Does Government Funding Suppress Nonprofits' Political Activity?

Author(s): Mark Chaves, Laura Stephens, Joseph Galaskiewicz Source: American Sociological Review, Vol. 69, No. 2 (Apr., 2004), pp. 292-316 Published by: American Sociological Association Stable URL: http://www.jstor.org/stable/3593088 Accessed: 16/10/2009 21:43
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=asa. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to American Sociological Review.

http://www.jstor.org

Does GovernmentFunding Suppress Nonprofits'PoliticalActivity?
MarkChaves University ofArizona Joseph Galaskiewicz University ofArizona the a Autonomyfrom state has been considered corefeatureofAmericancivil society, and understanding consequences the has ofperceivedthreatsto thatautonomy beena centralthemein social andpolitical theory.Weengagethis themeby examining a is on What the effectof governmentfunding nonprofit specific question: organizations' political activity?Extanttheoryand researchidentifysome mechanisms which by government fundingmightreducenonprofit political activityand othermechanisms by whichgovernmentfunding enhancesuch activity.Weinvestigate relationship this might withtwodata sets: a nationalsampleof religiouscongregations a longitudinal and in Paul.Resultsacross thesedata sets sampleof nonprofit organizations Minneapolis-St. are consistentand compelling: relationship The betweengovernment fundingand is eitherpositiveor null;governmentfunding not does nonprofit political activity suppressnonprofit political activity. At least since de Tocqueville,autonomy from the statehas been considereda core feature of American civil society and a source of its many virtues. Recent historical and sociological research,however,shows that civil society and the state are deeply intertwined(Salamon 1995; Skocpol 1999; Smith and Lipsky 1993). it Indeed,fromsome perspectives, is moreaccurateto say thatat least partsof civil society are dependenton the stateratherthan autonomous from it, promptingquestions about the consequences of that dependencefor the natureand functioningof civil society. But "dependence" and "autonomy," to mention"civil society," not arevague concepts thatrequirespecificationto develop sound knowledge about the consequencesfor civil societyof moreor less dependence on or autonomy from the state. We contribute knowledgeaboutthis relationship to by focusing on a specific sector of civil sociea ty (nonprofitorganizations), specific form of dependenceon the state(governmentfunding), LauraStephens University ofArizona

Direct correspondence to Mark Chaves, Department of Sociology, University of Arizona, P.O. Box 210027, Tucson, AZ, 85721-0027 Data collection for the (mchaves@u.arizona.edu). National CongregationsStudy was supportedby a major grant from Lilly Endowment, Inc., and by grantsfrom Smith RichardsonFoundation,Inc., the Louisville Institute,the Nonprofit Sector Research Fund of the Aspen Institute, and the Henry Luce Foundation,Inc. Data collection for the TwinCities projectwas mainlysupported the NationalScience by Foundation(SES 80-08570, SES 83-19364, SES

88-12702, and SES 93-20929), and also by the Program on Nonprofit Organizations at Yale the University, NonprofitSectorResearchFundof the the AreaFoundation St. of AspenInstitute, Northwest Paul,the CenterforUrbanandRegionalAffairsat the Universityof Minnesota,the GraduateSchool at the University of Minnesota, the Rockefeller Brothers Fund, and the American Association of Colleges. The authors thank ASR reviewers, Steven Smith, BurtWeisbrod,and participants the Universityof in Arizona'sSocial OrganizationsSeminarfor helpful commentson earlierversions of this article.

AMERICAN SOCIOLOGICAL 2004, VOL. (April:292-3I6) REVIEW, 69

GOVERNMENTFUNDING AND NONPROFITS' POLITICAL ACTIVITY

293

and a specific type of consequence thatmight flow from variationsin dependenceon government (amountof political activity). dependon govManynonprofitorganizations ernment funding, and this financial dependence on public money has increased in recent decadesas moregovernment fundedservicesare delivered via grants and contracts with nonIn governmentalorganizations.1 1997, government funding accountedfor 37 percent of the nonprofit sector'srevenue,up from 31 percent in 1977 (Salamon 2002). The extent of this financial dependence on government varies considerablyacross types of nonprofit organizations, but few types receive no government funding. Even among religious congregations, 3 percentreceivegovernment money in support of social service activity (Chaves 1999). There is a substantialempiricaland normativeliterature on the natureand implicationsof financial connectionsbetweengovernmentandnonprofit organizations (Boris and Steuerle 1999; Gronbjerg 1993; Salamon 1995; Smith and Lipsky 1993). Nonprofit organizationsalso play an important advocacy role in American civil society (Berry 2003; Boris and Krehely2002; Jenkins 1987; Reid 1999). They have provideda vehicle, at least somewhat distinct from corporaand tions, government, politicalparties,through which individualsvoice concerns and attempt to exertcollectiveinfluenceon politicalprocesses. Nonprofit advocacy takes many forms, includinglitigating,lobbying,researching, publishing, testifying, and organizing collective actions. Political activity is the primarypurpose of some nonprofitorganizations,such as the National Rifle Association or the National Organizationfor Women.But many othernonwhose primarypurposeis profit organizations,

1 Inthis refers paper, organization" to "nonprofit the"religious, and charitable, educational, scientific, literary" organizations encompassed category by of States codeandtherefore tax 501(c)(3) theUnited eligible to receive tax deductiblecontributions. are Conceptually, religious congregations "nonprofit"because contheyareeligiblefortax deductible even as tributions, if theyarenotofficially registered with Revenue 501(c)(3)organizations the Internal of Service. three Approximately quarters congregations have formal501(c)(3)status,eitherthrough theirdenominations on theirown. or

something other than political advocacy (e.g., social service agencies or religious congregations), also engage in political activityon occasion. We focus here on the latter type of organization:nonprofits whose primarypurpose is somethingotherthanpolitical activity.2 Substantialgovernmentsupportof nonprofit organizations raisesimportant questionsabout the consequences of this supportfor nonprofits' political activity.Are governmentfunding and political activity competing or complementary features of nonprofit organizations? Do government-supported nonprofits engage in politics as actively as those nonprofitorganizations that do not receive governmentsupport?Most broadly,does the currenttrajectory of welfare-statedevelopmenttowardcontracting out services rather than providing them directly threaten to undermine historically importantforms of nonprofit-basedpolitical activity in American society? Understanding the natureof American civil society, especially in the context of substantialcontractingout of governmentfunctionsand ongoing concern aboutdeclining civic engagementandpolitical participation,requires answering these questions about the consequences of financial dependenceon governmentfor political activity in the nonprofit sector. This issue has been particularlyvisible in debatesoverthe currentBush Administration's "faith-based" "charitable or choice" initiatives aimedat redirecting fundsto further public support congregations'and other religious organizations' social service activities. Some critics of these initiativeshave expressedconcernthat increasedavailabilityof public funds for relisocial service activitywill gious organizations' dampen religious organizations' "prophetic voice" by increasing their financial dependence on government(e.g., Wallis 2001). This concern is a special case of the more general concernexpressedby Salamon(1995) thatgovernment funding might threaten nonprofit with a "loss of autonomyor indeorganizations 2Interms Internal of Revenue Service categories, ourfocusis on 501(c)(3)nonprofit not organizations, on 501(c)(4)nonprofit The are organizations. latter tax exempt,but contributions themarenot tax to and deductible, nonpartisan political activity be may theorganization's exclusive activity.

294

REVIEW AMERICAN SOCIOLOGICAL

pendence, particularly [with] dilution of the sector's advocacy role" (p. 103). Our investigation of governmentfunding'seffect on nonprofit political activity is motivated,on the one hand,by generaltheoreticalinterestin the interplay between the nonprofit sector's advocacy role and its deep entanglement with government and, on the otherhand,by a specific policy debateaboutthe consequencesfor religious organizations'political activities of potentially increased access to governmentfunding. Our results speakboth to this specific debateandto issues theoretical deeperandmore longstanding the consequencesfor civil society of concerning more or less autonomyfrom the state. and Despite its theoretical policy importance, little systematic research has focused on the relationshipbetween governmentfunding and nonprofitpolitical activity.This connectionhas been explored directlyonly throughcase studies (e.g., Fish 1973; Harris2001; Helfgot 1974; Nowland-Foreman 1998), small-N comparative case studies (e.g., Gittell 1980; Kramer 1981; Kramerand Grossman 1987), and one survey of nonprofits large enough to submit financial returnsto the IRS (Berry2003).3 The connectionhas been exploredindirectly through surveys of nonprofitboard members'individual political activity(O'Reganand Oster2002) and surveys of executive directors'subjective assessmentsof how government fundingaffects their organizations' autonomy (e.g., Netting 1982;Monsma1996).4Claimsof any sortabout the relationshipbetween governmentfunding and nonprofit political activity currentlyrest

on a very thin empiricalbase. We examine the relationshipbetween governmentfunding and nonprofit political activity by measuring the key variables in two organizationalsamples, each of which representsthe full size range of the relevant population (not just the biggest organizations),and each of which enablesus to assess this relationshipin the presence of relevant statistical controls. One of these samples is a panel study,which allows us to introducea longitudinaldimensionin our analysis. We begin by drawingon existing theory and researchto build a simple theoreticalmodel of the relationshipbetween governmentfunding and levels of nonprofit political activity. helpsus to idenAlthoughthe existingliterature tify plausible mechanisms by which government funding might affect nonprofit political activity, it does not yield an unambiguous answerto a basic question about the direction of that relationship:Does governmentfunding suppress, enhance, or have no effect on nonprofits'politicalactivity?We addressthis question with a nationallyrepresentative sample of religious congregationsand a longitudinalrepin resentative sampleof nonprofitorganizations conMinneapolis-St.Paul.We obtainstrikingly sistent results. BACKGROUNDAND THEORY

Investigatingthe relationshipbetween government fundingandnonprofitpolitical activityis complicatedby variationin the goals nonprofit organizationspursuethroughpolitical activity, the tactics they employ in pursuitof those goals, and the types of nonprofit organization 3 Only 42 percent of all organizationsregistered engaging in political activity.Nonprofit organfinan- izations might use letter-writing telephone and submit withthe IRSas 501(c)(3)nonprofits the In cialreturns. Indiana, onlystateforwhichreldirectedat legislatorsto advocateon campaigns have evantcomparisons beenmade,the nonprofits behalf of an interestgroup; they might testify returns constitute 10 per- at financial only submitting to legislativehearingsor mobilize individuals cent of the totalnonprofit (Gronbjerg demonstrateat a state population to influencepolcapitol with 2002:1747, organizations less 1772). Nonprofit are to revenue notrequired reg- icy on behalf of a client populationthat they than $5,000inannual mission with less than serve; they mightpromotea particular ister with the IRS; organizations are to revenue notrequired sub- by supportingor resistingparticular in annual $25,000 policies or mit a financialreturn (Form990). Religiousconlegislative agendas;they might engage in nonto are gregations not required registeror submit partisan around politicalissues; publiceducation their whatever size. returns financial or they might endorsepolitical candidates; they 4Theinitial waveof thisresearch Fish1973; (e.g., dilemmas might sponsor demonstrationsand marches. the Gittell1980;Helfgot1974)examined of political activityfaced by 1960s' community Although this is not an exhaustive list of the that political activities in which nonprofits might publicfunds. organizations received

GOVERNMENTFUNDING AND NONPROFITS' POLITICAL ACTIVITY

295

engage, it illustratesthe rangeof goals, tactics, and organizational types encompassedby nonprofit political activity.We might expect governmentfundingto havedifferentconsequences across types of nonprofitorganizations,political goals, and tactics. We focus on nonprofit organizationswhose primarypurpose is somethingotherthanpolitical activity or advocacy, but this restriction still leaves substantial variation in organizationaltypes andin politicalgoals andtactics.We arenot ableto comprehensively studythe effects of governmentfunding on every type of political activity across every type of nonprofit We organization. are, however,able to measure a rangeof political activitiesamonga varietyof types of nonprofit organization.Since distintypesof politicalactivguishingamongdifferent ities and differenttypes ofnonprofits does not, withinthe limits of our dataandmeasures,produce importantlydifferentresults, we will, in this section, use an undifferentiated concept of cat"politicalactivity"and an undifferentiated egory of "nonprofitorganization"ratherthan distinguish specific types of political activity and specific sorts of organizations.
FUNDING How GOVERNMENT MIGHT POLITICAL NONPROFIT ACTVITY SUPPRESS DEPENDENCE. most obvious RESOURCE The

and direct mechanism by which government funding might suppress nonprofits' political don't-bite-the-handactivityis straightforward, that-feeds-you resource dependence. That is, nonprofitsthat depend on governmentfor part of their livelihood might refrainfrom oppositional political activity or advocacy not welcomedby thatfundingsource.As Wolch(1990) put it: "As public funding becomes more centralto organizational survival,these groupsmay be essentiallyco-opted andbecome quiescent"
(p. 215). This clearly happens in some times and

places. Harris(2001), for example, tracedthe relative passivity of some of Chicago's major African American churches during the Civil RightsMovementto theirfinancialdependence on Mayor Richard Daley's patronage, noting that the mayor "used the largess of federal Waron Poverty monies designatedfor Johnson's of andthe city's department HumanServicesto
undermine clergy dissent" (p. 153). And non-

profit executives sometimes report examples

of outrightpunishment for advocacy by such meansas reducingthe number clientsreferred of to the agency, excluding the organizationfrom importantmeetings, and discontinuing grant support(Ryan, Miller, and Weiss 2002:14). Instancesof outright (or punishment threatof punishment)by governmentfundersfor political activitymay be rare,but thatraritydoes not prevent nonprofit executives from worrying about the potential negative consequences of political activity.Nonprofit leaderssometimes expressboth anxiety overthe loss of autonomy they fearwill come with government moneyand fear that theirpolitical activity will drive away government funders (Monsma 1996; Netting 1982; Reid 1999:301-2). One case study describedan organizationthat eventuallysplit into two separateorganizations-one for political advocacy and the otherto receive government funding-because of perceived"tensions between [the organization's] historyof advocaand its role as a government contractor" cy (Stone 1996:79). The organizationwanted to continue its political work, but it "was obviously reluctantto bite the hand that fed them, as one interviewee reported"(p. 79). And a recently produced pamphlet intended to help religious organizations assess whether they should pursuegovernmentfunds in supportof social service activity simply asserts without commentthatone of the downsides of doing so is thatgovernment the money "undermines traditional role of religion as prophetic critic of government;like every othergovernment-subsidized group,religionwill be less likely to bite the hand that feeds it" (Keeping the Faith, n.d.:4). There are many such examples (also, Berry 2003:74,106; Hudson 2002:412). They show that at least some nonprofit leaders fear thatconfrontational politicalactivitywill endantheir governmentfunding. In the presence ger of this assumption, government funding will suppressat least some types of nonprofitpolitical activity, and that suppression will occur even if the assumption is unjustified because actual instances of retributionin response to political activity are rare. Resource dependence also operates more subtly than via punishmentor fear of punishment for unwelcome political activity. Government funding causes several sorts of changes in nonprofit organizationalstructure and behavior;it shifts board composition and

296

REVIEW AMERICANSOCIOLOGICAL

increases complexity, formalization, professionalization, and bureaucratization (Froelich 1999; Gronbjerg 1993; Kramer 1981; Smith and Lipsky 1993; Stone 1996). These changes occur as organizations reorient themselves towardtheir governmentfunders,acquiringor creating the technical expertise and administrativeinfrastructure necessaryto secure,manage, and sustain that funding. Similarly, the containsexamples of noncase studyliterature profit organizations reducing their political activity because they redirect organizational attentionand energy away from advocacy and towardactivities necessary to sustain the governmentfundingon which they depend.A qualitativestudyofnonprofits in CuyahogaCounty, Ohio, for example, found that increased governmentfundingled nonprofitsto redirectattention, energy,and resources away from service delivery and advocacy and towardadministrative activitiessuchas grant-writing, fundraising, and documenting the community's need for services. In this way, increased nonprofit relianceon governmentfunding"waschanging the natureof their services; it had substantially diminished their capacity to be political" (Alexander, Nank, and Stivers 1999:460). A qualitativestudy of rape crisis centersand battered women's shelters, to give anotherexample, found that government funding and,in the professionalizedthese organizations the "undermined distinctivelyprivate process, and ideological characterof these programs," changingthemfromagenciesthat"openlychallenged the establishedpolitical andprofessionto with "moreeducated, al order" organizations more experiencedstaff, more client equity,and greater standardization of services" (Smith 1989:227; also see Matthews 1994). These examples illustratehow the formalizationand that government professionalization accompany can move nonprofitorganizations away funding from political advocacy. A resourcedependenceperspectivethus suggests at least two paths by which government politicalactivfundingmightreducenonprofits' for ity: via fearof punishment unwelcomepolitical activity,and via redirectingorganizational attention, energy,andresourcesawayfrompolitactivities ical activityandtowardadministrative made necessary by governmentfunding. This lattermechanism can be understoodas a type of coercive isomorphism by which resource

dependence,ratherthan law, pressuresorganizationsto act in a certainway.Unlike the don'tmechanism, this bite-the-hand-that-feeds-you lattermechanismalso implies thatgovernment funding would reduce political activity of all sorts, not just activity that challenges the funder.
LEGAL ENVIRONMENT. COMPLEX Legal rules-

and, more subtly, (mis)perceptions of legal rules-also might producea negativerelationship between government funding and nonprofit political activity, and one by which all sorts of political activity would be negatively affected,notjust activitiesthatdirectlychallenge the funder.There are legal limits on the extent to which organizations granted501(c)(3) status by the United States InternalRevenue Service (andthereforeeligible to receive tax deductible donations) may engage in political activity, whetheror not they receive governmentfunds. or Some activities,such as supporting opposing candidatesrunning for office, are completely otheractivities,suchas lobbying,are prohibited; permittedas long as they arenot a "substantial" part of an organization'sactivities; still other activities,such as public educationor voterregistration,may be pursuedwithout limits. The regulatorysituationis morecomplex for nonprofit organizationsreceiving government funds.By IRS rulesthese organizations not may use governmentfunds to directly supportlobbying, but they still may engage in lobbying thatis supported otherfunds.Additionallayby ers of regulation emanate from the Office of Managementand Budget, from funding agencies, and from state and local authorities.This legal environmentis both uncertainand complex. Recent Congresses, for example, have considered,but so far have failed to pass, legall islationthatwouldprohibit politicaladvocacy by any nonprofit recipient of government grants.5

5 Motivated thebeliefthat law by existing doesnot

nonprofits' sufficientlyregulate politicalactivity, have to limit legislators attempted further thepolitthosereceiving ical voice of nonprofits, especially funds and 1999:159, 162; government (Brody Cordes The which Reid 1999:316-8). IstookAmendment, constrain wouldfurther by lobbying organizations is examgovernment grants, thebestknown receiving

AND POLITICAL FUNDING NONPROFITS' ACTIVITY297 GOVERNMENT Sometimeslegal rules have directlycoercive political activconsequencesfor organizations' Jenkins(1987), for example,describes1960s ity. Action Projects,which "foundthat Community servto federalsupport restricted particular was ices and imposed strictfiscal controlsthat frequentlycollided with advocacyand indigenous mobilization efforts"(p. 304). But beyond the directly coercive effect of anti-advocacyrules attached to government funding streams, the complexities and uncertaintiesinherentin the legal environment governingthe politicalactivity of nonprofitorganizationsmay decreasein a more subtle way the political activity of government-fundednonprofit organizationscontemplating political activity. It is rare for a nonprofit organizationto face legal sanction because of its political activity, but, as with beliefs thatpoliticalactivitymightthreaten governmentfundingvia retribution,it is common for nonprofitexecutivesandboardmembersto believethataccepting fundinglegalgovernment ly restrictstheir political activity more than it actually does. The widespread worryamongnonprofitleaders aboutthe legalityof a publiclyfundedorganization engaging in political activity is evident in surveysof nonprofitexecutives,in case studand ies, and in the literature discourseproduced by and for nonprofits.A 2000 survey of executives in nonprofitsthat filed financial returns with the IRS foundthattwo-thirds(68 percent) wrongly believed thattheir organizationswere not allowed to lobby if part of their funding came from the government (Berry 2003). Follow-upinterviewswith some of these executives confirmedthe survey results. One executive's comment was typical both for being incorrect and for being asserted with confidence: "[We can't] be involved in lobbying as a nonprofit because we receive government grants"(Berry 2003:59-60). A case study of an organizationdevoted to childrenmakesclear helping mentallyretarded thatthe executivedirectorandthe boardsimply to pie of such an effort. It was first introduced in detailed discussion the of For Congress 1995. more governingnonprofit complex legal environment see organizations' politicalactivities, Reid(1999), disand (2003).Our (1999),andBerry Brody Cordes is cussionof the legalenvironment basedon these threesources. assumedthatacceptingstatemoney meantthat it had to curtail its political advocacy. Both board members and top managers apparently believedthatthe organization faced a dilemma: "[H]owcouldit leadadvocacyeffortsagainstthe cuts [in the state'sspendingon communityprograms]and still receive more than $4.0 million dollarsfrom the state?"(Stone 1996:75).They eventually resolved this dilemma by splitting into two organizations-one thatwouldengage in political activity and one thatwouldhold the moretelling,theAnnie statecontracts.6 Perhaps E. Casey Foundationsponsoreda reportcalled Tips on Advocacy for Publicly Funded This report,explicitlyaimedat comNonprofits. to munitygroups attracted political activitybut "worriedthat speaking out will . . . endanger yourgovernmentcontractsor nonprofitstatus," asserts: "Many agencies that use government money to provide services are concernedthat speakingup will endangertheir legal statusor put philanthropicand governmentfunding in some (Duitchn.d.:2-3). Furthermore, jeopardy" nonprofits that intend to engage in politics decline to accept government funds because they believe the funds will threatentheiradvocacy. This is illustrated by the founder of a child-advocacy groupin Ohiowho notedthathis organization"won't accept governmentfunds [because] we want to be a pure voice for children"(Hallett 2002:1C). In the spring of 2003, uncertaintyaboutthe legal rules governing the political activity of publicly funded nonprofits was evident in a controversy over what political activities are permittedby the nonprofit operatorsof federally funded Head Startprograms.In the midst of of debatein Congress aboutreauthorization Head Start, an associate commissioner of the Head Start Bureau in the United States Department of Health and Human Services wrote to all local Head Startprogramsinformpoliticalactivitiesaregoving themthat,"Your erned and, in many ways, restrictedor limited by Federallaw,"and telling them thatan advocacy group's effort to encourage Head Start

6 Thisis the sameschismatic menorganization tionedearlier. example This illustrates resource that and each dependence legalambiguity reinforce may otherin pushinggovernment-funded organizations activity. awayfrompolitical

298

REVIEW AMERICAN SOCIOLOGICAL

to providers speakoutagainstaspectsof the proreauthorization"appearsto encourage posed to Head Startprograms use Head Startprogram funds and/or staff in a mannerthat is in direct violation of the laws that govern your political activities"(Hill 2003). The advocacy group in question, the National Head StartAssociation (NHSA), is a nonprofit association of Head of The Startproviders. president thatassociation respondedwith a letterstatingthat, contraryto the commissioner's warning, "there is no or 'restriction limitation'... on a HeadStart proor its staff, parents, or board members gram from expressing views on legislation to Membersof Congress ... providedthey do not use federal funds in expressing those views." She declaredthatthe commissioner'sletterhas "hadthe effect of chilling the exercise of free expression by Head Start programsand their (Greene2003). representatives" This example nicely illustratesboth aspects of the causal path producedby legal complexity. On the one hand, public funding carries with it some real legal restrictionson an organization'spolitical activity.On the otherhand, there is widespread misunderstandingabout just what those restrictionsare, and that misunderstanding,especially when exploited by interestedactors, induces political caution by The nonprofits. aboveexchangeclearlyassumes beliefs aboutthe legal thatHeadStartproviders' rules attachedto their federal funding directly affecttheirpoliticalactivity,andit assumesfurbelieve thattheywould therthatmanyproviders be wise to refrainfrom such activity.The associate commissioner is trying to reinforce that belief; the NHSA presidentis trying to counter it. Many examples exist, but these few suffice to establishthe dualpointthatgovernment-supported organizationsare in a more restrictive legal environment than are other nonprofits when it comes to political activity, and, more the importantly, legal complexities and uncertaintiesin this arenasometimescause leadersof publicly funded nonprofits to believe that the legal environment is more restrictive than it really is (also see Reid 1999:301). The consequenceis enhancedcautionaboutpoliticalactivity in the presenceof governmentfunding.This causalpathwayrests on whatnonprofitleaders (who weigh the costs andbenefits of organizational political action)believe aboutthe poten-

tial negative consequences of such action for their organizations,and it will operate even if those beliefs are unfoundedand alarmist.7
MIGHT FUNDING How GOVERNMENT POLITICAL ACvrrTY NONPROFIT ENHANCE

Kramer (1987) identified several features of government-nonprofitrelations that might be expected to mitigate the suppressingeffect of politicalactivfundingon nonprofit government ity. These include: "the payment-for-service form of most transactions,which involves less controlthangrantsor subsidies;the diversityof agencyincomesources,whichlessens voluntary dependency on any one; the countervailing power of a voluntaryagency oligopsony (few sellers) of a service requiredby a government agency for its clients; political influence of the voluntary agency; and the lack of incentives and capacityfor stricteraccountabilityby government" 247). Some of these circumstances, (p. and othersnot partof this list, might go beyond merelymitigatingthe potentialnegativeeffects of governmentfunding on political activity to set in motion causal chains by which government fundingactuallyenhancesnonprofitpolitWe ical activity. focus on two suchmechanisms.
MONETARY SELF-INTEREST. Nonprofit leaders

receive governmentfundwhose organizations have an objectiveincentiveto increasetheir ing participationin the political process to protect or enhancetheir own fundingstreamsor otherwise improve their working conditions. Of course,nonprofitorganizationsoften advocate on behalf of theirclient populationswhetheror not they receive governmentsupportto serve those populations.But the incentive to engage in political advocacy increaseswhen organizationsreceivegovernment fundingbecausemany, perhapsmost, policy changesaimed at improving the lives of clients also will augment the resourceenvironment. Nonprofit organization's like organizations, othersorts of organizations,

the leadersconcerning limitsplacedon nonprofit of simply virtue being activity by nonprofit political whatevertheir funding 501(c)(3) organizations, of See sources. Berry(2003)foranassessment this phenomenon.

7There is substantial also misunderstanding among

ACTIVITY GOVERNMENTFUNDING AND NONPROFITS' POLITICAL

299

do not necessarily react passively to their resource environments;they attemptto shape and control those environments.When a governmentfunderis an important of thatenvipart ronment, we might therefore expect political activity to increase (Galaskiewicz 1985:292).8 documentsthatthis The case studyliterature incentive operates at least in some times and places. In their study of several dozen San Francisco nonprofits, Kramerand Grossman (1987) observedthatit was "exceedinglydifficult [for government]not to renew a contractif the provider... can mobilize communitysupof port"(p. 43). Mobilizing "hundreds elderly or disabled clients to pack a legislative chamber" helped these nonprofits press for additional government funds and more favorable payment systems (Kramer and Grossman 1987:46). Smith (1999) describeda politically active social service organizationthatreceives 70 percent of its budget from governmentand whose "legislativeprioritiesfocus on obtaining rate increases for its programs" (p. 196). Nowland-Foreman(1998) studied a nonprofit organizationwhose governmentsupport "did not stop [it from] taking part (and sometimes taking a lead) in lobbying the government,not just or even mainly for its own grantsbut also about legislation that affected the people it served" (pp. 108-9). Fish (1973) examined a that Chicago communityorganization engaged intensively in various kinds of political action to obtain and protect its federal funding.Even if a nonprofitorganization's governmentfunding is straightforward political patronage, involvementin patronagenetworksmight discourage political activity critical of the patron at the same time as it encouragespoliticalactivrallies, drives,candidate ity-voter registration

and so on-in supportof the patron'spolitical machine (Harris2001). There are many such examples.Receivinggovernment fundingclearly creates an incentive for nonprofit political sometimes activity,andnonprofitorganizations act on that incentive.
GOVERNMENTDEPENDENCEON NONPROFITS.

Resourcedependencybetweengovernment and can be a two-way street. Even if a nonprofits nonprofit is dependent on governmentfor its funding,government agencies mayalso depend to on thatorganization deliverneeded services. Government agenciesresponsibleformanaging and contracts need qualifiedandcapable grants partnerswith which to connongovernmental tract.If the numberof such partnersis limited in a given community or in a given arena,the situationis one of mutualdependencebetween governmentand the nonprofit organization(s) thatcan deliverthe service.In a studythatquantified the extent of perceived resourcedependence between nonprofit organizationsand the state of New York, Saidel (1991) found that "public-sectoragencies and nonprofit sector organizationsreportedvirtuallyidentical [levels of] resourcedependenceon each other"(p. in 546). In case studiesof nonprofitcontracting Kramerand Grossman(1987) San Francisco, foundthat"moreoften thannot, governmentis confronted... by a marketconditionin which thereare relativelyfew sellers of social services that governmentis mandatedto supply"(p. 36). In situationsof mutual dependence,both governmentand nonprofits control resources valuableto the other,neithercan dominatethe other,and governmentagencies and nonprofit organizations approach each other on more agenciesareless likeequalterms.Government 8Another of could version thisargument be develly to punish unwelcome political activity by of Mueller's account the (1979:156) opedbyadapting who bureaucrat" faces many terminatingcontracts,and, moreover,govern"budget-maximizing what ment supportednonprofits have an incentive incentives seekgovernment to moneybeyond is neededto provideservices.If additional public to engage in political activity that other nonthis bureau- profits lack. When government is dependent moneyis available, budget-maximizing cratwill devoteresources (thatis, the organization on its nonprofitcontractees,government-fundand will engagein lobbying other activity) ed political organizationswill be more effective than it. this in aneffortto secure Mueller developed conotherwisemightbe at using politicalactivto bureaucrats seeking they ceptwithrespect government but ity to promotethemselves,theirclients, or their moneyfor theiragenciesfromlegislatures, it mission. to as seemsapplicable nonprofit executives well.

300

AMERICANSOCIOLOGICAL REVIEW

out of the box, and so we can assess whetherthe pathby which governmentfunding suppresses We surely have not identified every mechanonprofitpolitical activity is stronger,weaker, nism by which government funding might or balanced by the path by which government increase or decreasenonprofitpolitical activifundingenhancesnonprofitpoliticalactivity.In ty.9The mechanismswe have identified, howtermsof Figure 1, ourempiricaltask is to deterever,oughtto be sufficientto concludethat,even mine whether the sign attachedto arrow e is as governmentfundingsets in motion processnegative,positive, or null. es likely to decreasenonprofits'political activfundAssessingthe totaleffectof government ities, it also sets in motion processes likely to ing on the likelihood of engaging in political increase such activities.Which set of processactivityis complicated,of course,by priorvaries outweighsthe other,in the aggregate,forms ables thatmight producea spuriouscorrelation the empiricalquestionwe address. betweengovernment fundingandpoliticalactivFigure 1 summarizesthe mechanisms disity. The most importantprior variable,reprecussed above. Government funding sets two sented at the bottom of Figure 1, is the competing causal paths in motion. One, the underlying inclination of an organization to path througharrowsa and b in Figure 1, will engage civil society, political processes, and tend to decreasenonprofitpolitical activity;the the state. This sort of inclination could lead otherpath,througharrowsc and d in Figure 1, organizationsboth to seek governmentfunds will tend to increasenonprofitpolitical activiand to engage in more political activity,and it ty. Theory and researchaboutnonprofitorgancould produce a positive correlationbetween izations imply that both of these paths are these two organizational behaviors,even when operative,butno extantresearchestablishesthe there is no causal connectionbetween them. It relativestrengthof these paths.We do not have also couldproducea null correlation, even if the adequatemeasuresof relevantinterveningvaritrue causal effect is negative. ables, so we cannot assess the relative imporOur analyses address this concern in three tance of the specific mechanisms associated ways. First, and most directly,in each data set with each path, nor can we estimate the magwe construct a community involvement scale nitude of the effects associated with arrows a underlying throughd. We have drawna black box around intendedto proxy an organization's tendencyto engage with civil society, political this part of the diagramto indicate thatwe are processes, and the state, and we include that unableto investigateits innerworkings. We are able, however,to observe covariation variableas a control in all analyses. This is a betweenthe government strongcontrol in regressionsof political activfundinggoing intothis box and the nonprofitpolitical activity coming ity on government funding, and including it increasesourconfidencethatany observedpositive or null effects of governmentfunding on produced politicalactivityarenot spuriously by 9 One additional mechanism whichgovernby in wherethe activ- priorvariablesoperating a situation enhance mentfunding political might nonprofit of This truecausal effect is negative. Second,we know funding activism. ity is directgovernment in mechanism notseemimportant the1980sand a fair amount about which types of nonprofit does werecollected, wemention organizationaremore or less likely to manifest but 1990s,whenourdata and it herebecause its historical Several a tendencytowardcommunityinvolvement, of importance. of federal War Poverty on programs the 1960s,such we drawon that knowledge to introduceaddiActionProgram the Legal tional relevant control variables. Third, we and as the Community set fundcommu- exploit the longitudinal nature of the ServicesProgram, outto directly and effortsin poorneighnityorganizing advocacy Minneapolis-St. Paul data by controlling for borhoods (Haveman 1977).In recentdecades,the at time 1, therebyinvestigating about politicalactivity has federal government beenmoreconcerned whetherreceiving governmentfunds at time 1 thanfacilitating activity limiting nonprofit political in at time 2. it, thoughin the 1990s,"rights-oriented advocacy predicts change political activity Theselongitudinal also enableus to address data to theiractivity the to largeenough report groups" rev- the question of causal order. 26 of on IRSstillreceived, average, percent their Another sort of selectivity process might andMosher-Williams enuefromgovernment (Boris betweengovernment fund1998:496). pushthe correlation
SUMMARY

GOVERNMENTFUNDING AND NONPROFITS' POLITICAL ACTIVITY (BlackBox)

30I

v v

Resourcedependency Complexlegal environment

Government Funding

. e

Political Activity

c \

v

v

self-interest Monetary Government on dependence nonprofits

/
/

+

Toward World Tendency Engaging"Outside"

Pursuit Activities Government of Will Fund Organizational Figure 1. The Relationship between Government Funding and Nonprofit Political Activity

ing andpoliticalactivityin a negativedirection. Since government funds are more likely to directly support services than to encourage advocacy, government-supportednonprofits mightbe less politicallyengagedsimplybecause self-selectinto politicallyinactiveorganizations government fundingstreams-not becausegovernment fundingsets in motionany of the causal pathswe describeabove.We are less concerned with this sort of self-selection. If we have misspecified ourmodels by omittingvariablesthat measure an organization'sprior tendency to both seek governmentfunds and refrainfrom political activity,then our governmentfunding coefficients wouldbe biaseddownward the and effects of governmentfunding would be more positive thanwe observe.This possibility does not threatenour substantiveconclusions. We analyze the relationship between governmentfunding and nonprofitpolitical activity in two distinct data sets: a national sample of congregationsand a longitudinalsample of Minneapolis-St.Paulnonprofitsin severalsectors. These dataconcern very differentorganizationalpopulations, they measurepolitical and

activity and government funding in different ways. The resultsthey yield, however,are strikingly consistent. ANALYSIS
FUNDING GOVERNMENT ANDPOLITICAL
ACTIVITYAMONGCONGREGATIONS

SAMPLE. This analysis uses data from the National Congregations Study (NCS), a survey of a nationally representativesample of religious congregations in the United States. The NCS gathereddata via a 60-minute interview with one key informant-a minister, priest, rabbi, or other leader-from 1,236 congregations. The cooperationratewas 85 percent;the response rate was 80 percent. The probabilitythat a congregationappears in this sampleis proportional its size. Because to entered the sample by being congregations named by respondents to the 1998 General Social Survey(GSS) who attendreligiousservices, larger congregations are more likely to appearin the sample than are smaller congrewerenominated gations.Some congregations by

302

REVIEW AMERICAN SOCIOLOGICAL

more thanone GSS respondent. Weightedonly to account for duplicatenominations,univariate statistics from the NCS describe the characteristics of congregations in terms of the numbers of religious service attendees who with those characteristics. attendcongregations In this case each attendeeis given equalweight. When the data are weighted inversely proportional to congregationsize, however,each congregationis given equalweight,regardlessof its size, andunivariatestatisticsdescribethe characteristicsof congregationsas establishments. Both kinds of numbersare substantivelyinteresting, andwe reportdescriptivestatisticsfrom both the individualand congregationperspectives.10
DEPENDENT VARIABLES. NCS contains The

data on eight types of political activity: (1) whetherpeople at worship services have been told, within the past 12 months, of opportunities for political activity,includingpetitioning campaigns, lobbying, or demonstrating; (2) whether voter guides have ever been distributed to people through the congregation; (3) had whetherthe congregation a group,meeting, class, or event, within the past 12 months, to organize or participatein a demonstrationor or marchto support opposea publicissue orpolicy, (4) to registerpeople to vote, (5) to discuss politics, or (6) to organize or participate in efforts to lobby elected officials of any sort; and whether,within the past 12 months, (7) an elected official or (8) anyonerunningfor office spoke to the congregation.All of these variables are dichotomouslycoded.
INDEPENDENT VARIABLES. Our key independ-

ent variableis whethera congregationreceived any funds from local, state, or federal government in supportof their social service activities. 1 Overall, 3 percent of congregations received governmentfunds in 1998 in support of their social service activities.

As noted earlier,we wish to controlfor congregations' underlying inclination to engage with the worldaroundthem,whichwouldmake them more likely both to receive funds from governmentand to be politically active. In an attemptto directlycontrol for a congregation's underlying tendency to engage with its community,we constructa communityinvolvement measurethatis the first factorextractedfrom a factoranalysisof seven variables:(1) the number of social service programs sponsored by the congregation,(2) whetherthe congregation has planned or conducted an assessment of communityneeds, (3) whethersomeone from a social service agency has been a visiting speaker, (4) whether the congregation has held an event to organize or encourage people to do volunteer work, (5) whether an academic or professor has been a visiting speaker, (6) whetheran outsidegroup,program, eventhas or used space in the congregation'sbuilding, and (7) whetherthe congregationhas conductedor A used a communitysurvey.12 valueof 0 on this measuremeansthata congregation an averhas level of communityinvolvementas reflectage ed by these seven activities;a value of 1 means that its level of communityinvolvementis one standarddeviationabove the average. We also control for a variety of other conrelatedto congregagregationalcharacteristics tions' inclination to seek governmentmoney and be politically active. Liberaland moderate Protestantcongregationsare more likely to be civically engaged than are conservative and evangelicalProtestantcongregations;Catholic congregations fall in between (Chaves 2004, chap. 4). Consequently,we distinguishamong RomanCatholic, modthreereligioustraditions: and erate/liberal Protestant, conservative/evangelical Protestant. The distinction between moderate/liberaland conservative/evangelical moderateProtestants drawnin standard is fashion.13

10See Chaves al. (1999)formoredetailsabout et NCSdataandmethods. 11Analyses of usingthe percentage a congregation'soperating budgetcomingfromgovernment variable as sources thekeyindependent yieldresults here. to similar thosereported substantively

congregation'sactivity thepastyear. over 13The largestgroups in the moderate/liberal are: Protestant Church, category UnitedMethodist Evangelical Lutheran Church in America, Church Church, Presbyterian (U.S.A.),Episcopal in of The andUnitedChurch Christ. largest groups theconservative/evangelical Protestant are: category

12Eachof theseitemsaskedabout the

GOVERNMENT FUNDING NONPROFITS' AND POLITICAL ACTIVITY303 These categories, based on denominational affiliation, do not exhaust the religious differences associated with congregations'underlying tendencyto engage the world aroundthem. Anotherindicator religiousculturethatcrossof cuts these religious traditions is a congregation's theological orientation,measuredby the followingitem:"Theologically speaking,would be considered more on the your congregation conservative side, more on the liberal side, or rightin the middle?"Althoughmuch of the liberal/conservative variation among congregations is represented by differences in denomination-based religious traditions,there also is substantialtheological variationwithin denominations.This variation,over and above denominational affiliation, is associatedwith a congregation's likelihood of being civically engaged: Self-describedliberal congregations are more engaged than are self-describedconservativecongregations, whatevertheirdenominations. We thereforeinclude as controls two dummyvariablesbased on this item. We also know from previous researchthat AfricanAmericancongregations predominantly are more likely than are predominantlywhite congregations to be engaged in certain key kinds of political activities, and they also are more likely to be open to governmentcollaborationandfunding(BeyerleinandChaves2003; Chaves1999).Wetherefore includea controlfor a congregation'sracialcomposition:a dichotomous variablecoded 1 for all congregationsin which at least 80 percentof the regularparticipants areAfricanAmerican. Congregationsareaggregationsof individuals as well as more or less formallyconstituted organizations.Since it is likely that congregations' activities largely reflect the preferences and activitiesof the people in them, it is important to control variablesknown to be associated with more civically engaged individuals. Relevantvariablesincludethe percentof a congregation's people with four-year college or higher degrees (logged), with household incomes over $100,000 in 1998 (logged), who areunderage 35 (logged),andwho areoverage 60 (logged). We also include dummyvariables indicatinglocation in the south or in a city. Finally, congregationswith more organizationalcapacityaremorelikely bothto seek governmentfundingandbe politicallyactive,so we include controls indicating several aspects of organizationalcapacity: size (logged number of regular participants), founding date, and whether the head clergyperson has at least a four-yearcollege degree.
ANALYSISSTRATEGY. estimateeight logisWe tic regressions.Each model uses all cases with nonmissing data on the dependent variable. When there are missing data on independent variables,mean values (in the case of continuous variables)or zeros (in the case of dummy variables)are substituted.Missing value indicatorsare then includedin each model to confirm that this strategy does not affect the substantiveresults.14 The logistic regressions use nonweighted data,andwe used diagnostictests recommended by Winship and Radbill (1994) to look for misspecification errorrelatedto the probability-proportional-to-sizefeature of the sample. These tests indicatedthat,althoughinteraction termsinvolving size shouldbe includedin several of our models, the coefficients attachedto the government-funding variableareneversubstantively altered by including these interactions. Including all of these terms would unnecessarily complicate the presentationof our results, so we leave them out. RESULTS.Table 1 presentsbivariate crosstabulations comparingthe rate of engagementin with and politicalactivityamongcongregations without government funds. The first two columns present this comparisonfor individuals in religious congregations(that is, weight-

Southern Baptist Convention, Lutheran Church-Missouri of Synod,andAssemblies God. There 50 congregations theNCSthat either are in are non-Christianof indeterminate or affiliation. religious Thisset of 50 congregations includes Jewish synaand as gogues,Moslem mosques, Buddhist temples, wellas other A of congregations. dummy varitypes ableindicating non-Christian this is category included in themodels, we donotreport coefficient but its becausethe heterogeneity the category in renders thecoefficient substantively meaningless.

14Models exclude caseswith that all data missing on anyvariable results similar to yield substantively thosereported below.

304

REVIEW SOCIOLOGICAL AMERICAN

Funds: NationalCongregations Political Table 1. Congregations' Study, Activityby Receiptof Government 1998 Peoplein Congregations of EachFunding Type thatalso Do PoliticalActivity(%) of Congregations EachFunding Type thatalso Do Political Activity(%)

With With Without Without Government Government Government Government Variable Funding Funding Funding Funding 26 34 36 64** Toldpeopleat worshipservicesaboutopportunities for politicalactivity 13 17 43* 26 voterguides Haveeverdistributed Havehada group,meeting,class or eventto: 21 31 9 or 52** in or Organize participate a demonstration march 13 11 36** to Get peopleregistered vote 7 6 41** 12 Discusspolitics 6 5 11 39** in or Organize participate effortsto lobbyelected officialsof any sort Havehadas a visitingspeaker: 4 19 16** 6 for Someonerunning office 13 6 12 33** An electedgovernment official 59 41 86** 60 in Participated at leastone of the aboveeight politicalactivities in of tests wereperformed Note. Chi-square only for the differences the "percent people"partof the table.The in of numbers the "percent congregations" of the tableuse dataweightedto undothe probability-proportionpart tests for dataweightedin this to of al-to-sizefeature the NCS sample.It is not appropriate calculatechi-square way. * p <.05; **p <.01

ing the dataonly to accountfor duplicatenominations of congregations); the second two columns present the comparisonfor religious congregations as organizationalunits without respect to size (that is, applying a weight that undoes the probability-proportional-to-size natureof the sample). The pictureis clear. On all eight items, people in congregations that more fundsaresignificantly receivegovernment thanarepeople in congregationswithout likely that such fundingto be fromcongregations also in thatkind of politicalactivity.In every engage case the difference is sizeable and statistically significant. The table's bottom line tells the basic story.Eighty-sixpercentof people in congregations with government funds are from congregationsthat engaged in at least one of with only 60 these politicalactivities,compared withoutgovpercentof people in congregations ernmentfunds. The picturedoes not change much when we look at congregations as organizationalunits

without respect to size. Here, 59 percent of congregations with government funds have engaged in at least one of these political activities, compared with only 41 percent of congregations without government funds. The bivariate difference remains substantial for 5 of the 8 items.15 Overall, congregations with government funding appear to be more engaged in politics than do congregations without such funding. This basic pattern is sustained by the multivariate results reported in Table 2. There is much that is interesting in this table, but the primary results, and the ones relevant for our argument, appear in the first row. Receiving government funding for social service activity

15 A substantially reduceddifferencewhen applying the weight indicates an interactionbetween size andreceiptof governmentfundsin theirrelationship with those particular political activities.

GOVERNMENTFUNDING AND NONPROFITS' POLITICAL ACTIVITY

305

does not make congregations less politically for active,evenwhencontrolling a varietyof relevant prior variables, including a community involvement scale that is strongly associated with eachpoliticalactivity.16 the government All funding coefficients are positive; four are significant at least at the p < .05 level. A fifth coefficient (in the "political opportunities" model) approaches significance(p = .065). And the magnitude of these effects is notable: Controlling for other things, congregations fundsaremorethanthree receivinggovernment times more likely to organize a demonstration or march, for example, than are congregations without governmentfunds. Interpretedliberally, these results suggest that, at least for the political activities we have measured,congregations receivinggovernment funding engage in more political activity than do congregations fundreceivingno government ing, even with relevant variables controlled. Interpreted more conservatively, this set of results makes clear that congregationsreceiving government fundingengage in no less political activity than congregations without that support.There is no evidence here that government funding of human service activities suppressescongregation-based political activity.
GOVERNMENT FUNDING ANDPOLITICAL ACTIVITY TWINCITYNONPROFITS AMONG

survivors,respectively.Face-to-faceinterviews were conductedwith the chief executiveofficer or top administrator each organization. of Those organizationsthatwere small and/ornew were most likely to leave the panel (see Galaskiewicz andBielefeld 1998 for more detailon this data).
DEPENDENT VARIABLES. indicators of nonTwo

profits' political activity are availablein these data. Our primary dependent variable is the yes/no response to an item asking whetherthe organization"hadengaged in lobbying efforts on your own" in the previous four years. This questionwas askedin 1984, 1988, and 1992. In 1994, respondentswere asked about lobbying over the past two years. The question was not asked in 1980. In 1984, 1988, 1992, and 1994, 23 percent,24 percent,23 percent,and 29 percent of informants,respectively,said that their had organization engagedin lobbyingduringthe relevanttime interval. The other measure of nonprofits' political activity is a scale rangingfrom 0 to 3, indicating whetherorganizationshad launchedmajor public or communityrelationsefforts aimed at politicians,governmentagencies, or legislative bodies. Informants reported the presence or absenceof effortsfor each of thesethreetargets; the scale sums "yes"responses (yes = 1). This measureis availableonly for 1984 and 1988.17
INDEPENDENT VARIABLES. in the congreAs

data from a longitudinal study of 501(c)(3) organizationsin the Minneapolis-St. Paul metropolitan area. In 1980, Galaskiewiczand Bielefeld (1998) drew a one-in-five stratified systematic sample of 326 nonprofitsheadquartered theTwinCities in area. The sampling frame was metropolitan developed using the Cumulative List of Organizations published by the Internal RevenueService;neithercongregations prinor vate foundationswere includedin the sampling frame.Data were obtainedfrom 229 organizations, a responserateof 70 percent.Researchers returnedto the field in 1984, 1988, 1992, and 1994, interviewing 201, 174, 162, and 156 panel

SAMPLE. This analysis uses

gations analysis above, the key independent variable is receipt of government funding. informants the Organizational reported amounts of money theirorganization receivedfrom city, county, state, and federal sources for 1979, 1980, 1983, 1984, 1987, 1988, 1991, 1992, 1993, and 1994. Grants and contracts were for included,but reimbursements Medicaid or Medicarewere not. The amountsreceivedwere convertedinto 1994 dollars,and two-yearaverages were calculated to smooth year-to-year fluctuations.Thus we have data on the amount of money each organization receivedfromfour differentgovernmentsources for five two-year

the involvement in scale Including community the modelsreducesthe magnitude the governof mentfunding on coefficient, average, 25 percent. by

16

(SD= 1.13)for 1984and.977(SD= 1.18)for1988. Thereliability coefficients are (alphas) .792for1984
and .784 for 1988.

17Themeanof thispolitical scaleis .831 activity

Nati Table2. LogisticRegressions Congregations' of Activitieson Presence/Absence Governmental of Political Funding: Variables Independent Government Funding variable) (dummy ReligiousTraditionsa MainlineProtestant RomanCatholic Orientationb Theological Conservative Moderate Variables Demographic Percent with B.A. (logged) Percent (logged) rich Size (logged) Percent young(logged) Percent (logged) old Morethan80%African-American South Urban Clergywith college degreec date Founding Political Distribute Voter Groupto Guides Demonstrate Opportunities .656 (.356) .047 (.190) .621** (.201) .010 (.226) -.193 (.222) .074 (.183) -.184 (.159) .172 (.155) -.240 (.253) -.230 (.203) .706** (.212) -.217 (.138) .067 (.149) .055 (.161) -.001 (.001) .284 (.355) -.690** (.208) -.840** (.220) .233 (.250) -.256 (.252) -.337 (.195) -.104 (.176) .602** (.169) .473 (.295) -.223 (.222) .099 (.227) -.410** (.151) .025 (.161) .138 (.174) .002 (.002) 1.268** (.386) -.513* (.260)
1.171**

Voter Registration DiscussPolit 1.036* (.412) -.059 (.311) .388 (.306) .496 (.344) .028 (.339) -.435 (.286) -.166 (.252) .776** (.241) .742 (.475) -.261 (.323) 2.179** (.282) -.346 (.227) .614* (.266) -.573* (.248) -.001 (.002) .881* (.417) .635* (.293) -.096 (.328)

(.240) .727* (.320) .523 (.314) .216 (.251) -.188 (.202) .240 (.192) 1.326** (.406) -.286 (.273) .196 (.278) .013 (.179) .284 (.195) .233 (.214) .000 (.002)

-.776** (.297) -.481 (.278)

.211 (.319) -.023 (.248) .737** (.243) .058 (.435) -.071 (.336) (.318) -.411 (.225) .655* (.258) -.063 (.264) -.003 (.002)

.881**

GOVERNMENTFUNDING AND NONPROFITS'POLITICAL ACTIVITY

307

* c) 1:$

00

a

00

o c~

11 N 0,~ Cl C1l k 0 Cl (-I 0 ,-,
Cl4

CL,

.,,
d0~

0

~d
0 c
,t -4C1 1 * *-- r? Cl-,

0)OCa ;00

" C-)

0~ O)

Cl10 'I 01 0l C \ 0 0\ CIA

O

0o

o

0

0

* N Cl \Oc --( 0 C' 0\ C)r C 'IC 0 Cl Cl

0,
;-C

0CC~'C
"

0 ca 0
0 0
0) 0

o

?rr +-' o

o0

0

;-

00a;~
-00 O bD o )00)

.

-- , cli 4

~

-

Cl

00m

C0

0~Ob~c d

~

.,.
C.)

r

Cl4

C~ l
Cl4 N0)0)*)

N

C)

F1~~~~0
0)

-~
0) .8 o

OE-E
0)0o
=1 0 0) * 00 Cl 4 Cl "0 m~~~~00 00O tn Cl 05

ti~~

0

. ; Ci 4 06 C
I '-'0
Nc)

0) 0)

0 0
0 0) 0

00 -) 00 000

0
all

Cl rn 00 Cc N 00

00 OCl Cl4

0008)
a a, OF0)e

d

*8
"0 Cl4
C)

9

)

0
(U

;-00" N 00 N` c'i 00 Cl '.0 Cl Cl

0
rA

-~ 01

8 ~~ ~~0) ~~ ~8 "
0 )

~0d

0) Cl4
C'4

~0"

k

0)
0 0)0 600 0a

a

Cd)

0

9

~t
0

0

0\

000O

ClV

~
0)

8;-

o

-

0E 0)j

0)0)0

0) 0) 00 0)

periods: 1979 1980, 1983 1984, 1987 1988, 1991 1992, and 1993 1994.Forty-three percent of organizationsreceived governmentsupport in 1979 1980, 47 percent in 1983 1984, 48 percentin 1987 1988, 52 percentin 1991-1992, and 53 percentin 1993 1994. We use these datato construct threemeasures of governmentsupport: dummyvariableindia cating if the organizationreceived any public variableindicating perthe money,a continuous cent of an organization's budget coming from governmentsources, and a continuousvariable indicatingthe absolute dollaramountreceived from governmentsources.18 As in the congregational analysis,it is imporin tantto controlfor variation nonprofits' underlying inclinationto engage with the state. For 1984 and 1988, we construct a community involvement scale from items asking about major public or community relations efforts directedat the generalpublic, communityleadfunders,orthelocal media. ers, non-government Each of these targets was asked about separately,and the scale simply sums (range= 0 to 4) the numberof targetsat which efforts were we directed.19 Furthermore, exploit the longitudinal nature of these data by controlling whether the organizationlobbied (or directed public relations efforts at governmenttargets) in the previous observationperiod. We also controlfor othercharacteristics related to an organization's inclinationto seek governmentmoney andbe politically active. Most important,we include two measuresof organizational capacity: size (dollar expendituresin millions) and age. The organizationsin this data set operatein a wide range of functional arenas, and they serve differentpopulations.It seems likely that the connection between government funding and political activitywould work differentlyin different subsectors of the nonprofit sector. Health andwelfare organizations, example, for

0"

a

n
0 ) "
p4. C

0

k8
0) 10 0
OCCC

0 0 00 +-)
00 a).

0'.');~ 0 00
C 0 -)

~

0) "0l 0) 0) 0

z

U U

O~~~~(

18Analyses variables disusinga set of dummy fromfederal, of counstate, tinguishing receipt funds similar yieldresults ty,andcitysources substantively to thosereported below. 19 involvement scale Themeanof thiscommunity is 1.76(SD= 1.50)in 1984and1.80(SD= 1.53)in 1988.Thereliability coefficients are (alphas) .756in 1984and.781in 1988.

308

AMERICAN SOCIOLOGICAL REVIEW

both aremore likely thanculturalorganizations andbe politically to receivegovernment funding active. And organizationsserving clients or the generalpublicaremorelikelythanorganizations serving members both to receive government funding and be politically active. Member-oriented organizationstend to be more insular. To ensurethatourresultsarenot an artifactof the mix of arenasandbeneficiariesrepresented in this particulardata set, we use two sets of dummy variable controls. One set represents seven activityareas:healthand welfare,educaculture,science, tion, legal services, recreation, and housing and urban development. Respondentswere asked to rankthese activity areas in terms of organizational priorities;our dummyvariablesare coded 1 if an organization to named that arenaas "mostimportant" them. Manyorganizations a rankof 1 to morethan gave that one area,indicating morethanone areawas "most important."Consequently,these seven items are not mutuallyexclusive, so there is no We reference category. havethesedataforall time periods. Respondentsalso were asked "who benefits from the activities of your organization." memResponseswerecodedintofourcategories: the bers, clients/patients/students, public, and We other/mixed.20 representthis information withthreedummyvariables, usingmember-servas ing organizations the referencecategory. 20Insomeyears question open-ended; in this was were other respondents givenspecific response years categories.
Table 3.

RESULTS. 3 presentsthe bivariaterelaTable between governmentfundingand lobtionship bying activity by comparing,in each interval, the organizationsreceiving governmentfunds with the organizations receivingno such funds. Both for organizations receivedgovernment that fundingand for those that did not at time 1, we of showthepercentage organizations lobbied that betweentime 1 andtime 2. It is clearthat,except forthe 1988-1992 interval, that organizations had public money were significantly and substantively morelikely to engage in lobbyingactivity thanthose thatdid not have public support. The is in the same direction for the relationship 1988-1992 interval,but the magnitudeof the differenceis smallerand does not reachstatistical significance. a Table4 presents seriesof logisticregressions in whichthe dependent variables the dichotoare mous measuresof lobbyingin the yearspriorto 1984, 1988, 1992,and 1994.Withone exception, allpredictor variables werelaggedfouryearsin the firstthreesets of resultsandtwo yearsin the last set of results.The exception is the community involvement after scale,whichwas notmeasured 1988. For each time period,we estimatedthree models:one inwhichthekey independent variable is a dummyvariable the representing presenceor absenceof government fundingin the previous one variable period, in whichthekey independent is the percentageof revenuecoming from governmentsourcesin the previousperiod,and one in which the key independent variable is the absolute dollaramount millions)comingfrom (in in We government theprevious period. emphasize

Nonprofits' Political Activity by Receipt of Government Funds: Nonprofit Organizations in Minneapolis-St. Paul, 1980-1994 Public Funding in Stated Year No Yes No Lobbying between 1980 and 1984 (%) Lobbying between 1984 and 1988 (%) Lobbying between 1988 and 1992 (%)

between Lobbying
1992 and 1994 (%)

Year 1980

13 (109)** 35 (92) 16 (88)* 33 (86) 20 (80)
25 (80)

1984

Yes
1988 No

Yes
1992 No

17 (75)**
40(81)

Yes in Note: Numbersof cases appear parentheses.
* p < .05; ** p <.01 (two-tailed tests)

GOVERNMENTFUNDING AND NONPROFITS' POLITICAL ACTIVITY

309

thatthesemodelsincludecontrols bothfora general tendencytowardscommunityinvolvement and(exceptformodel 1)whether organization the lobbiedin the previousperiod.21
21

As in the congregational analysis, there is much interesting detail in this table, but we limit ourattentionto the findings (Table4, first
between 1976 and 1980 so we cannot include this controlin Model 1.

Recall that we do not have data on lobbying

Table 4.

Logistic Regressions of Public Charities' Lobbying Activities on Government Funding Variables: Nonprofit Organizations in Minneapolis-St. Paul, 1984-1994 Model la Lobbying in 1984 .553 (.502) .017** (.005) .143 (.180) .452 (.688) -.456 (.675) 1.719 (1.339) -7.059 (15.396) -.315 (.818) -.423 (1.446) .187 (.961) -.614 (.649) -.685 (.730) -.085 (.784) -.005 (.011) -.002 (.024) .467** (.145) .494 (.676) -.438 (.660) 2.394 (1.412) -7.338 (15.282) -.259 (.812) -.252 (1.284) .124 (.957) -.480 (.620) -.501 (.695) -.173 (.768) -.014 (.013) -.009 (.026) .525** (.142) .681 (.731) .157 (.609) 1.945 (1.278) 1.165 (.958) -.204 (.774) .127 (1.673) .144 (.977) .190 (.933) 1.541 (.952) 2.012* (1.010) .006 (.010) .001 (.022) .394* .442 (.748) .046 (.608) 1.714 (1.309) 1.106 (.958) -.047 (.755) -.174 (1.802) -.063 (.996) .099 (.927) 1.492 (.950) 2.142* (1.005) .009 .013 (.007) .126 (.159) .708 (.731) .153 (.609) 1.824 (1.305) 1.053 (.949) -.113 (.766) .154 (1.608) .137 (.986) .329 (.903) 1.561 (.945) 2.142* (.986) .005 (.011) .000 (.026) .410** (.155) 1.308** (.471) -3.849** (1.141) 151.424 173 Model 2a Lobbying in 1988

Independent Variables Government Funding in Previous Period Receives any government 1.006* (.417) funding Percent of revenue from public sources Amount from public sources (in $ million) Arena of Nonprofit Activity Health / welfare .567 (.691) Education -.490 (.677) 2.232 Legal (1.339) Recreation -7.429 (14.881) Culture -.515 (.825) Science -.306 (1.360) .080 Housing (.968) Intended Beneficiariesc Clients -.618 (.632) Public -.553 (.709) Mixed -.149 (.782) Other Controls -.011 Age Expenditures (in millions) Community involvement scaled Lobbying in previous period Constant -2 log likelihood Number of organizations (.011) -.005 (.026) .461** (.145)

-2.177** (.804) 170.797 198

-2.178** (.802) 166.388 198

-1.854* (.772) 176.217 198

(.010) .004 (.021) .408* (.158) (.155) 1.215* 1.115* (.477) (.479) -3.961** -3.876** (1.129) (1.136) 150.853 146.916 173 172

on (continued nextpage)

3Io

AMERICAN SOCIOLOGICAL REVIEW

Table 4. (Continued.)

Variables Independent Government Fundingin PreviousPeriod Receivesany government .381 funding .564) of Percent revenuefrom publicsources Amountfrompublic sources(in $ million) Arenaof Nonprofit Activity Health/ welfare .297 (.733) Education -.666 (.693) 1.977 Legal (1.123) Recreation -2.262 (1.490) Culture -.966 (.912) Science -4.215
(20.897)

Model 3a Lobbyingin 1992

Model4b Lobbyingin 1994
1.730*

(.677) -.006 (.008) .001 (.014) .280 (.732) -.624 (.691) 2.021 (1.141) -2.366 (1.486) -1.026 (.909) -4.192 (21.004) -.014 (1.065) .938 (1.669) .930 (1.704) 2.190 (1.738) .333 (.733) -.477 (.685) 1.496 (1.042) -2.135 (1.525) -1.058 (.939) -4.275 (21.047) -.333 (1.062) .727 (1.633) .777 (1.673) 2.090 (1.702) 1.207 1.041 (.793) (.767) 1.119 .954 (.716) (.696) -1.998 -1.529 (1.702) (1.636) -.261 -.829 (1.675) (1.771) .680 .971 (1.032) (.967) .009 -.057 (1.569) (1.647) .873 1.339 (1.730) (1.330) -1.067 (.997) 1.397 (.890) .082 (1.893) .012 (.014) .045 (.040) -.070 (.734) 4.192** (.764) -4.190** (1.245) 91.394 153 -1.086 (.970) 1.151 (.847) -.616 (1.861) .007 (.010) .095 (.246) 1.114 (.759) .975 (.698) -1.338 (1.587) -.880 (1.794) 1.042 (.972) -.099 (1.637) 1.363 (1.323) -1.036 (.956) 1.219 (.837) -.722 (1.853)

Housing Intended Beneficiariesc Clients Public Mixed OtherControls Age (in Expenditures millions) involvement Community scaled in Lobbying previous period Constant -2 log likelihood Numberof organizations

-.037 (1.075) .816 (1.575) .884 (1.612) 2.171 (1.650) .032** (.010) .003 (.012) .359* (.173) 1.880** (.525) -3.976* (1.703) 116.696 157

.031** .028** (.010) (.011) .003 .001 (.012) (.014) .289 .349* (.168) (.164) 1.958** 1.788** (.536) (.529) -4.093* -3.971 (1.785) (1.753) 116.567 115.235 157 157

.005 .002 (.761) (.015) .075 .075 (.080) (.089) .058 .071 (.195) (.190) 3.880** 3.848** (.697) (.688) -3.216** -3.170** (1.093) (1.094) 97.804 98.394 152 153

Note:Numbersin parentheses standard errors. are a variables laggedfouryears. are Independent b are variables laggedtwo years. Independent c The reference categoryis member-serving organizations. d The involvement scale is measured community only in 1984 and 1988.The 1984 valueis used in Model 1;the 1988valueis used in the othermodels.
* p < .05; ** p < .01 (two-tailed tests)

ANDNONPROFITS' GOVERNMENT FUNDING POLITICAL ACTIVITY 3II

three rows) directlyrelevantfor our argument. In short, the effects of governmentfunding on the likelihood of nonprofit lobbying are either significantlypositive (3 coefficients) or null (9 coefficients). Whetherexpressed as a dummy variable,as the percentageof an organization's budget, or as an absolutedollaramount,receiving governmentmoney at time 1 eitherhas no

effect on the likelihood of lobbying in subsequentyears, or it increases that likelihood. Table 5 contains a similar analysis but uses the count of public relations efforts aimed at governmenttargets in 1988 as the dependent variable.The firstthreerows show thatboth the percentage of an organization'srevenue that comes from governmentand the absolute dol-

of Table 5. OLSRegressions PublicRelationsEffortsAimedat Government (1984-1988) on Targets in Paul Government Variables: Funding NonprofitOrganizations Minneapolis-St. Variables Independent
Government Funding in Previous Period:

PoliticalActivity,1984-1988 .231 (.187) .010** (.003) .211** (.064) .095 (.252) .041 (.223) .300 (.534) .217 (.331) -.480 (.289) .793 (.521) .878* (.341) .005 (.259) -.049 (.293) .245 (.324) -.003 (.004) -.006 (.007) .390** (.053) .194** (.073) -.048 (.330) .361 172 -.136 (.246) -.064 (.215) .074 (.517) .114 (.318) -.452 (.268) .582 (.503) .701* (.329) -.087 (.244) -.108 (.278) .234 (.307) -.002 (.004) -.005 (.006) .383** (.051) .109 (.071) .120 (.318) .415 171 .003 (.242) -.013 (.217) .047 (.526) .134 (.323) -.507 (.270) .742 (.506) .765* (.332) .026 (.243) -.069 (.282) .282 (.307) -.007 (.004) -.008 (.007) .386** (.051)
.181** (.067)

Receivesanygovernment funding,1984 of Percent revenuefrompublicsources,1984 Amountfrompublicsources(in $ millions), 1984 Arenaof Nonprofit Activity(1984) Health/ Welfare Education Legal Recreation Culture Science Housing Intended Beneficiaries(1984)a Clients Public Mixed OtherControls Age, 1980 (in Expenditures millions), 1984 scale (1984-88) involvement Community Politicalactivity,1984 Constant Adjusted R-square Organizations (N) are errors. Note: Numbersin parentheses standard a The reference organizations. categoryis member-serving

.134 (.325) .396 172

312

AMERICAN SOCIOLOGICAL REVIEW

lar amountcoming from governmenthave significant positive effects on public relations efforts targetedat government. All in all, these results are strikinglysimilar to the resultsfor religiouscongregations. There is no evidence here thatgovernmentfundingof human service activities suppresses nonprofits' political activity.Thatrelationshipis either positive or null. The possibilityremains,however,thatreceiving government money at time 1 in partreflects an organization'sself-interestedlobbying at a previous point in time. We addressthis causal order issue by estimating a set of multiple regression equations in which the dependent variables are (1) whether the organization received public funds at time 2, (2) the total amountof incomeat time 2 frompublicsources, and (3) the percentageof income at time 2 from public sources. In each time period, each of these three dependentvariableswas regressed on our lobbyingindicator, then on the scale and of government-directed public relationsactivity. That is, government income in 1984 was level of politiregressed on the organization's cal activity between 1980 and 1984, government income in 1988 was regressed on the level of politicalactivitybetween organization's 1984 and 1988, andso on. Controlvariables,all measuredat time 1, arethe same as in previous analyses,exceptthatgovernment fundingmeas1 also are included.In all, we estiures at time mated24 equations:3 indicatorsof government funding regressed on each of 2 measures of political activity in each of 4 time periods.The results (table available upon request) do not paint a clear picture about whether nonprofit politicalactivityis an effectivetacticfor increasing government funding. In only 5 of the 24 equations did we find a significant (at the p < .05 level) positive effect of lobbyingon increases in a public fundingstream,1 coefficient was significantly negative, and the rest were null. These results add to our confidence that the positive effects of governmentfundingreported in Table 4 are not wholly produced by a reversed causal orderin which organizational lobbyinggenerates government fundingfornonprofits, but there is a hint of reverse causation that futureresearchmight profitablyexplore.

LIMITATIONSOF THE TWO ANALYSES We find thatgovernmentfundingdoes not suppress nonprofit political activity in either the Minneapolis-St. Paul sample or the congregations sample. Perhaps,however,we fail to find evidencefor suppression this activitybecause of of limitationson ourdataand analysis.We consider five potentiallyrelevantlimitations. First,we measureonly some types of political activity-eight types of political activityfor congregations and two for noncongregational nonprofits.It remainspossible thatgovernment fundingsuppressespoliticalactivitiesotherthan the specific ones we measure.A second limitationis related:We do not know the contentof any of this political activity,and so we are not able to assess the extent to which nonprofit political advocacy is narrowly self-interested versus broadly public-spirited.Perhapsif we were able to distinguish self-interested from public-spirited activitywe would find thatgovernmentfundinggenerallysuppressesthe latter but not the former. Although we measure only some types of political activity, and althoughit remainspossible that we have missed a type of political activitythatis suppressed governmentfundby ing, our measurescover a range of activitiesfrom lobbying to organizing demonstrations and marchesto sponsoringspeeches by candidates-and the consistencyof ourresultsacross all of these activitiesprovidessome confidence thatthepatterns observeherearenot specific we to any particular kind of political activity. Relatedly,although we have no direct knowledge about the specific content of any of this political activity,the resultsareno differentfor activities, like organizing demonstrationsor that marches, seem morelikelyto be public-spirited thanself-interested.Moreover,the distinction betweenpublic-spirited self-interested and actionin the nonprofitarenacan be difficult to draw.If a nonprofit'spolitical activity increases the funding streamflowing to that organization, but that increasedfundingalso meansthat more meals are served,more beds are providclasses areheld,andso on, ed,morejob-training it is not clear that such political activity should be considered self-serving ratherthan publicspirited. From this perspective, our lack of knowledge about the content of these organizations' political activity seems less limiting for our argumentthan it otherwise might.

GOVERNMENTFUNDING AND NONPROFITS'POLITICAL ACTIVITY

313

A thirdlimitationis that, with the exception of one ordinalscale, we measureonly the presnot ence or absenceof eachpoliticalactivity, the extent of it. It remains possible that government fundingreduces the intensitywith which nonprofits engage in political activity without reducing the likelihood that they do it at all. However,if governmentfundingin fact generally suppressedthe intensitywith which organizationsengagedin politics,it presumably would reducesome levels of intensityto zero, andthis should have left a trace in our dichotomous measures. Fourth, we have presented results only on linear relationshipsbetween the percentageof a nonprofit'srevenuethat comes from government and the likelihood thatthe nonprofitpursues political activity.But perhapsthe effect of politicalactivgovernment fundingon nonprofit becomes decreasingly positive (but never ity dependenceon that negative)as organizational source increases;or perhapsthe effect funding of government fundingon politicalactivityactually becomes negative at high levels of organizational dependence. We looked for these nuances by including in our models either a of transformation thepercentage-oflogarithmic revenue variableor its square,but we did not find consistent evidence for any sort of curvilinearity in the relationshipbetween the perrevenuethatcomes centageof an organization's and fromgovernment its level of political activity.This maybe becausewe do not have enough organizationsin either data set at high enough levels of dependence on governmentfunds to But discern curvilinearity.22 even if curvilinexistswhengovernment earity fundinglevels are our results show that at extraordinarily high, levels of relianceon governmentfundingcharof the and acterizing vastmajority congregations noncongregational nonprofits, government funding does not suppresspolitical activity. Fifth, we have data on noncongregational nonprofits only from one metropolitan area.
22In the congregationsdata,only 7 congregations

Although we would not claim that the Twin Cities arerepresentative Americancities, and of there is substantial variation across although cities in the structure the nonprofitsector,we of do not see any reason to suppose that the key we to relationships examineherearepeculiar the Twin Cities. Moreover,our argumentdoes not rest on the Twin Cities data alone. We rest our argument on the consistency of results from these two very different data sets, and across several differentpolitical activities, nonprofit subsectors,andtypes of government entityfrom which nonprofitsreceive financial support.In light of this consistency we are less concerned than we might otherwise be about limitations applying to either of these data sets if taken alone.23 CONCLUSION We have engaged the complex theme of the of natureandconsequences civil society'sautonfrom the stateby examining one sector of omy civil society (nonprofitorganizations), sort one of dependenceon government(financial), and one kind of consequence (political activity). We described mechanisms by which government funding might suppressnonprofitpolitical activity,andwe describedothermechanisms by which public funding might enhance that activity. Our results suggest either that these competing mechanisms balance each other or that the mechanismsby which public funding enhances political activity are somewhat stronger. Some observers believe that the resourcedependenceand legal ambiguitiesthat come with governmentfunding decreasenonprofitpolitical activity,but government-funded nonprofits also have incentives to be politically active-incentives that, as we noted earlier, encompass both self-interestand public spiritedness in ways difficult to disentangle.Add to these incentives government'sdependencyon nonprofits to deliver services, and also add otherdynamicswe have not elaborated, such as the 2000 surveyof a national sampleof nonprofits largeenoughto file with financial information theIRS.Thissurvey also founda nullrelationship between percentage the of an organization's budgetcomingfromgovernment it andtheextent which lobbies to officials government (Berry 2003:90).
23 We referred earlierto

of received morethan10percent theirincomefrom and in government, no congregation this sample of received morethan27 percent its annual income that In Cities from source. theTwin data, between only of in 15 and20 percent organizations eachperiod of received morethantwothirds theirincomefrom government.

3I4

REVIEW AMERICAN SOCIOLOGICAL

need fornonprofitexpertisewhen government's it comes to formulating policy, andthe result is a set of forces that seems at least to balance the resource dependencies,legal ambiguities, and otherfactorsthatpush in the oppositedirection. There is much to be said for "following the money" and attendingto the power inequities producedby resourcedependence.Nothing in this study contradictsthe sociological wisdom reduces that,all else equal,resourcedependence the capacityfor autonomousaction.But all else is not equal. When a nonprofit organization receives governmentfunding,forces otherthan resource dependenceare set in motion, and in the presence of such forces, resource dependence does not necessarilyproducepolitical quiescence. In this arenait appearsto be a mistake to inferloss of autonomyfromresourcedependence. of Whatever detailedoperations the mechthe anisms set in motion by governmentfunding, these mechanismsnet out to eithera positive or a null effect on political activity.We are tempted to conclude thatgovernmentfundingin fact enhances nonprofits' political activity, but a of morecautiousinterpretation ourresultsis that does not suppress it. Yet governmentfunding even this more cautious interpretation implies thatthe nonprofitsectorremainsa viable vehicle of citizen advocacy,even in the presence of its increasingreliance on governmentfunds to carryout its core work.At least in the specific manifestation have examined,civil society's we capacityforpoliticalactiondoes not seem to be reduced by its increased reliance on government funding. Ourconclusion is not affectedby our exclusive focus on nonprofits whose primary purpose is somethingotherthanpolitics-on 501(c)(3) ratherthan 501(c)(4) organizations.By focusing on 501(c)(3) organizationswe surely have understated the extent to which nonprofits but engagein politicalactivity, we do not see any reason that this focus should undermine our primaryconclusion aboutthe effect of government funding on nonprofit political activity. Indeed,it seems reasonableto conjecturethat, if governmentfunding does not reduce political activityamongorganizations which polfor itics is a secondaryactivity,neitheris it likely to reducepolitical activity in organizationsfor which advocacyis the top priority. Establishing

the truth of this conjecture,however,is a task for futureresearch. Whateverthe broadly theoretical import of our results, our congregations' results speak directly to ongoing debate concerning "chariinitiativesintendtablechoice"or "faith-based" social ed to directmorepublicfunds supporting services to religious organizations,including religious congregations.Critics of these initiatives, fromboththe political rightandleft, have expressed concern that increased government funding might threaten the mission and distinctiveness of religious organizations.Critics from the right worry that governmentfunding might threatenwhat some believe to be a distinctively "holistic"approachto social services manifest in religious organizations that maintain their autonomy from government.24 Critics from the left worry that congregations and other religious organizations might lose their "prophetic voice"-their capacity to engage in political activity critical of government-if they receivegovernment fundsto support their social service activities. Our results suggestthatthe lattercriticismis largelyunwarranted. Ourevidencesuggeststhatputtingmoregovernmentmoney into the "blackbox" of Figure 1 does not reducethe amountof political activity coming out of the box, but we have offered only limited insight into the innerworkingsof thatbox. Giventhe rudimentary stateof knowlabout the relationship between governedge mentfundingandnonprofit the politicalactivity, we takehereis a necessaryearlystep in the step effort to develop systematic knowledge about this relationship.Futureresearchshould move beyond our analysis to examine the relative of by importance the variousmechanisms which governmentfundinginfluencesnonprofitpolitical activity, the possibility that the relative importanceof those mechanisms shifts at different levels of nonprofit dependenceon government or for different types of nonprofits, andthe possibilitythatthe system changesover time as governmentalters the signals it gives aboutthe proprietyof nonprofitpolitical activity.Researchalongthese lines wouldcontribute

24 See Chavesand Tsitsos (2001) and Chaves to (2004,chap. forevidence 3) contrary thiscriticism fromtheright.

GOVERNMENTFUNDING AND NONPROFITS' POLITICAL ACTIVITY

315

valuable new knowledge about the complex relationship between civil society's autonomy from the state and its capacity to act in the political arena. Mark Chavesis ProfessorandHead of the Sociology ofArizona.Muchof his Departmentat the University workspans the boundarybetween the sociology of religion and thesociology of organizations.His most recent book, Congregations in America (Harvard UniversityPress), will appear in May, 2004. Laura Stephens is a doctoral candidate at the are University ofArizona.Her mainresearchinterests in the areas of religion and organizations.Her dissertation examines the impact of social service involvementon religious congregations. Joseph Galaskiewicz is Professor of Sociology and has a courtesy appointmentin the School of Public Administration and Policy at the University of Arizona.He is currentlydoing researchon the marketfor youth services in the Phoenix-Mesa metropolitan area withfundingfrom the National Science Foundation. He also serves as President of the Association Researchon NonprofitOrganizations for and Voluntary Associations and is co-editing (with Daniel Brass,HenrichGreve,and Wenpin Tsai)a speon cial issue of the Academyof Management Journal BuildingEffectiveNetworksthat will bepublishedin late 2004.

REFERENCES
Alexander, Jennifer, Renee Nank, and Camilla Stivers. 1999. "Implicationsof Welfare Reform: Do Nonprofit Survival StrategiesThreatenCivil Society?" Nonprofit and Voluntary Sector Quarterly28:452-75. Baptist Joint Committee on Public Affairs and The Interfaith Alliance Foundation. KeepingtheFaith: The Promise of Cooperation, The Perils of Government Funding: A Guide for Houses of Worship.N.d. Washington, DC: Baptist Joint Committee on Public Affairs and The Interfaith Alliance Foundation. Berry, Jeffrey M. 2003. A Voicefor Nonprofits. Washington,DC: Brookings Institution. Beyerlein, Kraig and Mark Chaves. 2003. "The PoliticalActivities of Religious Congregationsin the United States." Journalfor the ScientificStudy ofReligion 42:229-46. 1998. Boris, ElizabethandRachelMosher-Williams. "Nonprofit Advocacy Organizations:Assessing the Definitions, Classifications, and Data." Nonprofit and Voluntary Sector Quarterly 27:488-506. Boris, Elizabeth and Jeff Krehely. 2002. "Civic and Participation Advocacy."Pp. 299-330 in The

State of Nonprofit America, edited by L.M. DC: BrookingsInstitution. Salamon.Washington, Boris, Elizabethand C. Eugene Steuerle,eds. 1999. Nonprofits and Government:Collaborationand Conflict.Washington,DC: UrbanInstitute. Brody, Evelyn and Joseph J. Cordes. 1999. "Tax Treatmentof Nonprofit Organizaitons:A TwoEdged Sword?" Pp. 141-76 in Nonprofits and Government: Collaboration Conflict,editedby and E.T. Boris and C.E. Steuerle. Washington,DC: UrbanInstitute. Chaves,Mark. 1999. "Religious Congregationsand Welfare Reform: Who Will Take Advantage of 'Charitable Choice'?" American Sociological Review 64:836-46. . 2004. Congregations in America. Cambridge,MA: HarvardUniversityPress. Chaves, Mark, Mary Ellen Konieczny, Kraig Beyerlein,andEmilyBarman.1999."TheNational Methods,and CongregationsStudy:Background, Selected Results."Journalfor the ScientificStudy of Religion 38:458-76. Chaves, Mark and William Tsitsos. 2001. "Congregationsand Social Services: What They Do, How They Do It, andwith Whom."Nonprofit and Voluntary Sector Quarterly30:660-83. Duitch, Suri. N.d. Speak Up: Tipson Advocacyfor Publicly Funded Nonprofits. New York: Center for an UrbanFuture. Fish, John Hall. 1973. Black Power/WhiteControl: The Struggle of the WoodlawnOrganization in Chicago. Princeton, NJ: Princeton University Press. of A. Froelich,Karen 1999. "Diversification Revenue Strategies: Evolving Resource Dependence in and Nonprofit Voluntary NonprofitOrganizations." Sector Quarterly28:246-68. Galaskiewicz, Joseph. 1985. "Interorganizational Relations." Annual Review of Sociology 11:281-304. Galaskiewicz,JosephandWolfgangBielefeld. 1998. in NonprofitOrganizations anAge of Uncertainty: A Study of Organizational Change. New York: Aldine de Gruyter. Gittell,Marilyn.1980.Limitsto Citizen Participation: TheDecline of Community Organizations. Beverly Hills, CA: Sage. Greene, Sarah.M. 2003. Letterto Windy Hill, May 28, 2003. National Head Start Association, Alexandria, VA. Retrieved June 12, 2003 (www.saveheadstart.org/052003release.html). Kristen 1993. Understanding A. Nonprofit Gr0nbjerg, Funding: Managing Revenuesin Social Services and Community San DevelopmentOrganizations. Francisco,CA: Jossey Bass. 2002. "EvaluatingNonprofit Databases." AmericanBehavioral Scientist45:1741-77. Hallett,Joe. 2002. "OhioKids GainNew Advocate." The ColumbusDispatch, April 21, p. 1C.

316

AMERICAN SOCIOLOGICAL REVIEW JournalofPolicyAnalysis NonprofitContractors." and Management21:359-79. Reid, Elizabeth J. 1999. "NonprofitAdvocacy and PoliticalParticipation." 291-327 in Nonprofits Pp. and Government: Collaboration Conflict,editand ed by E.T. Boris and C.E. Steuerle.Washington, DC: UrbanInstitutePress. Ryan, William P., FrederickT. Miller, and Dorothy Weiss. 2002. "Government Fundingof Nonprofit Human-Service Organizations: A Review of Challenges and Opportunities." New York: Rockefeller Brothers Fund. Unpublishedmanuscript. Saidel, JudithR. 1991. "ResourceInterdependence: The Relationship between State Agencies and PublicAdministration Nonprofit Organizations." Review 51:543-53. Salamon,LesterM. 1995.Partnersin PublicService: Government(Nonprofit Relations in the Modern WelfareState. Baltimore, MD: Johns Hopkins University Press. 2002. "The Resilient Sector:The State of Nonprofit America." Pp. 3-61 in The State of Nonprofit America, edited by L.M. Salamon. Washington,DC: Brookings Institution. withoutMembers: Skocpol,Theda.1999. "Advocates The Recent Transformationof American Civic Life." Pp. 461-509 in Civic Engagement in AmericanDemocracy,editedby T. SkocpolandM. Fiorina.Washington,DC: Brookings Institution; New York:Russell Sage Foundation. Smith, Steven Rathgeb. 1989. "Federal Funding, Nonprofit Agencies, and Victim Services." Pp. 215-27 in Servicesfor Sale: Purchasing Health and HumanServices, edited by H.W.Demone Jr. and M. Gibelman.New Brunswick,NJ: Rutgers University Press. . 1999. "Government Financing of Nonprofit Activity."Pp. 177-211 in Nonprofits and Government: Collaboration Conflict,editand ed by E.T. Boris and C.E. Steuerle.Washington, DC: UrbanInstitute. Smith, Steven Rathgeb and M. Lipsky. 1993. State in theAge for Nonprofits Hire: The Welfare of Contracting. Cambridge MA: Harvard University Press. Stone, Melissa Middleton. 1996. "Competing Contexts: The Evolution of a Nonprofit Organization'sGovernance System in Multiple Environments." Administration and Society 28:61-89. Docile?" Wallis,Jim.2001. "WillAid MakeChurches Times,February3, p. A13. Op-Ed.New York Winship, Christopher and Larry Radbill. 1994. "Sampling Weights and Regression Analysis." Sociological Methodsand Research23:230-57. Wolch, Jennifer R. 1990. The Shadow State: Governmentand Voluntary Sector in Transition. New York:FoundationCenter.

C. Harris,Fredrick 2001. "BlackChurchesandCivic Traditions: Outreach, Activism, andthe Politics of Public Funding of Faith-Based Ministries."Pp. 140-54 in Can CharitableChoiceWork? Covering Religion s Impact on Urban Affairs and Social Services, edited by A. Walsh. Hartford, CT: Leonard E. Greenberg Center for the Study of Religion in Public Life. Haveman,RobertH., ed. 1977. A Decade ofFederal AntipovertyPrograms: Achievements, Failures, and Lessons. New York:Academic Press. Helfgot, Joseph H. 1974. "Professional Reform and of Organizations the SymbolicRepresentation the Poor." American Sociological Review 39:475-91. Hill, WindyM. 2003. Letterto Head StartProviders, May 8, 2003. Head StartBureau,Administration for Children and Families, United States Department of Health and Human Services, Washington, DC. Retrieved June 12, 2003 (www.saveheadstart.org/052003release.html). Hudson, Alan. 2002. "Advocacy by UK-Based Development NGOs." Nonprofit and Voluntary Sector Quarterly31:402-18. Jenkins, J. Craig. 1987. "Nonprofit Organizations and Policy Advocacy." Pp. 296-318 in The NonprofitSector:A ResearchHandbook,editedby W.W.Powell. New Haven, CT: Yale University Press. Kramer,Ralph M. 1981. Voluntary Agencies in the Welfare State. Berkeley, CA: University of CaliforniaPress. 1987. "Voluntary Agencies and the Personal Social Services." Pp. 240-57 in The NonprofitSector:A ResearchHandbook,editedby WW. Powell. New Haven, CT: Yale University Press. Kramer, Ralph M. and Bart Grossman. 1987. "Contracting for Social Services: Process and Social Management ResourceDependencies." Service Review 61:32-55. Matthews,Nancy A. 1994. ConfrontingRape: The Movementand the State. New FeministAnti-Rape York:Routledge. Sacredand Secular Monsma,StephenV 1996. When Mix:ReligiousNonprofit and Organizations Public Money. Lanham,MD: Rowmanand Littlefield. Mueller,Dennis C. 1979. Public Choice. New York: CambridgeUniversityPress. Netting, F Ellen. 1982. "Secular and Religious Funding of Church-Related Agencies." Social Service Review 56:586-604. Nowland-Foreman, Garth. 1998. "Purchase-ofServiceContracting, and Voluntary Organizations, Civil Society." American Behavioral Scientist 42:108-23. O'Regan, Katherineand SharonOster.2002. "Does Government Funding Alter Nonprofit Governance? Evidence from New York City

Sign up to vote on this title
UsefulNot useful