theoretical and empirical researchon this topic, the framework draws on the well-established lit-erature on tax evasion to construct a theoretical framework toguide the empirical analysis. The issues at stake are broadlysimilar given that tax evasion decisions also have both privateand public dimensions. Furthermore, there are also a numberof parallels between the decision-making process associatedwith whether or not to evade tax and whether or not to accu-rately reflect NGO activities.

In both instances compliance is expensive—in terms ofmoney, effort, and sometimes also ability 13—but noncompli-ance is risky as those deviating from these standards can ex-pect punishment when caught. Punishment can range fromfines to peer disapproval (Clotfelte, 1983; Cowell, 1985). Ifthe likelihood of getting caught is quite low, individuals maychoose to hide their noncompliance to avoid fines and to saveface (Allingham & Sandmo, 1972; Clotfelte, 1983). However,this deceit will have an individual cost, e.g., moral discomfort,guilt, or anxiety (Sandmo, 2004; Spicer & Becker, 1980).

The framework builds on the insights of the tax evasion lit-erature by acknowledging the importance of opportunities formisrepresentation. Slemrod (2007) emphasizes the strongempirical relationship between opportunities for tax evasionand the likelihood of tax evasion. 14 The literature also recog-nizes that misrepresentation cannot be fully understood byexamining the anticipated consequences of an action. It is vitalto consider how social norms may affect behavioral choices,and also to investigate the individual values, attitudes, andperceptions that may be linked to these norms (Sandmo,2004; Tanzi, 1993).

(a) First-stage decision: why not to comply with sector norms?

Consequently, the following framework is proposed: In thefirst stage, the NGO chooses whether to adhere to NGO sectornorms such as transparency and community involvement. 15

Because compliance has a cost, adherence to such norms willbe a function of effort and ability. In turn, effort is a functionof inherent and external motivation, with inherent motivationreferring to the extent to which the NGO endorses and sup-ports the sector norm and external motivation referring toexternal mechanisms to detect and punish noncompliance.The decision to not comply with sector norms in this first stagecan thus be depicted as:

y ¼ f ðe; aÞ; ð1Þ

e ¼ gðm; xÞ: ð2ÞWith y = [0, 1] (with 1 representing noncompliance) and0 < e, a, m, x < 1, feð�Þ 6 0, fa�Þ 6 0, gmð�ÞP 0, gxð�ÞP 0,

where fið�Þ ¼ Df ð�ÞDi

16 and where y represents the binary decision

outcome, e is effort, a ability, m internal motivation, and xexternal motivation.

(b) Second-stage decision: why misrepresent lack of compliance?

In the second stage, NGOs decide whether or not to concealtheir noncompliance in the first stage. 17 It is presumed thatNGOs will only be dishonest if they have something to hide,which in this framework amounts to having chosen againstthe norm in the first stage. Thus, we look only at the groupof organizations that chose against the norm, and ask underwhat circumstances these NGOs are likely to be honest abouttheir choice?

Following the contributions of Mazar et al. (2006) and Mar-tinelli and Parker (2009), the decision to misrepresent behavioris viewed as balancing external reward incentives with inherentmotivations. The framework includes inherent motivation bysuggesting that the likelihood of truth-telling will be a functionof the agent’s preference for honesty, here equated with theindividual agent’s commitment to honesty regardless of thespecific context or strategic concerns. The external rewardincentives are mediated via reputation loss—on the one hand,the reputation loss associated with revealing a deviation fromsector norms and on the other hand, the reputation risk asso-ciated with being caught lying. It is assumed here that lowerrevenue is part of the ripple effect of the loss of reputation.NGOs that are closely monitored or tightly regulated shouldhave less opportunity to misrepresent themselves. Regulationand monitoring are expected to improve outcomes by reduc-ing the lure of diverting funds or staff time, via better observa-tion of inputs and outputs and improved information. Thedecision about whether or not to misrepresent can be depictedas:

d ¼ Uða; rL; rT ; ‘Þ: ð3Þ

With d = [0, 1] (with 1 representing misrepresentation),U að�Þ 6 0, UrLð�ÞP 0, UrT ð�Þ 6 0 U ‘ð�Þ 6 0, 0 < a; ‘ < 1,rL < 0, rT < 0, ‘ < 0, r ¼ rT þ rL and where d is the decisionto misrepresent or not, a is the likelihood that the deceit is de-tected, rL is the drop in reputation associated with beingcaught lying, rT is the drop in reputation associated with theNGO “coming clean” about its digression from sector norms,and ‘ represents the agent’s aversion to lying.

It is assumed that the stakes of both truth-telling and lyingwould be higher for an organization with a greater and moreestablished reputation. The net effect of more reputation can-not be determined ex ante. Reputable organizations may havemore to lose from being exposed as liars than lesser knownorganizations, or organizations of disrepute, but similarly,having more of a reputation at stake will also increase theattractiveness of lying to “save face” (rather than admittingnoncompliance to widely held sector norms).

In line with what has been proposed in game theory modelssuch as those of Horner (2002) and Sobel (1985) and con-cluded in experimental investigations such as those of Camererand Weigelt (1988) and Bohnet and Huck (2004), it is here pre-sumed that an organization’s reputation is based on the re-peated interactions between the organization and its clients.The organization’s reputation can thus be thought of as thecumulative experiences of its clients and consequently re-garded as a function of perceived competence, accumulatedover years of existence:

r ¼ hðc; nÞ: ð4Þ

Figure 1. Two-stage framework for examining the economics of

misrepresentation.


With hcð�Þ > 0, hnð�Þ > 0, hnnð�Þ < 0 and where c represents theperceived competence, and n the number of years of existenceof the NGO.

Figure 1 represents the two-stage framework as a decisiontree, with the decision algorithms in reduced form. To summa-rize, the framework suggests that the four main determinantsof the second-stage misrepresentation decision are: havingsomething to hide (here appearing as a precondition for inclu-sion in the second-stage sample), a motive for hiding the truth(reputation loss), the ability or opportunity to hide the truth(represented by the extent of imperfect information and thelack of oversight), and the individual preference for concealingthe truth.

6. SUPPLEMENTARY HYPOTHESES ABOUTMISREPRESENTATION OF NONCOMPLIANCE

The paper also examines a second set of hypotheses pertain-ing to misrepresentation. In contrast with the hypotheses in-cluded in the main framework, these hypotheses have beendiscussed only tentatively in the literature and lack robustempirical evidence. Due to the speculative nature of thesehypotheses, they were not included in the framework, butare tested separately.

(a) Donor demands and funding pressure

The first of these pertains to donor demands and fundingpressure. Alexander (1998) highlights the complex and con-flicting demands NGOs often face trying to juggle their ownmissions and commitments, the requirements of donors, theneeds of beneficiaries, and the concerns of the wider commu-nity. One of the responses to complexity that Alexander hasobserved in her case studies is deflecting, where the NGO “lim-its the amount of external complexity that enters the organiza-tion” (1998, p. 277). Misrepresentation can be conceived as auseful deflection strategy in situations where NGOs may dis-agree with some of the demands of donors, because meetingthese demands may, for instance, put excessive strain on theorganization or conflict with the NGO’s core mission. In acompetitive funding environment, NGOs that want to resistunreasonable or intrusive donor demands may choose to mis-represent their activities instead of openly defying the donor’srequirements. This theory is in line with the observations of

Ebrahim, 18 who indicates that grass roots organizations in In-dia are skilled at devising strategies to subtly resist donor de-mands, while complying sufficiently to ensure that theycontinue to receive funds.

The same donor pressures may also intensify the inherenttrade-off between using information to enhance transparencyor to build reputation (Michaelowa & Borrmann, 2006).NGOs may feel under pressure to represent their activities ina more positive light and to gloss over their shortcomingsand failures. It is thus not surprising that a survey on NGOs’annual reports found them “overwhelmingly positive in tone,”with most of the text devoted to describing the organizations’achievements (Bolton, 2003). Chambers (1996) argues thatcompetitive pressure for grant funding may tempt NGOs tomisrepresent themselves. This resonates with the findings inthe tax evasion literature: Rice (1992) 19 concludes that corpo-rations under strain due to below-average profits appear to bemore likely to engage in such dishonest activities.

(b) Antagonistic government relationships

The second set of supplementary hypotheses explores theimpact of an antagonistic government. Many political sciencetheories of civil society emphasize the role of NGOs as govern-ment watchdogs and policy advocates. In this envisaged role,conflicts may often arise between NGOs and government;thus, it is often assumed that there will be a hostile relationshipbetween the sector and the government. Under such circum-stances it is plausible that NGOs may choose not to be trans-parent, but also attempt to not defy sector norms and donordemands publicly. As Florini (1999) notes, transparency doesnot necessarily protect the weak, it can expose the weak aswell. She argues that “secrecy can be the refuge of the weakagainst the strong, as in the case of human rights organiza-tions working in repressive countries.”

(c) Corruption and ineffectiveness

The final set of speculative hypotheses investigates whetherorganizations that have more to hide may be more likely tobe evasive and misrepresent themselves. The link between gov-ernment corruption and a lack of transparency is well-estab-lished (Florini, 1999; Rose-Ackerman, 1978, 1999; Stiglitz,1999), and here we examine whether such a relationship alsoholds for the NGO sector. In this paper, we look not only atcorruption but also include any behavior that is not in linewith the best interests of the beneficiaries, ranging from pre-meditated cases of fraud to general organizational ineffective-ness.

7. CONSTRUCTING AN ECONOMETRIC MODEL OFMISREPRESENTATION

One of the main objectives of this paper is to understandwhy NGOs resist transparency and misrepresent themselves.The framework defined in Sections 5 and 6 guides the empir-ical analysis. The paper considers misrepresentation in theform of two actions, namely financial transparency and com-munity participation.

Financial transparency is represented as the willingness andability of an NGO to provide information regarding their as-sets, income, and expenditure. Applying the framework to thefinancial transparency variable, the NGO decides in the firststage, whether or not to adhere to sector norms regardingtransparency. In the second stage, we consider only the


subsample of NGOs that were not transparent, and examinewhich of these opted to admit that they were not transparentand which claimed to be transparent.

In terms of the survey estimates: an NGO is viewed as mis-representing their financial transparency if they did not provideasset value estimates yet claimed they compiled accounts whichwere available to the public upon request, or if they did notprovide expenditure or income data yet told enumerators thatthe data were captured and available to the public. To ensurethat we are fair in the classification of misrepresentation, wecategorized the NGO as misrepresenting the transparency ofits cash flow only if they claimed that they kept records of theircash flow and these records were available to the public, butthen failed to provide an indication of total expenditure and to-tal income for either 2001 or 2000. Similarly, we classified anNGO as misrepresenting its balance sheet transparency if itclaimed that they kept a balance sheet and it was available tothe public, but then failed to provide evaluations for cars,equipment, or inventories. The financial transparency misrep-resentation variable was the sum of balance sheet misrepresen-tation and cash flow misrepresentation, i.e., if the NGO wasfound to misrepresent either its balance sheet transparency orits cash flow transparency, it was classified as misrepresentingthe transparency of its financial records.

For the community participation variable, the first-stagedecision is whether or not an NGO involved the communityin a project as sector norms dictated. In the second stage, welook only at those NGOs that neither asked the communityabout their needs beforehand nor requested feedback after-ward. 20 For this group of NGOs that failed to involve thecommunity, we study how truthfully they answer questionsregarding the NGO’s community participation initiatives.

To construct a measure of community participation we lookat whether the organization asked the community about theirneeds prior to commencing a project and requested feedbackfrom the community after a project. To ensure that we erredon the side of caution in our attempts to identify possible casesof misrepresentation, we classified an observation as misrepre-sentation only if the NGO reported that it both asked thecommunity about its needs and solicited feedback while thecommunity focus group said it did neither (i.e., the variableis constructed as the intersection of these two types of misrep-resentation).

Due to the self-selection of observations into the secondstage of the framework, the analysis begins with a Heckmanselection model. According to the framework, misrepresentinginformation is only a temptation for a subsample of observa-tions, namely those that deviated from accepted guidelines ontransparency and community involvement. Having decidednot to provide financial information and/or not to involvethe community, the NGO faces a decision regarding whetheror not to misrepresent the organization’s actions. The Heck-man selection model tests the hypothesis that the error termfrom the first-stage norm conformation decision affects thelikelihood of misrepresentation, after controlling for the inde-pendent variables in the model of the second-stage deceit deci-sion. Due to the binary structure of the variables discussed, weopt for a Heckman probit selection model (using maximumlikelihood estimation).

In defining the econometric model, the assumption is thatdifferent factors could determine whether NGOs pursue com-munity involvement and financial transparency, but the deter-minants of misrepresentation would be the same across thetwo dimensions. The paper presumes that the correlates ofcompliance will be specific to the norm in question. Intuitively,it makes sense that involving community members in projects

will require a different set of skills and capabilities than pro-viding stakeholders with access to accurate financial records.However, since both financial transparency and communityparticipation are strongly held norms in the sector it is likelythat the same considerations will motivate misrepresentationabout not complying with these norms.

Analytically, the exclusion restrictions are less problematicthan in traditional Heckman applications, because there is lit-tle reason to expect that the same set of factors would beresponsible for determining compliance and misrepresentationof compliance. For instance, while there may be some overlapin the correlates, the likelihood of making the organization’sfinancial data available when asked to, is not expected to bedetermined by the same factors that are associated with anhonest response about financial transparency.

In the translation of this theoretical structure to an empiri-cally estimable form, we need to make a number of additionalassumptions about the functional form of these relationships,and also empirical approximations of these concepts. In termsof the functional form, it is assumed that the impact of thevariables on the likelihood function is linear in its parameters.

(a) First-stage decision: why not comply with sector norms?

According to the framework, the first-stage decision, namelywhether or not to adhere to sector norms, y, is determined bythe following function:

y ¼ f ½gðm; xÞ; a�: ð5ÞAll the three coefficients are expected to be negative. Externalmotivation is represented by independent or third-party mon-itoring. To ensure that the variable that measures monitoringincludes direct observation by a third-party only, it is limitedto cases where the community reports that another NGO ororganization asked them for feedback about the NGO in ques-tion.

To capture internal motivation to be financially transparent,we follow the literature in arguing that it is plausible thatfinancial accounts may not be meaningfully interpreted byall stakeholders. It is presumed that donors are the most likelyusers of financial data and thus NGOs will have more of anincentive to align with this norm if the NGO has a grant.NGO alignment with this norm is thus represented by an indi-cator variable on whether or not the NGO had a grant.

Following a similar line of argument, the incentive to alignwith the norm of community involvement will be strongeramong those NGOs obtaining a larger proportion of theirfunding from the community, and we therefore include theproportion of funds from local donations and fees to capturealignment for this norm.

We also add a perceived altruism variable as a general indi-cator of internal motivation. This variable is constructed usingthe community focus group question regarding the perceivedselfishness of the NGO. 21

In the econometric model, the total number of staff membersand the age of the NGO manager are chosen to represent theability or capacity variable, a. To test the hypothesis that qual-ity of staff may be more important than quantity for high skilljobs such as accounting, we also include the share of profes-sionally qualified workers and whether or not the shortageof skilled staff was cited as a constraint by the NGO.

A measure of reporting fatigue is added to this model,namely the number of accounts the NGO was required to sub-mit per year.

Finally, to investigate the theory that NGOs may have a rea-son not to be financially transparent if they have a difficult or


hostile relationship with the government we include an indica-tor variable capturing whether or not the NGO considered itsrelationship with government to be antagonistic. It is for in-stance plausible that an advocacy organization that has anadversarial relationship with the government may need tokeep their funding information and their planned interven-tions hidden from the government to protect itself. Note thatthis argument is distinct from the supplementary hypothesisregarding the impact of antagonistic relationships with gov-ernment on misrepresentation (see Section “Antagonistic gov-ernment relationships”) that considers how such tensionscould influence the misrepresentation of transparency andaccountability.

(b) Second-stage decision: why misrepresent lack of compliance?

The second-stage decision is expressed as: 22

d ¼ U ½a; hðc; nÞ; ‘�: ð6Þ

The first and the last coefficients are expected to be negative.The expected effect of the remaining two coefficients is incon-clusive due to the counterbalancing effects operating via repu-tation (as described in Sections 5 and 6). The likelihood thatthe lie is detected, a, is represented by a third-party monitoringvariable. Perceived competence, c, is captured by a questionthat asked the community focus group whether they thoughtthe NGO staff were good at what they did. 23 The number ofyears of existence of the NGO is included in the econometricmodel, along with a squared term, as is the convention withage effects.

The preference for honesty, ‘, is proxied 24 by an indicatorvariable measuring whether or not the NGO had a religiousaffiliation. Religious affiliation is added as a proxy because itis here conjectured that ethical pragmatism will be less preva-lent among religious organizations due to the latter’s tradi-tional endorsement of rule-based ethics. Based on thefindings of Perrin (2000), Te Voert, Felling, and Peters(1994) and Cochran and Akers (1989), markers of religiosity,such as religious affiliation, are expected to be correlated torule-based ethics. Intuitively, it is argued that ethical pragma-tism may be an important influence because many may viewmisrepresentation to be a white lie under these circumstancesdue to the lack of tangible, direct, large, negative consequencesfor others. 25

We add the perceived altruism variable to investigate the im-pact of altruism on the preference for truth-telling. Given theemphasis on altruism in this sector, it is natural to ask whetheraltruism has a significant impact on misrepresentation.Wolfe’s (1998) review of altruism in the nonprofit sector dis-tinguishes between behavioral and motivational altruism.Evocative of the classical dichotomy of means versus ends,behavioral altruism considers the actual behavior of an orga-nization or individual, while motivational altruism focuses onthe goal or the intention. If the distinction between altruisticmeans and ends is acknowledged, it becomes clear that it isplausible that a higher level of perceived altruism may not nec-essarily translate to a higher aversion for dishonesty—espe-cially in the context of this situation where it is conceivablethat some may view this type of misrepresentation as a whitelie or “noble lie,” required as a means to a laudable end. Thelaudable end could, in this context, be retaining the NGO’sgood reputation, which may be viewed as critical for its sur-vival. Experiments carried out by Gneezy (2005) show thatmany individuals are pragmatic and consider the consequencesof their lies. In evaluating the fairness of a lie, individuals take

account of the size of the loss incurred by others as a result oftheir fibs.

8. EMPIRICAL ESTIMATES FOR SUPPLEMENTARYHYPOTHESES

The influence of disillusionment with donor rules is mea-sured by accounts and report submission fatigue. We testwhether the number of reports or accounts required annuallyby the NGO places a burden on them, affecting misrepresenta-tion. To capture the effect of exposure to competition for do-nor funding, a binary variable indicating whether or not theorganization had a grant at the time is included, as well as acontinuous variable showing the proportion of income derivedfrom grant and donor sources. 26

Four indicator variables are proposed to capture the influ-ence of government antagonism toward the NGO, namely,whether the NGO was involved in advocacy activities, whetherit viewed the government as a hindrance, whether the respon-dent felt that government staff was resentful toward the NGO,and whether the NGO believed that government dictates wereone of its main constraints.

It is a challenge to capture ineffectiveness and corruption. Itis difficult to identify such behavior, partly because NGOswould be inclined to conceal such markers. Additionally,many of the most corrupt or ineffective NGOs may lead ashadowy existence: present on the files of donors and theNGO register, but not to be found at the address or in thecommunities where they claim to work. Given that physicaltraceability was a precondition for inclusion in this survey, itis likely that some of the worst offenders are excluded fromthe sample. Given that the NGOs themselves are unlikely toreveal such markers of ineffectiveness and deceit, we examinea series of questions asked to the beneficiary focus groupsregarding the perceived impact and value of the NGOs. Thesevariables should identify severe cases of ineffectiveness andcorruption. 27

9. RESULTS

(a) Why are some NGOs dishonest?

Table 1 reports the findings from the Heckman selectionmodels. The paper presents two dependent variables, namely,whether the NGO misrepresents its community involvementand whether the NGO misrepresents its financial transparency.The first set of variables, reported in the top half of the table,are the factors associated with the misrepresentation of non-compliance. Because the focus is on misrepresentation, the sec-ond set of variables presented in the bottom half of the tableare discussed only briefly. The regressions show that commu-nity participation is associated with selfishness, number of staff,and third-party monitoring. The significance of the total num-ber of staff suggests that the ability or capacity to comply withthese norms may be important. Selfishness, as reported by thecommunity, is associated with a lower likelihood of involvingthe community, providing some credence for the hypothesisthat internal motivation matters. As the framework suggested,external motivation—here represented by monitoring—in-creases the likelihood of community participation.

A number of variables are shown to have a significant asso-ciation with financial transparency: the number of staff, the ageof the manager, the number of annual accounts the NGO is re-quired to submit, whether the NGO has a grant, the share of

Table 1. Heckman probit for community participation and financial transparency

Community participation Financial transparency

Coefficient P-value Coefficient P-value

Core regression: misrepresentation

Religious affiliation �0.844* 0.038 0.061 0.868Selfishness �0.089 0.698 �0.030 0.880Competence 0.150 0.456 0.022 0.924Years of existence 0.139* 0.046 0.046 0.547Years of existence squared �0.005* 0.029 �0.001 0.500Third-party monitoring 0.240 0.755 �1.259* 0.017Constant 0.286 0.873 �0.383 0.726

Selection: lack of compliance

Proportion of funds from local donations and fees 0.276 0.690Selfishness 0.187* 0.017 �0.160 0.214Total number of staff �0.0005* 0.022 0.000 0.902Age of NGO manager �0.015 0.203 �0.024* 0.073Third-party monitoring �2.003* 0.000 �0.309 0.311Number of annual accounts required to be submitted 0.099* 0.056Does the NGO have a grant? �1.549* 0.000Share of professionally qualified workers �0.767* 0.029Shortage of skilled staff cited as constraint 0.060 0.827Antagonistic relationship with government 0.597* 0.032Constant 0.511 0.408 1.364 0.017

Diagnostics

LR test of independent equations (rho = 0) prob. > chi2 0.02 0.892 0.16 0.688

Observations

Total number of observations 164 163Censored observations 80 103Uncensored observations 84 60

Note: Lack of compliance on community participation means that the NGO did not according to the community ask about their needs or ask for feedbackafter the project. Misrepresentation of community participation occurs when the community did neither, but claimed that they did both. Lack ofcompliance or financial transparency indicates that the NGO did not compile and share financial reports. Misrepresentation of financial transparencyoccurs when the NGO reported that they compiled financial data and that it was available to the public upon request, but they did not provide it whenrequested.* Coefficient significant at least at the 10% level of significance.


workers that are professionally qualified, and whether theNGO believed that the government was resentful toward it.

NGOs with grants are more likely to be transparent. In linewith the framework this is interpreted as evidence that incen-tives that align NGOs’ preferences with sector norms may beimportant. It has been argued that donors are the stakeholdersmost likely to require financial reports or, at least, some form offinancial oversight, or feedback. In this way, organizations thathave grants may have more incentives to keep careful financialrecords and make these available to the public upon request.

The results provide confirmation that capacity and abilitymay be important for determining how financially transparentan organization is. Both the age of the manager and the shareof workers that are professionally qualified are significant. Theregressions also suggest that NGOs that have fewer annualaccounting reports to submit are more likely to provide dataon the organizations’ financial positions, possibly indicativeof the impact of demands on these scarce, qualified human re-sources. Finally, NGOs that perceive their relationship withgovernment to be antagonistic are less likely to provide accessto this data.

Moving to the top half of Table 1, the determinants of mis-representation are discussed next. Based on the framework,the paper starts with the initial hypothesis that same factorsare associated with both the dimensions of misrepresentation.The results, however, suggest that there is no overlap in thefactors that are significantly associated across these twodimensions of misrepresentation.

NGOs that did not involve the community were more likelyto claim that they did so if they had existed for longer, whichthe theoretical model interprets as evidence of having more ofa reputation to risk by admitting that they did not. Religiousaffiliation, the proxy for the organization’s preference for hon-esty, is also significant.

In terms of financial transparency, the results imply that themain factor that matters is the likelihood of being caught—NGOs that are monitored independently are less prone tosay that they are financially transparent when in fact theyare not.

(b) Additional hypotheses

The next three tables test the additional hypotheses outlinedin Sections 6 and 7, starting with exposure to unreasonable do-nor demands and funding competition. Four variables areintroduced separately to approximate these influences:namely, having a grant, the proportion of revenue from thegrant, the number of reports the NGO is required to submitannually, and the number of annual accounts it is requiredto submit. Having a grant does not necessarily imply exposureto unreasonable donor demands, but it is used to capture theseeffects because it is a precondition for the influence of these ad-verse pressures. In line with expectations, NGOs that are reli-ant on grants (with a substantial proportion of revenue fromgrant income) are significantly more likely to misrepresenttheir financial transparency. Similar to the effect of reputation,

Table 2. Examining influence of exposure to unreasonable donor demands and funding competition


Religious affiliation �0.764* (0.052) �0.874* (0.025) �0.826* (0.022) �0.717* (0.045) �0.156 (0.667) �0.158 (0.670) �0.133 (0.715) �0.101 (0.801)Selfishness �0.043 (0.769) �0.121 (0.359) �0.123 (0.348) �0.148 (0.272) �0.032 (0.856) 0.050 (0.795) 0.026 (0.878) 0.067 (0.686)Competence 0.140 (0.545) 0.129 (0.527) 0.172 (0.357) 0.137 (0.497) 0.100 (0.643) 0.206 (0.317) 0.102 (0.650) 0.178 (0.395)Years of existence 0.276* (0.002) 0.214* (0.004) 0.134* (0.037) 0.103 (0.106) 0.052 (0.313) 0.046 (0.434) 0.076 (0.197) 0.051 (0.301)Years of existence squared �0.010* (0.004) �0.008* (0.004) �0.005* (0.017) �0.005* (0.033) �0.001 (0.336) �0.001 (0.424) �0.002 (0.275) �0.001 (0.270)Third-party monitoring 0.662 (0.476) 0.612 (0.450) 0.344 (0.591) 0.301 (0.615) �1.170* (0.037) �0.061 (0.858) �1.221* (0.031) �1.520* (0.004)Does the NGO currently have a grant? �1.578* (0.000) 0.903 (0.122)Proportion of revenue from grant �1.209* (0.005) 2.365* (0.002)Number of annual reports required to be submitted �0.045 (0.447) 0.010 (0.877)Number of annual accounts required to be submitted 0.277* (0.043) 0.117 (0.201)Constant 0.530 (0.656) 0.619 (0.546) 0.424 (0.656) 0.608 (0.557) �0.552 (0.598) �1.218 (0.239) �0.715 �1.108 (0.285)Observations 89 89 89 89 66 66 66 66Prob [Wald] > chi- 0.009 0.013 0.035 0.004 0.076 0.087 0.261 0.035Pseudo R2 0.281 0.206 0.142 0.181 0.126 0.113 0.105 0.140

* Coefficient significant at least at the 10% level of significance.

Table 3. Examining influence of ineffectiveness or fraud


Religious affiliation �0.830* (0.020) �0.771* (0.027) �0.728* (0.039) �0.778* (0.022) �0.120 (0.742) �0.232 (0.538) �0.130 (0.720) �0.244 (0.509)Selfishness �0.098 (0.447) �0.074 (0.613) �0.176 (0.178) �0.121 (0.348) 0.063 (0.718) �0.006 (0.970) 0.019 (0.906) �0.029 (0.869)Competence 0.175 (0.444) 0.137 (0.494) 0.299 (0.150) 0.202 (0.268) 0.011 (0.967) 0.207 (0.316) 0.132 (0.630) 0.251 (0.302)Years of existence 0.123* (0.043) 0.121* (0.047) 0.130* (0.033) 0.126* (0.042) 0.077 (0.175) 0.100* (0.079) 0.077 (0.173) 0.121* (0.057)Years of existence squared �0.005* (0.020) �0.005* (0.022) �0.005* (0.017) �0.005* (0.020) �0.002 (0.256) �0.003 (0.121) �0.002 (0.260) �0.003* (0.073)Third-party monitoring 0.377 (0.543) 0.325 (0.595) 0.353 (0.574) 0.368 (0.565) �1.276* (0.022) �1.129* (0.044) �1.222* (0.031) �1.100* (0.083)Beneficiary community satisfied with NGO? 0.045 (0.835) 0.129 (0.536)NGO could disappear without being noticed? �0.152 (0.251) 0.198 (0.112)NGO important part of life? �0.223 (0.251) �0.077 (0.791)Perceived value added �0.164 (0.740) �0.856 (0.149)Constant 0.199 (0.826) 0.863 (0.424) 0.82 0.357 (0.712) �0.908 (0.406) �1.633 (0.138) �0.484 �0.860 (0.441)Observations 89 89 89 87 66 66 66 63Prob [Wald] > chi2 0.043 0.042 0.034 0.058 0.195 0.067 0.261 0.291Pseudo R2 0.138 0.155 0.155 0.133 0.109 0.136 0.106 0.136


1272W

OR

LD

DE

VE

LO

PM

EN

T

Table 4. Examining influence of an antagonistic relationship with government


Religious affiliation �0.666* (0.059) �0.885* (0.017) �0.821* (0.024) �0.603* (0.098) 0.001 (0.999) �0.077 (0.839) �0.108 (0.769) �0.363 (0.356)Selfishness �0.194 (0.166) �0.102 (0.469) �0.118 (0.365) �0.057 (0.684) �0.025 (0.875) �0.007 (0.965) 0.034 (0.841) �0.021 (0.902)Competence 0.209 (0.351) 0.144 (0.464) 0.193 (0.313) 0.149 (0.473) 0.082 (0.708) 0.140 (0.579) 0.071 (0.745) 0.102 (0.678)Years of existence 0.075 (0.229) 0.123* (0.041) 0.124* (0.035) 0.114* (0.168) 0.060 (0.269) 0.073 (0.145) 0.077 (0.172) 0.063 (0.296)Years of existence squared �0.004* (0.077) �0.004* (0.041) �0.005* (0.018) �0.004* (0.217) �0.001 (0.348) �0.002 (0.180) �0.002 (0.247) �0.001 (0.447)Third-party monitoring �0.051 (0.934) 0.432 (0.482) 0.367 (0.562) 0.391 (0.546) �1.359* (0.021) �1.337* (0.030) �1.337* (0.019) �1.535* (0.018)Does NGO engage in advocacy activities? 1.051* (0.004) 0.398 (0.265)Does NGO describe government as a hindrance? 0.398 (0.259) �0.455 (0.202)Does NGO feel that government staff is resentfultoward it?

�0.022 (0.952) �0.293 (0.369)

Does NGO cite government dictates as one ofmain constraints?

�0.062 (0.861) 0.668* (0.085)

Constant 0.348 (0.755) 0.235 (0.828) 0.334 (0.737) 0.318 (0.773) �0.716 (0.496) �0.578 (0.638) �0.472 (0.396) �0.804 (0.504)Observations 89 84 89 78 66 61 66 59Prob [Wald] > chi2 0.002 0.075 0.044 0.479 0.171 0.252 0.208 0.181Pseudo R2 0.209 0.142 0.138 0.068 0.118 0.147 0.113 0.134


PR

OM

OT

ING

TR

AN

SP

AR

EN

CY

INT

HE

NG

OS

EC

TO

R1273


it is conceivable that having a grant, or being dependent ongrants, may increase the stakes for NGOs. It may generallybe easier for a third-party to detect and demonstrate misrepre-sentation in terms of community involvement than in terms ofa lack of financial transparency. The only evidence of report-ing fatigue having an impact is the significant and positivecoefficient on the number of accounts that has to be submittedper year in the model for community participation misrepre-sentation (see Table 2).

Table 3 explores the relationship between misrepresentationand corruption. The variables used are beneficiary assessmentsof the impact and value of the NGO, including questions onhow important the NGO is in the community, how satisfiedthe beneficiaries were with the NGO’s performance, whetherthey would notice if the NGO disappeared, and the estimatedvalue added by the NGO. These variables are meant to cap-ture serious cases of corruption or inefficiency that would beobservable to beneficiaries. There is only one significant vari-able: beneficiary satisfaction reduces the likelihood of misrep-resentation of community participation. The lack of othersignificant variables may be because the most serious casesof corruption have already been eliminated from this sample.Many briefcase NGOs without an office or permanent staff of-ten had to be dropped from the sample because they were dif-ficult to pin down and may have ceased to exist in allobservable ways for long periods between grants. Further-more, all NGOs that were not known in the parishes wherethey said they worked had to be removed because there wereno community focus group observations that could bematched to these organizations.

Finally, Table 4 considers whether hostile government rela-tionships can shape truth-telling and openness. This influenceis assessed using four variables, namely, the NGO’s engage-ment in advocacy, the NGO’s description of the governmentas a hindrance, the NGO asserting that it felt that governmentstaff was resentful, and the NGO citing government dictates asone of its main constraints. Advocacy activities appear to havea positive and significant effect on lying regarding communityinvolvement, while viewing government dictates as one of themain constraints has a positive and significant impact on thelikelihood of misrepresenting financial transparency.

10. CONCLUSIONS

The findings of this work caution against an overly naıveview of NGOs and also, specifically, an over-reliance on re-ported information when regulating, monitoring, or surveyingNGOs. They also emphasize the value of third-party monitor-ing and regulation in a sector where the effectiveness andappropriateness of monitoring and regulation are often dis-puted, at least partly because of the conventional association

of this sector with pro-social values and altruism. Even if thesetraditional notions about altruism were not contested, itwould not provide an alibi for secrecy and concealment. Goodintentions do not provide insurance against human fallibility.Individuals and organizations should be transparent andaccountable because such checks and balances can help iden-tify and rectify blunders, missteps, and ineffectiveness.

According to this analysis, there is no evidence that per-ceived altruism has a significant relationship with misrepresen-tation. The insignificance of altruism may suggest that manyselfless NGO employees may often be pragmatic in their ap-proach to ethics. The main focus of these altruistic NGOsmay be on their mission, with less emphasis on the means usedto reach that mission. At the very least it appears that altruis-tic NGOs have some tolerance for what they may view to beminor digressions in service of the greater good.

The analysis also suggests that the motive for misrepresenta-tion may depend on the norm in question. With regard tofinancial transparency, the analysis indicates that motives forlying are less robust correlations of the likelihood of lying thanthe opportunity for misrepresentation, i.e., monitoring. Forcommunity involvement, NGO misrepresentation is associatedwith the organization’s desire to “save face,” and a proclivityfor rule-based ethics.

Results are mixed for the supplementary set of hypotheses,but these tentative results may still be valuable, given that thisarea of research is characterized by a lack of empirical evi-dence. The results indicate that tension in relations with gov-ernment may be an important reason for misrepresentation.It is important to remain mindful of this legitimate concernregarding transparency in the sector when interpreting pat-terns of misrepresentation and also of course when formulat-ing transparency policies and guidelines for the sector.

The analysis also shows that there is partial and tentativesupport for a view that NGO misrepresentation can be attrib-uted to excessive and unrealistic donor demands. There is noevidence of a relationship between misrepresentation and theset of corruption and ineffectiveness indicators. This may bepartially due to the difficulties relating to surveying briefcaseNGOs and corrupt NGOs given that these organizations oftenexist only or mainly on paper.

Given the centrality of transparency for good governance ofthe sector, it is vital to know more about the prevalence oftransparency and to better understand why some organiza-tions may be reluctant to become more transparent andaccountable. Given the difficulty of generating robust resultsfrom a heterogeneous, cross-sectoral NGO data set with a rel-atively small number of observations, these results are viewedas exploratory and the authors want to encourage otherresearchers to investigate and test these findings using otherdata sources.

NOTES

1. Quoted in The Economist (Living with the enemy), 7th August 2003.

2. Veron et al. (2006) find that 20% of beneficiaries interviewed in theirinvestigations of corruption did not view the corruption in a negative light.They have accepted that bribes were necessary for securing service delivery.Platteau and Gaspart (2005) report that community members did not viewthe siphoning off of funds by a community leader as corrupt or illegitimate,as they argued that they have benefited from the leader’s efforts, and withouthis intervention their situation would not have improved.

3. Desai and Yetman (1995) emphasize the substantial discretion thatmost nonprofit organizations retain over the use of their funds. They notethat nonprofit organizations can decide how much they want to spend oncharitable activities vis-a-vis, for instance, administrative activities, fund-raising expenses, and staff compensation.

4. See Gibelman and Gelman (2004) for a survey of cases of nonprofitsector wrongdoing reported by the media. They suggest that there may bean association between cases of misallocation (or fraud) and a lack of


oversight. They report that board members often attributed theirignorance of such activities to their NGO’s philosophy of trust.

5. For instance, Mocan’s (2000) study of wages for child care workersfinds that there is a nonprofit wage and compensation premium. However,others such as Ruhm and Borkoski (2000) find that the nonprofit sectordoes pay less. Ruhm and Borkoski (2000) report a slight negative paydifferential (2–4%) adding that it is only observed in certain subsectors.

6. A large proportion of the labour force of developing countries work inthe informal sector for a small and/or variable wage or no pay. While thismay make it difficult to know where to draw the line between underem-ployment and unemployment, this conceptual question has no bearing onthe key argument here, namely that individuals in developing countriesoften have fewer options to generate income and may thus be more likelyto enter the NGO sector for monetary gain rather than for altruisticreasons.

7. Since the early 1990s, nongovernmental organizations (NGOs) havebecome major actors in the process of economic development: leadingcampaigns on pro-poor policy issues and delivering services to disadvan-taged communities. According to the Overseas Development Institute(1996) 10–15% of all aid to developing countries is channeled throughNGOs. Last year, both the UK government and OECD allocated 5% oftheir development budgets to NGOs. This growth in funding to NGOs inpart reflects frustration and impatience with failures of governments todeliver public goods and services to the poor, and a willingness of donorsto fund NGOs to by pass corrupt or inefficient governments

8. If they did not know the NGO, the group would evaluate anotherNGO.

9. In an appendix, available on request, bias introduced by the loss ofobservations due to the matching is investigated. Based on a number ofkey observables, there is little evidence of such bias. This does not,however, exclude the possibility that there may be bias based onunobservable factors.

10. The matched sample refers to the sample resulting from merging theNGO questionnaire sample with the community focus group interviewsample.

11. It is important to note that the authors do not think that the entirediscrepancy is attributable to misrepresentation. There are also otherexplanations for these discrepancies. For instance, it may be because ofgeographical variation: the NGOs’ answers may outline standard ortypical practices across the locations where the organizations work,whereas the community focus group answers were based on the impres-sions of a number of representatives from a specific parish. Some of thedifferences may also be due to misunderstanding the questions. However,these two explanations cannot account for the size and asymmetry of thisdifference.

12. There are four such cases for community needs assessments, andthree for feedback.

13. Tanzi and Shome (1993) lists lost time, stress, payments to taxaccountants and lawyers and visits to the tax office as part of the cost ofcompliance.

14. The evidence he cites includes the positive relationship between therate of compliance and enforcement mechanisms (e.g., information reportsand employer withholding), the correspondence of noncompliance rates toproxies for the traceability, deniability and ambiguity of items (Klepper &Nagin, 1989) and the higher deficiency rates in private companies(Hanlon, Mills, & Slemrod, 2007).

15. Transparency and participatory development are both included asaspirational norms in the code of conduct of the Ugandan NGO sector.Uganda has a strong tradition and long history of participatoryapproaches tied in with its culture of accountability. Uganda is celebratedfor these democratic norms and is associated with crucial advances intransparency and accountability to the community such as the landmarkexpenditure tracking survey in education and the follow-up experimentwith posting budgets on school gates.

16. The notational definition is altered to accommodate the binarydependent variables.

17. We opt for a two-stage process to reflect the chronological orderingof the two decisions. It is viewed as excessively cynical to assume that theNGO staff members anticipate having to lie about their choice in thefuture when they make the decision about compliance. Additionally, inmany NGOs the decisions may be taken by two different individuals.

18. Cited by Christensen (2004).

19. As cited in Slemrod (2007).

20. In terms of needs assessment the survey question asked how theNGO found out about the needs of the communities it helped, asking therespondent to circle all that applied from a list of eight options. Fourresponses (learn from the local government; learn from opinion leaders inhost community(ies); surveys run by the NGO; participatory workshopswith community members) were interpreted as a NGO claim that it hadassessed community needs. In the focus groups, the facilitator asked if theNGO has ever asked the community what activities it should undertakeand if yes, how they went about asking. To ensure that we erred on theside of caution all listed options were interpreted as verification that suchneeds assessment had occurred. In terms of community feedback thesurvey questionnaires asked how the NGO evaluated whether it fulfilledthe needs of the communities it assisted. The respondent had to circle allthat apply from a list of nine options and five responses (feedback fromopinion leaders in host communities; surveys run by the NGO; surveys runby other organizations; participatory review with community members;beneficiary assessment) were interpreted as the NGO reporting that it didask the community for feedback. The correlate for the focus group wascreated using a question that asked the community members whether theNGO in question has ever given your community an opportunity toprovide feedback on its work and how this was done. Again, to err on theside of caution all options were seen as verification that the NGO hadasked the community for feedback.

21. The variable is coded on a Likert scale, which is coded 1 if there isstrong disagreement with the statement that the NGO exists to serve thepurposes of its own staff rather than to help the community, 2 if there issome disagreement with the statement, 3 if there is neither agreement nordisagreement, 4 if there is some agreement, and 5 if there is strongagreement with the statement.

22. Although the decision-making process was implicitly modelled usinga utility framework, the observable equivalent, namely the decision itself,will be binary. This is in line with Spicer (1986). He emphasized the binarynature of the tax evasion decision—citing findings from Friedland, Maital,and Rutenberg (1978) and Spicer and Thomas (1982). He argues that thedecision to evade tax or not appears to precede the choice about whatproportion of tax to evade.

23. The answer is coded on a Likert scale.The variable was coded on aLikert scale, which is coded 1 if there is strong disagreement with thestatement that the NGO representatives are good at what they do, 2 ifthere is some disagreement with the statement, 3 if there is neitheragreement nor disagreement, 4 if there is some agreement, and 5 if there isstrong agreement with the statement.


24. In formal notation, if the “true” model is

d ¼ amþ bqþ cl; ð7Þ

it is assumed that the only influence that is problematic to capture with theavailable data is q and the candidate proxy for this variable is #. There aretwo conditions a suitable proxy must meet. Firstly, the variable needs dobe redundant in the structural or “true” model. In formal notation, thiscan be represented as

Eðdjm; q; lÞ ¼ Eðdjm; q; l; #Þ: ð8ÞIf a candidate proxy variable did not meet this criteria and it was signifi-cant when added to the “true” model, it would mean that the variable hasan impact and interpretation over and above that of a proxy for the omit-ted variable, which implies that it is not an appropriate proxy. A secondcondition, which is imposed frequently, but not always, requires that afterthe inclusion of the proxy variable, the other regressors should not be par-tially correlated with the unobservable variable for which it intends to

serve as a proxy. The coefficients will not be consistent without the secondcondition. However, even if the second condition does not hold, there maystill be an argument for using the proxy variable if its inclusion may reducethe bias of the estimates. According to Wooldridge, “This can usually beargued if [the proxy variable] is a reasonable proxy” (2002, p. 64). Also seeWickens (1972) in this regard.

25. This is in line with the work of Yankelovich, Skelly and White, Inc.(1984) cited in Spicer (1986), which reports that tax evasion was associatedwith a flexible definition of honesty.

26. Where values were missing for this variable, we replaced such valueswith subgroup averages, with two identified subgroups: NGOs with, andwithout grants.

27. Financial variables are avoided due to the potential high correlationbetween data availability, and corruption and ineffectiveness.

REFERENCES

Alexander, V. D. (1998). Environmental constraints and organizationalstrategies: Complexity, conflict, and coping in the nonprofit sector. InW. W. Powell, & E. S. Clemens (Eds.), Private action and the publicgood. New Haven: Yale University Press.

Allingham, M. G., & Sandmo, A. (1972). Income tax evasion: Atheoretical analysis. Journal of Public Economics, 1, 323–338.

Atkinson, R. (1989). Altruism in nonprofit organizations. Program on Non-Profit Organizations. Working Paper 145. Institution for Social andPolicy Studies, Yale University.

Barr, A., & Fafchamps, M. (2006). A client-community assessment of theNGO sector in Uganda. Centre for the study of African economiesworking paper, Oxford University. Journal of Developing Studies,42(4), 611–639.

Barr, A., Fafchamps, M., & Owens, T. (2005). The governance of non-governmental organizations in Uganda. World Development, 33(4),657–679.

Barr, A., Fafchamps, M., & Owens, T. (2003). Non-GovernmentalOrganizations in Uganda. Draft report to the Government of Uganda.

Bohnet, I., & Huck, S. (2004). Repetition and reputation: Implications fortrust and trustworthiness when institutions change. The AmericanEconomic Review, 94(2), 90, Papers and proceedings of the onehundred sixteenth annual meeting of the American economic associ-ation, San Diego, CA, January 3–5, 2004 (pp. 362–366).

Bolton, M. (2003). Setting the record straight. The Smart Company.Botchway, K. (2001). Paradox of empowerment: Reflections on a case

study from Northern Ghana. World Development, 29(1), 133–153.Burger, R. (2005). What we have learnt from post-1994 innovations in

pro-poor service delivery in South Africa: A case study-based analysis.Development Southern Africa, 22(4), 483–500.

Callamard, A. (2006). NGO accountability and the humanitarianaccountability partnership: Towards a transformative agenda. In L.Jordan, & P. Van Tuijl (Eds.), NGO accountability: Politics principlesand innovations. London: Earthscan.

Camerer, C., & Weigelt, K. (1988). Experimental tests of a sequentialequilibrium reputation model. Econometrica, 56(1), 1–36.

Chambers, R. (1996). The primacy of the personal. In Beyond the magicbullet: NGO performance and accountability in the post-cold war world.West Hartford: Kumarian Press.

Christensen, J. (2004). Asking do-gooders to prove they do good. New YorkTimes.

Civicus Civil Society Index Team (2006). Civicus Civil Society Index:Preliminary Findings Phase 2003–2005.

Clotfelte, C. T. (1983). Tax evasion and tax rates: An analysis ofindividual returns. The Review of Economics and Statistics, LXV.

Cochran, J., & Akers, R. (1989). Beyond hellfire: An exploration ofthe variable effects of religiosity on adolescent marijuana andalcohol use. Journal of Research in Crime and Delinquency, 26(3),198–225.

Cowell, F. A. (1985). The economic analysis of tax evasion. Bulletin ofEconomic Research, 37(3), 163–193.

Edwards, M., & Hulme, D. (1996). NGO performance and accountability.In Beyond the magic bullet: NGO performance and accountability in thepost-cold war world. West Hartford: Kumarian Press.

Fama, E., & Jensen, M. (1983). Agency problems and residual claims.Journal of Law and Economics, 26, 327–350.

Florini, A. M. (1999). Does the invisible hand need a transparent glove?The politics of transparency. In Paper prepared for the annual worldbank conference on development economics, Washington, DC.

Friedland, N., Maital, S., & Rutenberg, A. (1978). A simulation study ofincome tax evasion. Journal of Public Economics, 10, 107–116.

Gibelman, M., & Gelman, S. R. (2004). A loss of credibility: Patterns ofwrongdoing among nongovernmental organizations. Voluntas, 15(4),355–381.

Gneezy, U. (2005). Deception: The role of consequences. AmericanEconomic Review, 95(1), 384–394.

Hanlon, M., Mills, L., & Slemrod, J. (2007). An empirical examination ofcorporate tax noncompliance. In A. Auerbach, J. R. Hines, Jr., & J.Slemrod (Eds.), Taxing corporate income in the 21st century. Cam-bridge: Cambridge University Press.

Hansmann, H. (1980). The role of nonprofit enterprise. Yale Law Journal,89, 835–901.

Hansmann, H. B. (1987). Economic theories of nonprofit organization. InW. W. Powell (Ed.), The nonprofit sector: A research handbook. NewHaven, London: Yale University Press.

Horner, J. (2002). Reputation and competition. The American EconomicReview, 92(3), 644–663.

Kihato, C. & Kabemba, C. (2002). Enhancing policy implementation:Lessons from the education sector. Centre for Policy Studies ResearchReport 92. Johannesburg, Centre for Policy Studies.

Klepper, S., & Nagin, D. (1989). The anatomy of tax evasion. Journal ofLaw, Economics, and Organization, 5(1), 1–24.

Mansuri, G., & Rao, V. (2004). Community-based and -driven develop-ment: A critical review. World Bank Research Observer, 19(1), 1–39.

Martinelli, C., & Parker, S. W. (2009). Deception and misreporting in asocial program. Journal of the European Economic Association, 7(4),886–908.

Mazar, N., & Ariely, D. (2006). Dishonesty in everyday life and its policyimplications. Journal of Public Policy & Marketing, 25(1), 1–21.

McGann, J., & Johnstone, M. (2006). The power shift and the NGOcredibility crisis. The International Journal of Not-for-Profit Law, 8(2),65–77.

Michaelowa, K., & Borrmann, A. (2006). Evaluation bias and incentivesstructures in bi- and multilateral aid agencies. Review of DevelopmentEconomics, 10(2), 313–329.

Mocan, H. N. & Tekin, E. (2000). Nonprofit sector and part-time work: Ananalysis of employer–employee matched data of child care workers.NBER Working Paper 7977.

Mordaunt, J. (2006). The emperor’s new clothes: Why boards andmanagers find accountability relationships difficult. Public Policy andAdministration, 21(3), 120–134.


ODI (1996). The Impact of NGO Development Projects, OverseasDevelopment Institute Briefing Paper 1996 (2).

Olken, B. A. (2005). Monitoring corruption: Evidence from a fieldexperiment in Indonesia. National Bureau of Economics ResearchWorking Paper 11753.

Olken, B. A. (2006). Corruption perceptions vs. corruption reality. NationalBureau of Economics Research Working Paper 12428.

Perrin, R. D. (2000). Religiosity and honesty: Continuing the search forthe consequential dimension. Review of Religious Research, 41(4),534–544.

Platteau, J.-P. & Gaspart, F. (2005). Disciplining local leaders incommunity-based development. Centre for Research on the Economicsof Development Working Paper.

Rice, E. (1992). The corporate tax gap: Evidence on tax compliance bysmall corporations. In J. Slemrod (Ed.), Why people pay taxes. AnnArbor: University of Michigan.

Rose-Ackerman, S. (1978). Corruption: A study in political economy. NewYork: Academic Press.

Rose-Ackerman, S. (1996). Altruism, nonprofits, and economic theory.Journal of Economic Literature, 34, 701–728.

Rose-Ackerman, S. (1999). Corruption and government: Causes concernsand reform. Cambridge: Cambridge University Press.

Rowat, C., & Seabright, P. (2006). Intermediation by aid agencies. Journalof Development Economics, 79, 469–491.

Ruhm, C. J. & Borkoski, C. (2000). Compensation in the nonprofit sector.NBER Working Paper 7562.

Sandmo, A. (2004). The theory of tax evasion: A retrospective view.Norwegian School of Economics and Business Administration,December 2004. Discussion Paper 31/04.

Slemrod, J. (2007). Cheating ourselves: The economics of tax evasion.Journal of Economic Perspectives, 21(1), 25–48.

Sobel, J. (1985). A theory of credibility. The Review of Economic Studies,52(4), 557–573.

Spicer, M. W., & Becker, L. A. (1980). Fiscal inequity and tax evasion: Anexperimental approach. National Tax Journal, XXXIII, 171–175.

Spicer, M. W. (1986). Civilization at a discount: The problem of taxevasion. National Tax Journal, 39(1), 13–20.

Spicer, M. W., & Thomas, J. E. (1982). Audit probabilities and the taxevasion decision: An experimental approach. Journal of EconomicPsychology, 2(3), 241–245.

Stiglitz, J. (1999). On Liberty, the Right to Know, and Public Discourse:The Role of Transparency in Public Life, Oxford Amnesty Lecture,Oxford, UK.

Tanzi, V. & Shome, P. (1993). A primer on tax evasion. In Internationalmonetary fund staff papers (Vol. 40, No. 4, pp. 807–828).

Te Voert, M., Felling, A., & Peters, J. (1994). The effect of religion on self-interest morality. Review of Religious Research, 35(4), 302–323.

Tendler, J. (2000). Why are social funds so popular?. In S. J. Evenett, W.Wu, & S. Yusuf (Eds.), Local dynamics in an era of globalization.Oxford: Oxford University Press.

The Economist, Living with the enemy (2003).Thompson, E. (1980). Charity and nonprofit organizations. In K.

Clarkson, & D. Martin (Eds.), The economics of nonproprietaryorganizations. Greenwich: JAI Press.

Veron, R., Williams, G., Corbridge, S., & Srivastava, M. (2006).Decentralized corruption or corrupt decentralization? Communitymonitoring of poverty-alleviation schemes in Eastern India. WorldDevelopment, 34(11), 1922–1941.

Weisbrod, B. A. (1998). Institutional form and organizational perfor-mance. In W. W. Powell, & E. S. Clemens (Eds.), Private action and thepublic good. New Haven: Yale University Press.

Wickens, M. R. (1972). A note on the use of proxy variables. Econome-trica, 40(4), 759–761.

Wolfe, A. (1998). What is altruism? In W. W. Powell, & E. S. Clemens(Eds.), Private action and the public good. New Haven: Yale UniversityPress.

Yankelovich, Skelly and White, Inc. (1984). Tax attitudes study: Finalreport. Prepared for the Internal Revenue Service.

Zadek, S., & Gatward, M. (1996). Transforming the transnational NGOs:Social auditing or bust?. In Beyond the magic bullet: NGO performanceand accountability in the post-cold war world. West Hartford: KumarianPress.

Available online at www.sciencedirect.com

burger 2010

Documents