Should We Have Faith in Faith?

Section 104 of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (“PRWORA”) was the first of a series of legislative measures that are often collectively referred to as “Charitable Choice” laws—legislative efforts to encourage greater numbers of religious, or “faith-based,” organizations to bid on government contracts to provide social services to the needy. When George W. Bush was elected President in 2000, he made a Faith-Based Initiative expanding upon those legislative measures a centerpiece of his administration’s domestic policy. Both the Bush Faith-Based Initiative and the various pieces of legislation that preceded it generated substantial political opposition, with most critics focusing upon the First Amendment issues involved.

The policy issues involved, however, go well beyond arguments about constitutionality. Charitable Choice and the Faith-Based Initiative are efforts to encourage larger numbers of explicitly religious organizations, including congregations, to compete against other religious and secular service providers for government contracts to provide social services to those in need. That goal implicates several important, and contentious, policy areas: the definition and design of social welfare programs; the practice of providing government services through third-party surrogates or intermediaries rather than through employees, often referred to (incorrectly) as “privatization,” and the management challenges attendant to such delivery mechanisms; methodological difficulties and resource constraints that complicate (or preclude) the evaluation of government program effectiveness generally, and social service delivery mechanisms particularly; and—centrally—stormy political, religious and ideological debates over the causes of, and appropriate remedies for, poverty in America.

It has been ten years since the passage of PRWORA and Section 104. What, if any, conclusions can we draw based upon research and experience since then?

The public focus on the political and constitutional concerns involved has tended to divert attention and analysis from other, equally important, issues; it has also obscured the existing reality of social service provision in the United States, where government agencies have paid religious organizations to house, clothe and counsel the poor since the earliest days of state involvement in social welfare programs. Religious organizations providing government-financed social services range from 501(c) 3 affiliates of denominational entities, like Catholic Charities and Lutheran Social Services, to “pervasively sectarian” organizations like the Salvation Army, to individual congregations. Indeed, many proponents of “Charitable Choice” legislation and President Bush’s “Faith-Based Initiative” describe those policies simply as an attempt to level a playing field that already includes significant numbers of religious players. They argue that fear of overzealous application of the First Amendment has kept many religious providers from bidding on government contracts, and describe the legislation as an effort to ensure that government officials do not inappropriately require participating “faith-based” contractors to diminish or eliminate religious components of their services. Whether that assertion is accurate or disingenuous, the fact that public tax dollars have been used for decades to purchase social services from sectarian providers is beyond debate.

Charitable choice implicates two important arguments over the proper role of the state: the obligation of government to provide social welfare and the proper place of religion in the “public square.” Welfare policy disputes have an even longer pedigree than does the modern idea of the limited state; the origins of today’s policy debates over social welfare can be traced at least as far back as 1349. In that year, England enacted the Statute of Laborers, prohibiting citizens from giving alms, or charity, to those who had the ability to work—that is, to “sturdy beggars” (Handler and Hasenfeld 1997). This first attempt to deal with what we would later call welfare was emphatically not about providing assistance; it was about forcing people to work. Not until 1601, in the reign of Elizabeth I, would a tax be levied to provide material assistance to the poor. The Elizabethan Poor Law established three categories of needy people: children without parents (or at least without parents who could care for them adequately); the able-bodied; and the incapacitated, helpless or “worthy” poor. Vagrants, or able-bodied persons who refused to work could be “committed to a house of correction, whipped, branded put in pillories and stoned, or even put to death”(Kennedy, 2005). Help—however meager—was limited to the deserving, or “worthy” poor. That distinction between the “deserving” and “undeserving” poor would eventually be carried to the British colonies and reproduced in the laws of virtually all American states. It was the model that settlers brought to the New World; it was the approach adopted by the original thirteen colonies, and as people moved west, it was the approach incorporated in the Ordinance of 1787, which prescribed rules for governing the Northwest Territory. To a significant extent, the distinction between the deserving and undeserving poor, the “categorizing” of people needing assistance, and the emphasis upon work have remained the primary framework through which the general public and federal and state policymakers continue to view social welfare and poverty.

It is a paradigm that found considerable support in religion. The belief that poverty is evidence of divine disapproval—that virtue is rewarded by material success—was held in one form or another by a number of the early Protestants who settled the colonies; it is a theological perspective that has continued to influence American law and culture. It was not until the Great Depression that American lawmakers would widely acknowledge the need for some sort of social safety net. Even then, it would be a mistake to assume that the dislocations of the 1930’s or the passage of New Deal legislation changed Americans’ deeply-rooted beliefs about poverty, welfare, or their own history of self-reliance. As social historian Stephanie Coontz has written, “Most Americans agree that prior to federal ‘interference’ in the 1930’s, the self-reliant family was the standard social unit of our society. Dependencies used to be cared for within the ‘natural family economy’ and even today the healthiest families ‘stand on their own two feet.’” (Coontz 1992, 69)

Coontz and others have demonstrated that this widely-held belief in self-sufficiency is inconsistent with the reality of the American experience. From the massive federal land grants that helped early Pioneers, to early twentieth century government construction of dams and federal subsidies to irrigation projects, to Depression-era government electrification projects, government investment has been critical to the survival of the family farm. In the 1950’s, the GI Bill and the National Defense Education Act that subsidized the educations of a whole generation, and the Federal Housing Authority and the Veterans’ Administration allowed Americans to purchase homes with artificially low down-payments and subsidized interest rates. Billions of dollars of government-financed inventions, production processes, and research have enabled businesses to flourish, and by the 1970s, Social Security had virtually wiped out the historic tendency for the elderly to be the poorest sector of the population. These programs have enjoyed wide popularity and support; welfare for the poor, however, continues to carry a considerable social stigma. It is simply not possible to understand Charitable Choice laws or the President’s Faith-Based Initiative without understanding the historic context within which welfare policy debates occur, and the sources of the contending perspectives involved.

Charitable Choice legislation rested—as all laws do—on assumptions both articulated and tacit. They included the assumed existence of the “armies of compassion” so frequently alluded to by the President—voluntary efforts to alleviate poverty that would assist with that effort once unnecessary roadblocks had been removed, and the presumed governmental discrimination against “faith” had been remedied. And they quite explicitly included a belief that “faith-based” organizations are more effective than secular ones in combating poverty (and the poor “values” to which such poverty is often attributed).

Ten years after the passage of PRWORA’s Section 104, do we know whether the “armies of compassion” have materialized? Do we know whether faith-based efforts to alleviate poverty are more or less effective than secular ones? Or, as opponents of these measures have charged, is the emphasis on faith-based contracting diverting necessary resources, eroding constitutional boundaries, and distracting policymakers from the real issues?

Can we even agree on what “success” and “effectiveness” and “help” (let alone “faith”) look like? Is the highly polarized discourse and conceptual incoherence that characterize this particular policy arena simply another example of the increasingly partisan nature of policymaking in the United States—a reminder of the truth of the old admonition not to look too closely at the process of making either sausages or laws—or is this debate a “special case,” and if so, why?

A review of the literature available in 2000 suggests the quite limited nature of what was known about faith-based social service provision when PRWORA passed. There was then—and is now—virtually no credible empirical research on the comparative performance of religiously-affiliated organizations and secular ones; or of the comparative efficacy of providers that accept government contracts and those who do not. For that matter, there had been relatively little evaluation of the efficacy of any social service programs. Government agencies rarely evaluated their contractors to determine the utility of the programs being funded, or to compare the performance of contractors. Existing government partnerships with religious organizations had generated very little constitutional litigation or commentary. What information existed tended to be descriptive: how many religiously-affiliated organizations were government contractors and what percentage of contractors’ budgets came from government (Netting 1982, 1984); the extent of congregational (as opposed to religiously affiliated nonprofit organizations) interest in providing social services under such contracts and the capacity of such congregations to deliver those services (Chavez 1999; Grettenberger 1997). To suggest that faith-based social services of the sort contemplated by Charitable Choice were “understudied” would be a considerable understatement.

While research directly on point was sparse, studies having varying degrees of relevance to the questions facing researchers in the wake of Charitable Choice did exist. Several articles by Mark Chaves were particularly helpful, given the new emphasis on outreach to congregations in addition to religiously-affiliated social service agencies (and the President’s later assertions that “Armies of Compassion” represented significant untapped resources that could be made available to help fight poverty). Chaves mined the National Congregational Study to provide a wealth of information about American congregations, including information relevant to the question of congregational capacity to provide social services. His data suggested that most congregations are severely limited in their capacity to operate formal programs. Only twelve percent of the congregations in his study operated social service programs under their own auspices—that is, using their own staff and financial resources. It was far more typical to find congregations supporting programs delivered by others. Chaves found congregations to be adept at mobilizing small groups of volunteers to conduct well-defined, periodic tasks. Churches and synagogues are particularly well-suited to providing short-term responses to immediate needs, and much less able to sustain long-term programs requiring consistent staffing and supervision (Chaves 1999). An earlier study by Susan Grettenberger, limited to United Methodist congregations in Michigan, had come to similar conclusions; Grettenberger found that congregations in her study provided services primarily in response to emergencies, and that they simply did not possess the necessary resources to implement programs requiring more sustained service delivery. (Grettenberger 1997)

In a later article, Chaves once again used data from the National Congregational Study to assess the prospects for Charitable Choice and the Faith-Based Initiative. He noted that only three percent of congregations surveyed in 1998 received government funding, that larger congregations—those having more than 900 regular attendees—were more likely to operate social service programs, that ethnic composition was the single most important predictor of willingness to partner with government, and that congregations in the South were less willing than congregations elsewhere to apply for government funding. He also found that “secular integration” (defined as the extent to which a congregation operates within a relatively more secularized institutional environment), was a factor in its willingness to work with government agencies. Chaves noted that Catholic churches and liberal Protestant denominations expressed significantly more willingness to enter into government contracts than did more conservative, or evangelical, churches. (Chaves 1999)

In a 1998 Policy Paper written by Stephen Monsma and issued by the Center for Public Justice, Monsma referenced the long history and current, widespread practice of government co-operation with religious social service providers—a history and practice which he characterized as “largely positive.” Monsma based his findings on a national survey of religious agencies providing services to children and families, which he had conducted in 1993-94. Ninety percent of his respondents reported receiving at least some public funding. Of those who characterized themselves as “clearly religious” 82% got government money, and a strong majority of those reported no problems with government officials over religious practices Most who had experienced such problems had been requiring religious participation as a condition of receiving the service involved—a clear violation of the First Amendment. (Monsma, 1998:5-6)[1] . (Monsma’s results have been recently confirmed by other scholars (Kramer, De Vita & Finegold 2005)).

In 1999, McCarthy and Castelli published a summary of the literature on religious social services then available. “Religion-Sponsored Social Service Providers: The Not-So-Independent Sector” was issued as part of their working paper series by the Aspen Institute’s Nonprofit Sector Research Fund. The introduction noted that “political leaders are…making dramatic proposals and even enacting new laws without the benefit of an understanding of the extent and nature of existing religion-sponsored social service providers” (McCarthy and Castelli, 1999: 3). Their review of available research reinforced Chaves’ finding that religious social services primarily take the form of emergency interventions, and that congregations most likely to provide services are those that are ideologically liberal and large, high-income, suburban or black. They found that religious social services were almost exclusively local, and were typically provided without reference to the religion of the beneficiary. They also found that government agencies and religious providers worked together in a number of non-contractual contexts (for example, by providing referrals, consultations, and facilities). With respect to the issue of comparative effectiveness of faith-based and secular providers, they flatly stated “[W]e have been unable to locate a single credible study assessing the relative effectiveness of religion-sponsored social services.” (5) In language anticipating the belief that “Armies of Compassion” were waiting to help government provide for the needy, they were similarly blunt. “We do not believe, given the evidence we have examined, that it is possible for America’s religious institutions to dramatically expand their provision of social services.” (6)

Even those who did hope for increased congregational participation in social service delivery as a result of faith-based outreach efforts recognized that religious congregations would not and could not replace the nonprofit and religious organizations that have traditionally provided the bulk of American social services. In 1996, in the wake of PRWORA, several scholars published articles addressing that issue and looking specifically at the capacities of nonprofit and religious providers to succeed in the expanded role envisioned for them as a result of the various Charitable Choice measures being enacted.

In 1999, John Bartkowski and Helen Regis published a “Feasibility Study of Faith-Based Welfare Reform in Mississippi,” which surveyed religious involvement in social services in that state. The study analyzed the strengths and weaknesses of the organizations involved; they found that religious providers tended to be more “holistic,” but also noted limits on such organizations’ capacity imposed by inadequate resources, as well as a “potential for exclusivity and insularity.” (Bartkowski and Regis 1999) Also in 1999, Jennifer Alexander published an analysis of the effects of devolution on nonprofit organizations in the Cuyahoga County, Ohio area. She concluded that the capacities of community and faith-based organizations to fulfill the expectations raised by Charitable Choice were severely limited by a lack of human and financial resources. She also noted that many of those working at nonprofit agencies believed they had already begun to absorb added costs as a result of welfare reform and “load shedding” by government. (Alexander 1999)

Similar results were reported in a study of Washington, D.C. faith-based service providers; in that study, a majority of respondents said they lacked adequate resources, facilities, staff and funds to meet increases in need. (The survey also found that large numbers of nonprofit service providers could not account for the resources they had, or for their income and expenditures, because they did not keep records.) (Printz 1998)

Many of the articles and pamphlets published in the wake of PRWORA saw Charitable Choice as an opportunity. One of those was aptly titled “The Charitable Choice Opportunity.” Issued by the Center for Public Justice, it described the old rules (i.e., before PRWORA) as “restrictive” and the new ones as “inclusive,” and cheered passage of the law for “clearing away barriers to full involvement by faith-based organizations.” (Carlson-Theis, 2000:2-3). Others—including many who were supportive of participation by faith-based providers in welfare programs—raised red flags.

In 1997, the Nonprofit Sector research fund published “Government Promotion of Faith-Based Solutions to Social Problems: Partisan or Prophetic?” by Thomas J. Harvey, President Emeritus of Catholic Charities USA. Harvey’s experience, he wrote, led him to “believe that religiously-affiliated agencies, other voluntary agencies, and the public at large” should assess Charitable Choice efforts carefully, in the context of government’s unique role in promoting the general welfare (Harvey, 1997: 5-6). Among Harvey’s concerns was the “values” rhetoric frequently employed by proponents of the measure—rhetoric that tended to characterize poverty as a spiritual deficit rather than a material one. In that formulation, people in need lacked jobs not because the local factory had closed, or because educational requirements had changed, or because day care was unavailable, but because they possessed an inadequate work ethic or otherwise lacked the “middle-class values” demanded by the job market.

“Interpreting the problems of the poor as spiritual also subtly perpetuates a centuries-old interpretation that the poor are lazy and without worth…This prejudice implies that government assistance to help them will only perpetuate the problem. This bias represents a form of social Darwinism—survival of the fittest—which has been well documented and analyzed by historians and social commentators as being deep-seated, but without foundation.” (Harvey, 1997:9)

Harvey also advised the religious community to resist the temptation to see public programs as evidence of altruism, rather than as evidence of “a political process [that] often reflects vested interests more than a commitment to the common good” (p.21).

Several articles addressed the issue of “load shedding,” warning that nonprofit agencies, whether faith-based or secular, lack the resources to substitute for the “eroding public safety net.” (Edin and Lein, 1998) Other writers focused upon the potential effects of government contracting on religious and community-based organizations. Alexander, Nank and Stivers investigated the impact of welfare reform generally on nonprofit agencies in Cuyahoga County, Ohio, and concluded that smaller nonprofits simply lacked the capacity to adopt the organizational and business practices required by government agencies (Alexander, Nank and Stivers, 1999.) Their research confirmed several earlier articles, including a 1994 study by Kearns, who had found “public sector conceptualizations” (i.e., business practices) “ill-suited” to nonprofit agencies. (Kearns, 1994: 185) In “Black Churches’ Involvement in ‘Charitable Choice’ Programs: The Promise and Peril,” Ryden endorsed the effort to direct added resources to urban African-American churches, but warned that such churches would have to learn to walk “a fine line” to avoid a number of legal pitfalls. (Ryden, 2000:2) Chambre studied four faith-based AIDS organizations in New York, and found that the religious identity of the providers had become attenuated in the course of the contract relationship; two of the organizations she studied secularized, while the other two adopted more “ecumenical” forms of faith. While Chambre did not attribute these changes solely to the organizations’ receipt of government funds, her research suggested that such funding was a significant element driving the documented shift (Chambre, 2001).

Arthur E. Farnsley was an early, acute observer of Charitable Choice efforts. In “Ten Good Questions About Faith-Based Partnerships and Welfare Reform” he identified important and at that point unanswered questions, among them “how will this work?” “do congregations have the necessary administrative capacity to use public funds?” “will congregations change individuals’ values, and are we comfortable with that?” and “who will benefit, and who will lose?” (Farnsley, 2000) “Ten Questions” was followed by an article in Christian Century, in which he pointed out that “many interested parties have different goals” for Charitable Choice, and that it is difficult if not impossible to decide whether a program is successful when there is no agreement on what that program is intended to do, or what “success” might look like. Farnsley identified four interest groups with highly disparate agendas that had supported Charitable Choice legislation: (1) advocates of privatization, who believed outsourcing would save money, and whose motives were largely economic; (2) social service and foundation professionals who were frustrated with social service failures and the seeming intractability of poverty, who saw Charitable Choice as a “bold social experiment” that might work; (3) “reformers” for whom separation of church and state has gone too far, who wanted government to be more ‘accommodating’ to religion and religious expression; and (4) “Salvationists” who saw an opportunity to press a Christian mission agenda with money from the government. (Farnsley, 2001)

These wide differences in perspective explain the equally different approaches to Charitable Choice in the scholarly literature available. Romzek and Johnston offered an explanatory model of accountability mechanisms and focused on problems likely to be faced by public managers operating in the “hollow state” described by Milward (1994) and others (Romzek and Johnson, 2001). Kennedy and Bielefield addressed the issues the law’s ambiguities raised for public managers (Kennedy and Bielefeld, 2001). Nank and Stivers also explored public administration issues arising out of the “hollow state” (Nank and Stivers, 2001). Monsma continued to produce careful, workmanlike descriptions and case studies of partnerships between religious organizations and government, and to catalogue their approaches to service delivery (Monsma, 2002), while others brought social work and social service network perspectives to bear (Cnaan and Boddie, 2002; Wineburg, 2001). Still other research efforts focused upon particular faith communities, or on specific social services (Choi and Tirrito 1999; Thomas, Quinn, Billingsley and Caldwell 1994). The picture that emerged from these and later studies, although far from complete, was of a wide variety of nonprofit and religious social service organizations engaged in a daily struggle to stretch limited resources and meet existing demands while assessing the pros and cons of closer relationships with government.

As evidenced by this brief and necessarily incomplete summary, much scholarship was relevant to aspects of the research conducted in the wake of Charitable Choice by this writer and others, even if very little of it was directly on point. It was when we looked for empirical research comparing the performance of secular and faith-based nonprofits, however, that we came up dry. Once we excluded anecdotal reports and public relations efforts by agencies promoting their own programs, we found—as McCarthy and Castelli had predicted—only four articles of even arguable relevance. All of them involved substance abuse, and two of the four did not actually involve comparison of religious and secular treatments, but did explore the role of religion in substance abuse programs.

One of the two investigations of religious elements in substance abuse treatment was a 1978 evaluation of a Salvation Army alcohol treatment program. The researchers concluded that the “social climate” and “treatment approach” of the Salvation Army contributed to the success of treatment for those participants who actively participated, and further concluded that the Salvation Army’s program was as successful as (but not more successful than) other, similar programs. (Moos, Mehren and Moos 1978) The other was a study entitled “Individual Religiosity, Moral Community and Drug User Treatment,” reporting the results of research to examine the relationship between reductions in alcohol and drug use and religiosity. Despite the authors’ original hypotheses that such connections existed, and despite some evidence suggesting that participation in a “religious moral community” could positively affect use, the researchers could establish no “obvious relationship” between dependent behavior and individual religiosity. (Richard, Bell and Carlson 2000)

Two studies compared secular and religious programmatic effectiveness. A 1981 article by David Desmond and James Maddux reported on research involving programs for heroin addicts. Eighty-seven percent of the addicts who participated in the study were Hispanic. Over a twelve-year period, 11% of them had entered religious treatment programs, and 10% had participated in a methadone program that also had a religious component. The rest had gone to providers offering traditional treatment modalities. Those addicts who willingly joined the religiously intensive programs were more successful than those who participated in conventional programs; however, the religious methadone program was less successful than the secular programs. (The authors theorized that many of those who signed up for the methadone program did so only to obtain the substitute drug, not to reduce their drug dependence.) The researchers concluded that the variable most likely to predict success was personal motivation—that those who entered programs with positive attitudes and openness to the treatment approach were more likely to succeed. The evidence did not support a determination that effectiveness was attributable to the religious character of the programs. Such a conclusion would be particularly problematic, according to the authors, since Hispanic clients were likely to self-select religious treatment programs. (Desmond and Maddux 1981)

The other comparative study was a 1993 report on church-based smoking cessation programs. The authors compared the efficacy of “intensive, culturally relevant and spiritually-based church interventions” to the effectiveness of a “self-help, cold turkey” approach. The church interventions used sermons, testimony, volunteer counselors and follow-up evaluations; these mechanisms—not surprisingly—were found to be more effective than the “tough it out” approach (Stillman, Bone, Rand, Levine and Becker 1993).

Studying Charitable Choice: A Challenge

If the lack of consensus on who the intended targets of outreach might be, and on what problem the policy was intended to solve, makes productive deliberation difficult, it also poses significant challenges for research methodology and process. Charitable Choice laws raised a number of preliminary questions: How are the faith organizations targeted by these measures different from Catholic Charities, Lutheran Social Services, the Salvation Army, and government’s many other long-time religious partners? What are “faith-based” organizations, and how do they differ from other non-profit organizations? The legislation did not address these issues, and the White House Office on Faith-Based and Community Initiatives expressly declined to do so (Hiatt, 2001). That was not the only definitional dilemma; it was by no means clear what was meant by “programmatic success” and “efficacy.”

Furthermore, the effort to recruit new faith partners was not accompanied by additional funding for social services. With no new money, some scholars charged that Charitable Choice and the Faith-Based Initiative would simply end up shifting funds from one set of religious providers to another—presumably, from government’s traditional religious partners (who generally operate in accordance with applicable professional norms) to providers more focused upon the "personal transformation" of clients? There were a number of other issues implicated by funding realities: What happens to small, grass-roots faith-based organizations when diminished public resources make funding streams unreliable? Will government funding affect the character or mission of small religious organizations new to the contracting regime? If so, how?

Finally, there was the question that had occasioned the most political controversy: would constitutional norms be followed? The First Amendment does not prevent government from doing business with faith organizations, but that doesn’t mean that any program run by a religious provider will pass constitutional muster. There is a constitutionally significant distinction between programs that are offered by a religious provider or in a religious setting, and programs in which religious observance or dogma are integral to service delivery. What mechanisms are proposed to ensure that services are delivered in a constitutionally appropriate manner? How will constitutional accountability be ensured? And what are we to make of the assertion that religious providers are more effective than secular ones? Many faith-based and religious organizations do important, often exemplary, work—but when Charitable Choice was passed, there was no evidence either supporting or rebutting the claim that religious organizations as a category are more effective than secular ones. And finally, how will the multitude of management issues be resolved? The law left numerous ambiguities and unanswered questions. The question for public managers has never been whether government should partner with religious organizations to provide social services. It always has, and undoubtedly always will. The question is "under what circumstances are such partnerships appropriate and when they are, how should they be structured and monitored?" Similarly, the question is not whether, in the abstract, religious programs or secular approaches are preferable. The question for government program managers is "what organizational characteristics are most likely to predict successful program delivery, and how can I determine which of the bidders for this contract possesses those characteristics?" “How will I define and measure accountability?”

One Set of Questions

This article is an effort to reflect on faith-based contracting in the aftermath of a research project that focused upon five specific questions:

· In a federalist system, how will state implementation approaches vary? One of the frequently-cited virtues of a federalist system is the freedom such systems afford individual states to act as “laboratories of democracy”—to take different, hopefully innovative, approaches to the national agenda from which the country as a whole can learn and benefit. It was inevitable that states would approach Charitable Choice and the Faith-Based Initiative differently; we wanted to examine those differences and see what policy conclusions, if any, we might draw. Since our resources were limited, we chose three states to serve as a primary focus for that part of our inquiry, although we did monitor publications, news items and lawsuits from all of the states.

· Is there a measurable difference in performance efficacy between faith-based and secular services? A significant part of the political justification for faith-based initiatives was the often-articulated belief that religious providers are more effective, that religious providers do a better job than their secular counterparts. We looked at faith-based services provided to TANF (Temporary Assistance to Needy Families) recipients—specifically, job training and placement programs—to determine whether FBOs were, in fact, more successful in meeting observable, objective benchmarks: were their clients more likely to obtain and hold employment? Did they earn more? And if so, if there were measurable differences in performance between faith-based and secular organizations providing government services, was faith the variable that accounted for those differences?

· What organizational characteristics and capacities correlate with effectiveness? The success of any government program depends significantly upon the organizational and management capacities of those in charge; in the case of Charitable Choice, successful implementation depends upon the fiscal, programmatic and constitutional capacities of both government agencies and the faith-based service providers with whom they contract. Analysis of those capacities requires answering several questions: How do the states define and locate FBOs and what criteria do they use to assess their ability to deliver the necessary services? Are small, grass-roots FBOs able to absorb the transaction costs associated with government contracting? (Gronbjerg 1997). What about their access to the accounting and legal resources needed to assure compliance? If FBOs do not have these resources, are states assisting with “capacity-building” measures? What about the states’ own capacities to manage contracts and monitor for fiscal, programmatic and constitutional compliance? These narrow questions go to the broader issues of public management, definition of terms and funding.

· What are the effects of contracting on faith-based organizations? For-profit and nonprofit organizations that work with government agencies exhibit a variety of adaptations as a result of those partnerships (Smith and Lipsky 1993; Kettl 1993; Frumkin, 2000) A significant concern voiced by religious critics of Charitable Choice and the President’s Faith-Based Initiative has been that small faith-based organizations accepting government funding would thereby lose independence—that they would eventually be co-opted, and the distinctive and important “prophetic voice” of the church would be muted. If, as some proponents of Charitable Choice assert, the larger religiously-affiliated providers that have worked with government for decades have become “too secularized” as a result of those partnerships, what are government agencies doing to ensure that smaller FBOs will not experience the same organizational and behavioral metamorphosis? (See, for example, Alexander, Rank and Stivers, 1999; Chambre, 2001).

· How accountable are public managers and contractors? Outsourcing generally, and Charitable Choice specifically, implicate important issues of constitutional, programmatic and fiscal accountability. What are the capacities of grass-roots religious organizations and their government partners to provide the required services, account for government funds, evaluate program outcomes, and adhere to Constitutionally- mandated boundaries? Accountability requires clarity of definition, implicates funding and management issues, and is at the very core of concerns over the constitutionality of Charitable Choice programs.

We focused our research on one type of social service, job training and placement. Because of the difficulty in data collection (a difficulty that helps to explain the paucity of empirical research), the portion of our research comparing the effectiveness of faith-based and secular providers was limited to providers in just one state. How representative our findings were, how generalizable, is thus a fair question.

Our findings must also be understood in the context of the overall response by religious providers to the invitation of Charitable Choice and the Faith-Based Initiative. To be blunt, there has been no dramatic (or even significant) change in the social welfare landscape. The hypothesized “armies of compassion” have thus far not materialized; in states where such information is available, it would appear that the composition of social service contractor pools has changed very little. What we cannot say with any degree of confidence is why. It seems reasonable to assume that fiscal realities have played a role: the states have been in fiscal crisis almost from the time the Administration began encouraging them to find faith-based partners, and the lack of added federal funds for social services has prevented any expansion of social service programs. To the contrary, in many cases, scarce resources have resulted in the termination of programs not required by state or federal law, and fewer dollars for the programs that have survived. It has not been an auspicious time for new entries into the already tenuous mix of government, for-profit and non-profit organizations that make up America’s tattered “safety net.” Another possible disincentive is theological and/or cultural: many religious nonprofits and congregations, for a variety of reasons, look with suspicion on partnerships with government. Other small FBOs do not want the legal restraints and reporting burdens that accompany contractor status, while still others have been lured into a contract relationship and after a brief time, determined that it was not in their interest to continue. It is also possible that those religious contractors who really had an interest in working with government on social welfare issues were already doing so. Whatever the reason, the response from religious organizations to Charitable Choice and the Faith-Based Initiative has been decidedly tepid, despite a variety of efforts to lure such organizations into the contracting pool.

The specific research questions we addressed, and the methodologies employed, have been the subject of several publications; and for purposes of this reflection on the policy implications overall, are simply summarized below.

The Effects of Federalism

Perhaps the simplest task was the investigation of differing state approaches to implementation. This was a descriptive endeavor: we chose three states, based upon their different political cultures and religious landscapes. Massachusetts is a northeastern state with a liberal political culture. Religiously, it is heavily Catholic. Massachusetts government is emphatically hierarchical—the state does not even have counties, and most social service programming and contracting are conducting at the state level. North Carolina is a southern state, with a political culture that remains heavily influenced by its origins in southern agrarian life. Unlike Massachusetts, government is decentralized, and social services have been “devolved” to nearly 100 county governments. County governments contract for social services, and state officials often do not have information on the identity of contracting parties. North Carolina is heavily populated with Baptists. Indiana is a Midwestern state, with what Daniel Elazar has called an “Individualistic” political culture (Elazar, 1994), characterized by a pragmatic, “quid quo pro” attitude. Indiana has no single dominant religious tradition. There are 92 counties, and those counties administer social services, but the state government is the contracting party and (theoretically) maintains copies of all contracts and monitors compliance. (Such maintenance and monitoring is not uniform in practice, but it is the intended system.)

As hypothesized, these different states did approach Charitable Choice implementation differently. Despite the centralized nature of the welfare system, and the federal reach of the regulations promulgated under PRWORA, state approaches reflected differences in local political cultures and governmental structures. Massachusetts read the statute to require transparency and equal treatment in the contracting process; the state had completed a massive overhaul of its contracting procedures barely two years before PRWORA, and took the position that its process was already transparent and non-discriminatory. State officials pointed to the large numbers of religiously-affiliated providers already doing business with the state, and concluded that Massachusetts was in compliance, and that no further effort needed to be made. (Jensen, 2003)

North Carolina, lacking a central contracting structure, opted for a string of “demonstration projects” funded at the state level. The impetus for several of these came from the religious community, rather than from state government. While comprehensive data is not readily available from North Carolina, it is safe to assume that a significant percentage of county-level contracts had always been with religious organizations, and that Charitable Choice further encouraged that practice. (Queen, 2003)

Indiana hired a consultant, Crowe Chizek, and paid it $250,000 each year for three years to conduct statewide workshops encouraging faith organizations to bid on contracts and to engage in capacity building activities, including grant-writing, contract compliance and similar skills. The state aggressively moved to implement outreach to congregations and other religiously-affiliated providers, and state officials assisted new contractors with the intricacies of billing and reporting (Thelin, 2003).

A survey of literature from studies of other states displayed a similar array of responses. Interestingly, despite the wide variation in the intensity and scope of state efforts at “implementation,” no state of which we are aware experienced a significant increase in the percentage of religiously-affiliated contractors providing social services.

Comparative Efficacy

We found no evidence to support the contention that faith-based contractors are more effective than secular providers. As indicated above, there is relatively little research in the area of provision of social services by faith-based organizations. Since our original literature review in 2000, Chaves and Tsitsos (2001) have examined what social services religious organizations provide and how they do it. Kramer, et al. (2002), asks similar questions but in the context of provision of employment-related services. Monsma and Mounts (2002) have examined how faith-based welfare-to-work programs differ from their secular counterparts in terms of funding from government and services offered. However, we have found little published literature that examines differences in outcomes for clients who receive social services from faith-based versus secular providers.

Providers of job training services were categorized as faith-based or secular based upon the results of surveys measuring eight dimensions related to the influence of faith in the organization (Bielefeld, et al., 2002). Organizations scoring positively on one or more dimensions were classified as faith-based in this study and their value on an eight point scale was used to designate the degree to which faith influenced the organization.

The period during which we collected our sample coincided with the initiative by the state to encourage religious social service providers to bid on contracts to provide job training and other social services to welfare clients. As a result of the state’s efforts, a number of new providers had been added to the existing set of job training providers available to welfare clients. We identified five faith-based providers who were either less than five years old and/or had contracted with the state for two years or less. Because these providers are relatively young and new to the welfare landscape, and because it was possible that differences between their capacities, capabilities, and resources and those of existing religious providers could contaminate the comparison, we conducted our analysis both with and without them.

In the sample including all faith-based providers, 40 percent of clients who engaged in job training were placed into jobs after training. Of those who were placed, 44.5 percent worked full-time when full-time is defined as at least 35 hours per week (that fell to 35.4 percent if full-time status is defined as working at least 40 hours per week). They earned an average of $6.88 per hour. Almost six percent of these individuals were covered by employer health insurance plans. These characteristics do not change substantially when the new faith-based providers are removed from the sample. The individuals in our samples were disproportionately female (77 percent) and non-White (20 percent White). Slightly over half of the clients had high school diplomas, and seven percent had at least some additional education.

Among those who were placed, clients who received their training at more strongly faith-based providers were significantly less likely to work full-time. Regardless of whether full-time work is defined as working 35 or 40 hours, clients of faith-based providers were, on average, less likely to be employed full-time. For full-time defined as 35 hours, clients of faith-based providers were 7 percentage points less likely to be employed full time for each 1 point increase in their provider’s faith-influence scale value. For full-time defined as 40 hours, the percentage point decrease in the likelihood of full time work was 6.1 for each 1 point increase in provider faith orientation. Wages were not significantly different for clients who received training from faith-based providers compared to those who received training from secular providers. Finally, we found that clients of faith-based providers were significantly less likely to receive health insurance compared to clients of secular providers. The decreased likelihood in this case was 2.1 percentage points for each 1 point increase in provider faith orientation.

The effects of provider type on labor market outcomes for the sample in which new faith-based providers are excluded were qualitatively identical to those in which they are included. Clients of more strongly faith-based job training providers, as compared to those of secular providers, were equally likely to be placed into jobs and, among those with jobs, likely to earn similar wages. But they were less likely to work full-time and have health insurance.

Other research on comparative effectiveness of religious and secular providers is underway, and will either support our research conclusions or come to different ones, but such research as has emerged thus far suggests our results are not an anomaly. (One recent investigation found no difference in the “ethical considerations” motivating religious and secular providers, workers & volunteers “Even secular programs have sacred roots” (Lockhart, 2001)). While different researchers define “effectiveness” differently, when we limit the definition to the programmatic goal of placing people in jobs that will allow them to achieve self-sufficiency, faith-based organizations do not do a better job, and on some dimensions, they do not do as well as their secular counterparts.

Effects on Faith-Based Organizations

Critics of Charitable Choice, and even some supporters, worried about potentially negative effects on faith-based providers as a result of doing business with agencies of government. Those concerns ranged from the practical (cash-flow problems, lack of resources to adequately report and account for funds and clients) to the systemic (co-option of the “prophetic voice” of religious organizations, mission-creep resulting from growing dependence upon government dollars). We found that the more strongly faith-based organizations were likely to include explicitly religious elements in programming, and as a result, to experience more negative reactions from clients. They were also more likely to provide services for which they were not being reimbursed. We found no positive correlation between faith and holistic service provision, although such a connection has been widely hypothesized. In addition, research by others supports the contention that contracting with government affects nonprofits’ missions. Studies suggest that “sustaining the distinctive normative climates that ensure uniqueness and selectivity is increasingly difficult” (Lynn, 2002) for nonprofits (not just faith-based nonprofits) that contract with the government. These and similar findings lend strength to the hypothesis that there may be measurable outcome differences between faith-based services funded by private contributions and those provided by a faith-based government contractor.

These organizational research results underline certain of the conceptual difficulties with Charitable Choice. There are significant organizational differences between congregations and other faith-based nonprofits, and those differences deserve more recognition than current outreach efforts accord them. Congregations are—first and foremost—communities of worship. They are not formed to be social service providers, and some of the demands made upon them by the latter status are inconsistent with the mission of the former. Pastors and other church leaders are rarely equipped to deal with the multiple and difficult issues of poverty (Farnsley 2003). Furthermore, research does not support the assumption that congregations are more rooted in and familiar with their neighborhoods and communities. (Interestingly, members of African-American churches are even less likely to live in the neighborhood where the church is located than are members of white churches (Farnsley, 2003: 52)).

Organizational Effectiveness

Our study did find a relationship between management challenges and the influence of faith, but the challenges were not limited to the effects of a faith orientation. As might be expected, we also found relationships between management challenges and organizational age and size. According to our research, FBOs experience both management strengths and weaknesses. More strongly faith-based providers indicated greater challenges primarily in the areas of governance and external relations. The former includes recruiting and keeping effective board members and managing board/staff relations. The latter includes strategic planning and the delivery of high quality services. Faith-based organizations also faced more challenges with their internal operations in terms of managing facilities. A number of these challenges were also related to age or size.

The governance challenges may be a reflection of a dual authority structure, in that religious leadership may have difficulties recruiting and working with nonreligious board members. Staff may also have similar difficulties. The relevant literature suggests that newly established or growing faith-based organzations may encounter issues with external actors and competing standards, and that this may account for the difficulties those in our study had experienced with strategic planning and facilities management. The strategic planning process may be more of a challenge for faith-based organizations because they may not be as familiar as other nonprofits with the process, especially the assessment and evaluation of complex external environments. Managing their facilities/space may pose more of a challenge because faith-based groups may face demands for their facilities from their congregation or other community programs.

On the other hand, more strongly faith-based providers reported fewer challenges with mission accomplishment and a number of internal processes including internal communication, internal working relationships, anticipating financial needs, obtaining funding, using technology for service provision, recruiting/keeping qualified administrators, and managing programs. None of these was related to size and only the last was related to age.

The finding that faith organizations have fewer internal problems may indicate that, while it may be problematic for governance, a dual authority structure does not seem to be creating internal stresses and strains. More specifically, achieving their mission may be less of a challenge for religious organizations as they usually have a well-articulated and focused mission related to their faith. Communicating internally and developing/sustaining good working relationships may be less of a challenge because their employees and volunteers are likely to be of the same faith and religious culture, a shared background which would facilitate communications. Similarly, obtaining funding and anticipating financial needs may be less of a challenge because they often have the support of a congregation or other religious institution.

Constitutional Accountability

Critics of Charitable Choice warned that faith-based contracting would lead to wholesale constitutional violations, a concern that our study suggests is not entirely misplaced. In addition to the three states we studied in depth, we conducted a survey of contracting practices nation-wide, and monitored court decisions and other legal actions involving Charitable Choice and the Faith-Based Initiative. We found that state-level agencies lack the resources—fiscal and human—to conduct even the most cursory monitoring for such violations. No state of which we are aware has effective oversight mechanisms in place, and most make no effort to suggest otherwise. The most common “enforcement mechanism” in states claiming to have one is a contract provision requiring the contractor to obey all relevant laws.

The pressure on state agencies to demonstrate that they are not discriminating against religious providers has led to accusations that they are erring in the other direction, and improperly favoring religious contractors. Allegations of proselytizing, of coercing clients into participation in religious rituals and observances, have begun to surface (although it is difficult to know whether these practices have become more common, or whether the visibility of the Charitable Choice debate and thus of religious contracting has made people more alert to them.) (Pierce, 2005; Fried, 2004; Neary, 2003; Sager, 2005) Few public managers or managers of small faith-based organizations display a sound understanding of the First Amendment constraints that accompany tax dollars, and a small study we conducted of congregational leaders in one mid-sized Indiana municipality disclosed a widespread misunderstanding of First Amendment application and requirements. (It would not be appropriate to claim statistical validity for that study; the response rate was not adequate to establish such validity. Nevertheless, the responses were suggestive—and disheartening. Over seventy percent of those who did respond were unaware of the most basic principle of state action—that is, that the Bill of Rights constrains only government action. Almost an equal percentage saw no problem using government dollars to proselytize.)

While it would not be accurate to characterize hiring practices based upon religious criteria as constitutional violations until and unless the Supreme Court rules on the Charitable Choice exemption from civil rights laws, such hiring practices raise numerous policy concerns. And in addition to the issues specific to the First Amendment, there are the broader constitutional concerns raised by third-party government in general. (Kennedy, 2001)

Policy Implications

What do these and other research findings tell us? One clear conclusion—a conclusion that is unavoidable for anyone familiar with the discourse and emerging literature on Charitable Choice and the Faith-Based Initiative—is that attempts to understand the competing points of view involved are hampered if not made entirely futile by the fact that those engaged in the discussion frequently inhabit different realities. Participants in the political discussions are in large measure talking past each other. The choice of the term “faith-based” is an example: the term was undoubtedly chosen in an effort to be inclusive of all religious providers, but it betrays a quite narrowly Protestant conception of religion. (Judaism and Catholicism, among other religious traditions, are “works based.”) Glib and fashionable political references to a “Judeo-Christian” America ignore profound religious differences. Indeed, several scholars have suggested that faith-based efforts are not intended to privilege religious over secular service providers, but rather are meant to privilege a particular, religiously-rooted view of poverty over competing, equally religious beliefs about the nature of social justice. At one end of the religious spectrum are those, like Catholic Charities or the Jewish Welfare Board, who approach poverty as a sign that social institutions are not functioning properly, and believe they have an obligation to work for systemic remedies—and in the meantime, an obligation to assist the needy. At the other end of the spectrum are some—but certainly not all—Evangelical views that poverty is evidence of moral defect, and that the way to eradicate it is to bring the poor to Jesus.

Those who are heavily invested in the arguments over Charitable Choice at either end of this religious/ideological spectrum will undoubtedly remain wedded to their perspectives, but that does not mean that a more engaged and productive discourse is impossible. The organizations and individuals most involved in the day-to-day efforts to provide social services are rarely invested in either of these polarized approaches, and are open to more productive discussions about how to improve the system. Based upon what is known ten years after PRWORA, how might such a discussion proceed?

To the extent possible, productive conversations should focus on “how” rather than “whether.” There is no reason—constitutional or prudential—that government should exclude (or favor) religious social service providers. Religious organizations have always been an important part of the social service contracting mix, and they will undoubtedly continue to be. Those who are opposed to—or leery of—a role for faith-based providers need to acknowledge that reality. For their part, those who favor an increased role for FBOs need to acknowledge that there are some faith-based organizations and/or services that do fall outside the constitutional pale. Government can and does purchase secular services from religious vendors; but government is constitutionally prohibited from purchasing religious services, or from entering into transactions that will have the effect of financially supporting or endorsing religion. Most Charitable Choice and other faith-based initiatives are neither “constitutional” nor “unconstitutional” per se; their propriety depends upon the manner in which they are implemented. If we can focus the national conversation on the appropriate questions, we stand a much better chance of getting useful answers.

Suggestions for reframing the discussion of Charitable Choice have begun to emerge. As David Campbell has suggested (2002), treating contracting choices as a zero-sum game in which government and faith-related organizations represent alternatives is unproductive. Campbell suggests that we focus more on the difficult task of integrating fragmented services at the community level, improving communication and co-operation among the various grass-roots organizations involved. A number of people have advocated for greater emphasis on capacity-building for grass-roots organizations—whether secular or faith-based—and for the formation of umbrella organizations that would act as intermediaries between government agencies and small, grass-roots providers. Such umbrella organizations could handle the accounting and legal compliance reporting that so many small organizations (whatever their faith component) find costly and burdensome, and they could also provide valuable assistance in integrating social service networks.

Once we refocus the discussion—if, indeed, that can be done—the need for credible, empirical data becomes clear. If we argue in the abstract, or from ideological perspectives, data is irrelevant; if we are genuinely concerned with concrete performance, we need to agree on basic definitions, and on organizational and performance indicators. For those who support Charitable Choice and the Faith-Based Initiative because they want to improve the safety net available to those in need, replacing hypotheses with evidence should be a priority. Similarly, if the common goal is to provide more help for our impoverished citizens, then the first—and presumably easiest—point of agreement should be ensuring the adequacy of resources available to support programs that are found to be effective. In an article posted on the website Beliefnet in early 2005, a former White House aide and ardent proponent of the Faith-Based Initiative made precisely that point. David Koh wrote that “the faith-based initiative and compassionate conservatism” had been left at “precisely the place Gov. Bush pledged it would not go.”

“ [I]t has done the work of praising and informing but it has not been given the ‘resources to change lives.’ In short, like the hurting Charities it is trying to help, the Initiative has been forced to make bricks without straw…The Administration…has pushed an ambitious domestic agenda; three huge budgets have been submitted, each of which had billions of dollars for other domestic priorities but lacked any new money to pay for ‘compassionate agenda’ promises, which are ever more in need of fulfillment. After all, there are now more poor Americans than ever before.”

One of the reasons there has been so little evaluation of the efficacy of government-sponsored social services is that there are not even enough resources to provide the promised services to all who qualify—let alone funds for evaluation and proper program monitoring. If this seems penny-wise and very pound foolish, it is; without the ability to evaluate providers and determine the outcomes and comparable merits of their programs, we have no way to know whether we are spending public dollars wisely or throwing good money after bad. In the absence of such information, it is much more likely that public managers will favor contractors with whom they are comfortable, or those whose beliefs (programmatic, religious or otherwise) they find compatible.

If Charitable Choice and the Faith-Based Initiative provide an impetus for addressing these questions, for examining and improving America’s social safety net, it will have proved to be a success. If it merely provides a temporary diversion from the hard questions, it won’t be.