America’s Subnational Welfare State

The decentralization of social benefits and the turn to nonstate providers has provoked a surge of interest in the subnational dynamics of welfare states.^{Footnote 9} Efforts to understand how states and local governments have approached their social policy responsibilities have tracked and explained variations in spending on benefit programs across the states.^{Footnote 10} One of the central findings of this literature is the correlation between race and spending. States with larger numbers of African Americans and Latinos exhibit lower levels of spending for social welfare.^{Footnote 11}

There is no doubt that state spending levels provide one crucial indicator of variation in social benefits—for both entitlement programs and social services. By recent estimates, human service nonprofit organizations derive 65 percent of their revenue from public dollars.^{Footnote 12} After the federal government devolved and cut federal social programs during the 1980s, states not only acted as pass-throughs for federal dollars, they also took on greater responsibility for benefit levels in key programs.^{Footnote 13} The federal government allows states exceptional leeway in devising programs for the poor as authors including Bruch, Meyers, and Gornick observe.^{Footnote 14} Examining eleven core safety net programs for which states and the federal government share responsibility, they find that programs where states have the greatest authority exhibit the most variation. Local governments spend far less on redistributive purposes but even these small amounts can provide important sources of nonprofit income.^{Footnote 15}

But, we suggest, spending represents only a piece of the puzzle. The organizational infrastructure for serving the poor also varies on several other critical dimensions that determine the scope of social programs, their accessibility, and the politics that surround them. By themselves, spending patterns do not tell us enough about the configuration of public and private institutional relationships that organize the politics and delivery of public benefits for low-income people. As Esping-Andersen’s classic comparison of national welfare states shows, institutional relationships anchored in distinct political and social connections can create policy variation even within similar levels of spending.^{Footnote 16} Moreover, as institutionalist theorists argue, configurations of institutions and ideas can work to reinforce one another over time, so that spending may itself be influenced by institutional arrangements.^{Footnote 17}

We argue that the delegated state should be seen as a set of organizational relationships that co-evolve as governments and nongovernmental organizations grow and build relationships over time.^{Footnote 18} This perspective, as articulated by Morgan and Orloff, sees states “as encompassing multiple institutions, varying forms of interpenetration with civil society, multiple scales of governance, and potentially contradictory logics.” In a similar vein, Clemens characterizes the relationship between state and social organizations as symbiotic.^{Footnote 19}

Most configurative analyses of social policy take the nation-state as their focus. In these analyses, the United States represents a low-spending liberal model that outsources extensively to nonprofit organizations.^{Footnote 20} But we know that sharp differences have characterized regions of the country since the Founding. On the eve of the New Deal, Cybelle Fox identified three worlds of welfare linked to distinct labor markets, political contexts, and racial systems.^{Footnote 21} Scholars recognize that even the most politically assertive initiative of the Great Society, the War on Poverty, relied on local organizations to carry out federal goals. States and localities won greater autonomy in shaping social policy after the 1970s as the federal government devolved decision-making and reduced spending on programs that assist the poor. Have different paths of regional development in the United States created enduring distinctions in our approach to social welfare? Or have federal initiatives since the New Deal worked to lessen regional institutional variation?

We argue that federal policies have allowed regional differences to flourish by granting state and local governments extensive discretion in implementing assistance to the poor. After the 1960s, Washington moved away from what Desmond King calls a “forceful federalism” exemplified by the War on Poverty’s categorical grants and vigilant monitoring of local implementation.^{Footnote 22} The devolution and block granting of federal social policy in the 1970s combined with support for outsourcing in the late 1960s and for-profit privatization in the 1980s, gave state and local institutions broad leeway in how they chose to assist the poor. The 1996 welfare reform (Temporary Assistance to Needy Families or TANF), which replaced the federal entitlement to cash support with block grants to the states, further enhanced state and local discretion.^{Footnote 23} TANF also magnified the importance of services as access to cash assistance became sharply limited.^{Footnote 24} In the process, two very different local models for assisting low-income residents emerged.

The first, the civic-public model, features a robust nonprofit sector, strong local philanthropy, and close bonds between these civil society organizations and local governments. Large experienced nonprofit organizations rely on public funds and can expand their resources by applying for grants from governments as well as philanthropic foundations. For example, Reckhow shows that cities with more nonprofit organizations secure more federal grants than those without regardless of need.^{Footnote 25} Similarly Lowry and Potoski show the central role that locally-based interest groups play in securing competitive grants.^{Footnote 26} As we will see, many of these nonprofit organizations are religiously affiliated or have religious origins. But in accepting public funds, they enter the civic domain and abide by its rules, which prohibit proselytizing.

A robust nonprofit ecology is significant because it can defend or even expand public dollars targeted at low-income residents through organized advocacy.^{Footnote 27} Scholars have long recognized the reciprocal relationships that can develop between nonprofits and government.^{Footnote 28} Government funding can create vested interests that lobby to retain their funding when it is threatened; these interests may also engage in broader advocacy campaigns aimed at increasing the overall supply of resources.^{Footnote 29} Studies show that many nonprofits engage in advocacy, either on their own or as part of broader coalitions.^{Footnote 30}

Philanthropy matters both because it provides financial resources for nonprofits, but also because it can serve as an organizational node for the nonprofit community. Community foundations and federated organizations such as the United Way play a special role in this regard since they have an explicit geographical mission. These organizations supply information, training, and organizational connections that build a community of expertise around human services and affordable housing. Their combination of geographic reach and information resources makes these organizations especially well situated to identifying new needs and helping to start new organizations to address them.^{Footnote 31}

The second model, the religious-private model, displays a significantly smaller nonprofit sector, weaker ties between government and nonprofits, and less robust philanthropic institutions. Instead, this configuration of institutions features a larger role for private religious charity and for-profit government contracting. Religious charity is an essential element of the safety net in all parts of the country. Yet it is no substitute for public funds. A wide literature has shown that congregations are inherently limited in the services they can provide.^{Footnote 32} In the repeated waves of the National Congregations Study, sociologist Mark Chaves found that congregations tended to supply small-scale short-term emergency assistance. In 2012, only 14 percent of congregations surveyed had a staff person who devoted at least one-quarter of their time to social services. Only 4.9 percent had applied for a government grant in the past two years and only 8.9 percent reported starting a separate nonprofit for human services or ministerial outreach in the past two years.^{Footnote 33}

The religious-private model also relies more heavily on contracting with for-profit firms to deliver social services and affordable housing. A large presence of for-profit stakeholders creates a distinctive politics surrounding assistance to the poor. Since firms defend their own contracts, they are less likely to join broader coalitions of nonprofit social service providers to advocate against cuts or to press for additional public dollars. Because they need to make a profit, these contractors avoid entering less remunerative service functions, leaving these services with fewer political allies. For example, few private providers enter the field of serving the homeless mentally ill.^{Footnote 34} The ties that for-profit contractors forge with governments are more narrow and particular than those built by nonprofits. In contrast to nonprofits, for-profits also do not attract philanthropy and do not allocate resources to help cross-subsidize less profitable but worthy activities.^{Footnote 35} Inadequate contract design and weak government monitoring capabilities, which often accompany enthusiasm for contracting with for-profits, exacerbate the risk of rent-seeking and cream skimming that abandons the least promising clients.^{Footnote 36}

Critical questions about the trajectories of social support in different metropolitan areas can only be understood by taking into account these distinct organizational configurations. How did they emerge? Will the nonprofit laggards come to resemble the leaders or are they following fundamentally different tracks? Even if spending increased, would significant differences endure across metros? Organizational configurations also provide leverage for understanding the justification and ideas that inform the politics of social welfare in different places. Finally, organizational relationships illuminate the challenges that external actors, such as the federal government or national philanthropic foundations, are likely to confront if they attempt to influence local capacities to implement social welfare programs.

Varieties of Local Infrastructure Serving the Poor

To investigate variation in the organizational configuration of social services and affordable housing programs, we first probe the historical development of social infrastructure for assisting the poor in different parts of the country, showing how the distinctive alchemy of European immigration, religion, race, and the timing of urbanization created divergent institutional legacies for assisting the poor. In the North and Midwest the civic-public model thrived after the 1960s as increased federal spending combined with devolution of authority to states and localities and delegation to nongovernment organizations. In the South and Mountain West, a much weaker nonprofit infrastructure and a preference for assisting the poor through religious congregations constrained the civic role. The later timing of urbanization also meant that these jurisdictions increasingly turned to for-profit firms to deliver social programs. The organizationally and politically weaker civic infrastructure inherited from the past, together with the tremendous growth of these regions, has created a religious-private model of social assistance quite distinct from that in the North and Midwest.

Having documented these distinct patterns, we then test how these distinct historical legacies continue to exert influence by devising a Metropolitan Institutional Support Index. The Index ranks twenty-five metropolitan areas in different regions of the country on the public, civic, and private characteristics of the organizations that serve low-income residents. The metros are listed by region in table 1.

Table 1 Population five largest metros in five regions of the U.S, 2005-2009

Source: U.S. Census Bureau, 2009a.

The Roots of Regional Differences

Institutional variations in support for the poor trace their roots to sharply distinct regional histories rooted in patterns of European immigration, religion, race, and the timing of urbanization. In the North and Midwest, European immigration, religious organizations, and government intertwined over time to create a virtuous cycle between robust civil society organizations and local political support for serving low-income residents. In the Pacific West, San Francisco, and to a lesser extent, Seattle, shared a similar history. By contrast, the South and the Mountain West, and other metros in the Pacific West exhibited very different patterns. In the South, the combination of low levels of immigration and legalized racial segregation constrained the development of civil-society organizations focused on the poor and sharply circumscribed the public sphere. Likewise in the Mountain West, limited European immigration, civic exclusion of Mexican migrants, and the political dominance of Anglo economic elites provided little impetus for developing civil-society organizations or a public sector focused on the poor.^{Footnote 37}

Comparative analyses of welfare states highlight the enduring impact of different religious traditions on the organization of social policy and its particular importance in programs that serve the poor.^{Footnote 38} Religion shapes beliefs about when and how to assist the poor but it also provides an organizational infrastructure that states can use as they build welfare systems.^{Footnote 39} In national comparisons, the United States embodies the Calvinist beliefs that prize personal responsibility and view personal or religious charity as the best way to assist the poor.^{Footnote 40} But the United States was a site of fierce religious competition that played out in state and local political arenas. This competition led to a variegated pattern of public religious relationships that supported a diverse set of organizations to assist the poor. As a consequence, regionally distinct religious patterns stemming from immigration had a profound impact on the infrastructure of support for the poor and its relationship to politics.

The intertwining of European immigration, religion, and government created a potent mix in the large cities of the North and Midwest. Competition between mainline Protestant elites and the largely immigrant Catholic and Jewish communities drove the creation of new organizations. Seeking to tame the social ills associated with large-scale immigration and to create services free from often-corrupt local machines, mainline Protestants established independent charitable organizations, such as Charity Organization Societies and later, community philanthropy. Immigrant Catholic and Jewish communities countered by creating their own social assistance organizations.^{Footnote 41}

Over time, however, organizations with religious roots became more secular. Protestant congregations spun off separate secular civic organizations, and Catholic and Jewish social assistance institutions became religiously-sponsored, but government-friendly, agencies. Even before the New Deal, the highly centralized Catholic Church, with its large infrastructure of orphanages and hospitals, forged strong relationships with local politicians and gladly accepted local public funds.^{Footnote 42} When the New Deal offered cities federal dollars, Northern and Midwestern cities used these resources to assist European immigrants, and in the process strengthened social service organizations with federally supported employees.^{Footnote 43} Mainline Protestant, Catholic, and Jewish organizations also established national federations, positioning these organizations to grow when the federal spigot opened to fund private social service agencies in the 1960s.^{Footnote 44} Over time, these religious federations became highly professionalized and gained “extensive experience implementing government contracts.”^{Footnote 45}

African Americans and Latinos could not count on equal treatment in Northern and Midwestern cities but they could expect their claims for public assistance to receive some attention.^{Footnote 46} Black churches and missions offered support to those excluded from white-serving settlement houses as did the social-support institutions created by secular black reformers.^{Footnote 47} Yet, it was not until the War on Poverty that independent black and Latino neighborhood organizations flourished with the support of federal dollars.^{Footnote 48} Often church-related, these groups, such as New York’s Abyssinian Baptist Church, were well positioned to command large-scale public funding in the era of devolution.^{Footnote 49}

In the South and the Mountain West, by contrast, limited European immigration, legalized racial hierarchy, and different religious traditions led to a very different configuration of social organization and government support for the poor. Although Progressive reformers created some social-welfare organizations in cities outside the North and Midwest, these regions lagged far behind in the number of secular benevolent organizations.^{Footnote 50} In the South, African Americans created their own religiously-linked social-welfare institutions, but political disenfranchisement meant little or no public funding. Segregated social institutions blocked the organizational cross-fertilization that was emerging in the North and Midwest in the 1960s.^{Footnote 51} It likewise limited the development of community-based philanthropy. Comparing the South to the North and Midwest, David Hammack notes that “in religion and in other configurations of philanthropy and civil society the Gilded Age South was another country, or two other countries, white and black.”^{Footnote 52} Most of the philanthropic dollars that entered the South came from northern foundations, such as the Rockefeller Foundation and the Julius Rosenwald Fund, both of which provided vital support for black education.^{Footnote 53} In the Mountain West and the Pacific West, Mexican immigrants faced deportation as the Great Depression deepened, leaving little institutional legacy on which to build organizations to assist the needy.^{Footnote 54} Both regions remained influenced by Progressive-era reforms that included voting restrictions and elite-dominated politics well into the 1960s.^{Footnote 55}

Low levels of European immigration also gave the South and Mountain West a distinctive religious profile. In place of the Catholics, Jews, and mainline Protestants that dominated in the North and Midwest, evangelical Protestants thrived in the South and the Mountain West. The rise of evangelical Protestantism coincided with the formation of a distinctive southern identity in the antebellum South.^{Footnote 56} As table 2 shows, evangelical Protestants continue to dominate in the South and the Mountain West to this day.^{Footnote 57} These religious differences had significant implications for the growth of an organizational infrastructure to assist the poor. With their emphasis on personal salvation, evangelical denominations historically shunned government funding, preferring to pursue their charitable mission through their congregations.^{Footnote 58} Yet, as noted earlier, congregations are quite constrained in the types and amount of social services and housing they can deliver. Moreover, the organization of evangelical religions makes them especially poor vehicles for redistribution. Compared to mainline Protestants and Catholics, the much less centralized structure of evangelical churches makes it difficult for them to organize services beyond the congregation and share resources across communities. Charitable giving as a percentage of income is higher within evangelical denominations than mainline Protestants and Catholics.^{Footnote 59} However, the preference for delivering services through congregations and reluctance to engage with government means that the high levels of personal generosity do not result in expansive services to the poor.

Table 2 Religious affiliations in the twenty-five largest metros in five regions of the U.S., 2005-2009

Sources: Association of Religion Data Archives 2010; author’s calculations.

Federal efforts to help build organizational capacity in these metros have achieved mixed results. The War on Poverty, the clearest example of “forceful federalism,” succeeded in creating new organizations to serve poor communities in the South and Mountain West. Poor minority communities, long ignored by mainstream social-service agencies, especially benefitted from the new organizations.^{Footnote 60} Even so, resistance from local elites meant that War on Poverty programs spent less in the South than in other regions.^{Footnote 61} These metros remained ill equipped to benefit from the flow of funds after 1967 when the federal government authorized nongovernmental organizations to receive federal funds.^{Footnote 62}

Seeking to promote more engagement of religious organizations in assisting the poor, presidents, starting with Bill Clinton, implemented “charitable choice,” a set of incentives designed to draw more religious organizations into delivering social services in the aftermath of welfare reform. Yet there is little evidence that these measures have prompted evangelicals to overcome their reluctance to accept public funds by accepting the prohibition on proselytizing that comes with them.^{Footnote 63} African American congregations mark a striking departure from this pattern. As in the North, African American churches in the South and Mountain West exhibit much greater willingness to accept government funds.^{Footnote 64} Indeed, in 1998, Chaves found that African American congregations were the most willing of all congregations to apply for government funds.^{Footnote 65} Likewise in a survey of Atlanta-area religious leaders, Owens found that the racial composition of the congregation was the single most important factor in predicting willingness to apply for public funding: predominantly black churches were five times more likely to express support for receiving public funds.^{Footnote 66} However, black congregations that can provide a broad array of services remain exceptions, especially in low-income neighborhoods.^{Footnote 67}

Even with the exception of black churches, the legacies of the past left a much smaller infrastructure for assisting the poor in the South and the Mountain West. In the decades after World War II, evangelical leaders opposed government funding of religious organizations to deliver social services. They believed—with some accuracy—that public funding disproportionately benefited Catholics and mainline Protestants. Although their opposition faded by the 1980s, and large evangelical institutions such as universities and hospitals gladly accepted public funds, the evangelical infrastructure for assisting the poor continued to lag behind.^{Footnote 68}

Combined with the later timing of urbanization, the weak nonprofit infrastructure in the South and Mountain West facilitated the move to for-profit outsourcing. Most of the metros of the South and Mountain West were, at most, small cities at the end of World War II. Some were barely dots on the map. Much of their growth—both in population and in low-income population—occurred after 1980. Yet it is precisely during this time that for-profit organizations began to compete in earnest with nonprofits for government contracts focused on low-income people.

The welfare reform law of 1996, which transformed welfare into block grants and gave states wide discretion, has been accompanied by increased outsourcing to private firms. As states embraced the idea that markets could solve social problems better than government, large corporations, such as Lockheed Martin, entered the field of social services.^{Footnote 69} Outsourcing to for-profits could more easily be justified as the logical decision in a context where nonprofit capacity was weak; moreover, a weaker nonprofit sector could not put up a fight against outsourcing to for-profit organizations.

The Metropolitan Institutional Support Index

How do these distinct organizational legacies affect the contemporary institutional configuration of support for low-income Americans in metropolitan America? To find out we use four indicators to assess the organizational infrastructure for assisting the poor in different metropolitan areas. They include nonprofit spending focused on low-income residents; spending by local philanthropy and federated giving organizations; the size of the workforce that delivers services to low-income residents; and the percentage of the private sector workforce employed by for-profits. Because our interest is in understanding regional differences, our data set includes the five largest metropolitan areas in the North, Midwest, Pacific West, Mountain West, and South.^{Footnote 70} The metros are listed by region in table 1.

To gauge the strength of the civic sector, our first indicator measures the level of spending by nonprofit organizations that serve low-income populations. It relies on 2010 data from the National Center for Charitable Statistics to examine fourteen categories of organizations that specifically target low-income residents. These categories include food aid, homeless shelters, affordable housing, and human services.^{Footnote 71} We did not include health because it is difficult to disentangle spending on low-income residents in this category. We cleaned the data to eliminate nonprofits that do not serve the local community or that do not serve low-income residents.^{Footnote 72} We norm the aggregate expenditure for all nonprofit organizations serving low-income residents with respect to the size of the population (dollars per person).^{Footnote 73} We rely on the 2005–2009 American Community Survey for population estimates corresponding to our uniformly defined metropolitan areas over time.^{Footnote 74}

Our second indicator provides information about the strength of the local philanthropic sector, examining the spending of community-based philanthropy and federated giving programs, such as the United Way. Evidence for the impact of philanthropy on the density of the nonprofit sector is mixed. Studies that measure individual “philanthropic propensity” show no impact. Measured as “grants and donations given through private foundations or individual citizens,” however, Lecy and Van Slyke show that philanthropy contributes positively to nonprofit density, albeit at a much lower level than government spending does.^{Footnote 75} But philanthropic organizations also matter for civic capacity because they can serve as an organizational coordinating point for the nonprofit community.^{Footnote 76} Community foundations and federated organizations such as the United Way play a special role in this regard since they have an explicit geographical mission. These organizations provide information, training, and connections that build a community of expertise around human services and affordable housing. As such they help nurture the supply of organizational founders and administrators whose presence contributes to a robust nonprofit sector.^{Footnote 77} The combination of geographic reach and information resources makes these organizations especially well situated to identifying new needs and helping to start new organizations to address them.^{Footnote 78} Recognizing their importance for civic capacity but their limited impact on spending, we combine these two measures into a single indicator relying on data from the NCCS data set.

Our third and fourth indicators assess the extent and character of the social-service sector, providing measures of the total size of the social-services workforce and distinguishing whether the services are delivered by employees in the public, nonprofit, or for-profit sector. The total size of the social assistance workforce captures the scope of this sector regardless of its public or private character. If some metropolitan areas compensate for small nonprofit sectors with large public or for-profit social-services sectors, we would see that reflected in this indicator. Using Quarterly Workforce Indicators, we aggregate public and private employment data in three social assistance categories. Within the social assistance classification, we selected three of four sub-categories—family services; community food and housing, and emergency and other relief services; vocational and rehabilitation services.^{Footnote 79} We excluded childcare services because it includes organizations serving all income groups.

The percentage of private-sector employees in the for-profit sector captures the prevalence of for-profit versus nonprofit service providers. Using the Economic Census, we aggregate employment in the social assistance sector. We selected the same social assistance categories as those used for the previous workforce indicator.

As a first strategy for identifying regional patterns, we construct a simple Metropolitan Institutional Support Index to rank our twenty-five metro areas on each of these measures. The index assigns a 1 if the metro is above the median in nonprofit spending and a 0 if equal to or below; 1 if the spending of community foundations and federated giving programs is above the median and a 0 if equal to or below; 1 if the metro is above the median in total social-service employment; and 1 if the metro is below the median in percent for-profit and 0 if it is equal to or above the median. The scores are then aggregated so that metros receive a final ranking between 4 and 0. A score of 4 represents a strong civic sector, robust community philanthropy, a large social service and housing support sector, and relatively low for-profit outsourcing. A score of 0 represents the weakest social assistance sectors with low nonprofit capacity, weak philanthropy, and high levels of for-profit outsourcing.

As table 3 shows, the index displays a clear regional pattern corresponding to the two models of organizational infrastructure. Eight of the metros that received scores of 3 or 4 were in the North and the Midwest, with two (San Francisco and Seattle) in the Pacific West and one (Tucson) in the Mountain West. Eight of the eleven metros that received scores of 0 or 1 were in the South and Mountain West, with three in the Pacific West (San Diego, Los Angeles, and Riverside). Table 4 presents the raw data and table 5 the Z scores for the indicators. With the exception of Tucson, these measures confirm the general regional pattern of a much stronger civic sector as defined by the level of nonprofit spending and philanthropic giving in the North and Midwestern metros and a much weaker civic sector in the South and Mountain West metros. Pacific West metros exhibited a mixed pattern with San Francisco and Seattle ranking in the top group. The employment indicators for outsourcing to for-profits versus nonprofits show a similar regional pattern, with such outsourcing far more prevalent in the South and Mountain West. Three metros, Denver, Detroit, and St. Louis, receive the middle score of 2.

Table 3 Metropolitan Institutional Support Index, 2010

Sources: National Center for Charitable Statistics 2010, U.S. Census Bureau 2009a, 2012b, 2014; author’s calculations.

Note: Shading distinguishes metropolitan areas with a high, medium, and low score on the Index: Light grey= 3 or 4; Medium grey=2; Dark grey=0 or 1.

Table 4 Metropolitan Indicators, 2010

Sources: National Center for Charitable Statistics 2010. U.S. Census Bureau 2009a, 2012b, 2014; author’s calculations.

Note: Shading distinguishes metropolitan areas with a high, medium, and low score on the Index: Light grey= 3 or 4; Medium grey=2; Dark grey=0 or 1.

Table 5 Metropolitan Indicator Z-scores, 2010

Sources: National Center for Charitable Statistics 2010. U.S. Census Bureau 2009a, 2012b, 2014; author’s calculations.

Note: Shading distinguishes metropolitan areas with a high, medium, and low score on the Index: Light grey= 3 or 4; Medium grey=2; Dark grey=0 or 1.

To further test these regional patterns we constructed a second version of our index incorporating a measure of independent foundation spending into the philanthropy indicator. This data was drawn from the Foundation Center’s research data set of the 1,000 largest independent foundations.^{Footnote 80} The new philanthropy indicator combined standardized variables for community foundation, federated giving program, and independent foundation spending in order to equally weight spending by any of the three types of foundations because of the limited number of foundations in the Foundation Center data. We do not report the alternative philanthropy indicator here because it is impossible to know whether the limitations of the Foundation Center data introduce bias. The alternative measure resulted in only minor changes to the indicators for two metropolitan areas—St. Louis’s philanthropy indicator flipped from 1 to 0, and Boston’s philanthropy indicator flipped from 0 to 1.

In sum, the regional patterns that first took root over a hundred years ago remain evident in contemporary configurations of institutional support for low-income people. Yet the metros that depart from their regional pattern and the metros in the middle category invite questions about the broader trajectory of these models in the future.

Trajectories of Change

What will drive the future trajectory of the two models? Will the civic-public model and the religious-private model steadily come to resemble one another or will they follow distinct paths? One possibility is that federal social-policy retrenchment will cause the two models to converge with the civic-public model eroding as public funding shrinks. Another possibility is that the two models will move in different directions as the organizations in the civic-public model use their power to enhance the resources going to nonprofits and restrict the role of for-profit firms in social service delivery. By contrast, regions with the religious-private model may delegate more resources to for-profit firms even if they increase spending. Finally, as regions confront economic and demographic change, the boundaries between the two models may begin to blur.

When we examine change in nonprofit spending between 1990 and 2010, metros in all parts of the country display growth but with a few exceptions, the highest nonprofit spenders in the South and Mountain West remain below those in the North and Midwest, evidence of the enduring significance of the civic-public and religious-private models for the nonprofit infrastructure for serving the poor.

Yet, the experience of Detroit suggests that the civic-public model can become vulnerable. Like St. Louis, it sits in the middle of our index. Unlike St. Louis, however, Detroit had a much higher level of nonprofit spending in 1990, and over the next twenty years nonetheless lagged behind the growth rates of other metros in the North and Midwest.^{Footnote 81} A weak local economy is part of the explanation. Detroit’s CBSA, experienced an 8.5 percent drop in median household income between 1990 and 2010, the most severe decline of all our metros.^{Footnote 82} Indeed, when it comes to nonprofit spending, economic growth matters: our 25 metros show a .46 correlation between the growth of median household income and a rise in nonprofit spending over these twenty years.

In addition to declining income, Detroit points to the ongoing role of race in influencing the infrastructure of support for the poor. Depleted by white exodus and decades of decline, Detroit’s nonprofits have strained to support low-income residents, while the city itself fell into bankruptcy. Detroit also highlights the importance of the principal city for the vitality of the nonprofit infrastructure in the region as a whole. When largely black central cities have been left to struggle with rising rates of poverty and declining tax bases, the nonprofit sector of the entire region suffers. Indeed, among our North and Midwest metros, the cities of Detroit and St. Louis have the two largest black populations at 82.7 percent and 49.2 percent respectively.^{Footnote 83}

State spending also shapes the trajectory of the models. Data from the Census of Governments places Michigan near the bottom of all our metros in state spending growth on public welfare and housing and community development from 1992 to 2012.^{Footnote 84} But local factors can create large differences even within states, as the case of California suggests. The three metros in California—San Francisco, Los Angeles, and Riverside—reveal the difficulty of building a strong nonprofit sector in metros whose history diverges from that of cities in the North and Midwest. San Francisco, whose nineteenth- and early twentieth-century history resembles that of the North and Midwest, sits at the top of our index, Los Angeles in the lower category, and Riverside near the bottom. As important as state spending is, these California metros indicate the importance of specifically local factors in building a robust nonprofit sector to serve low-income residents.

European immigration was central to the historical development of the civic-public model, but it is not clear how contemporary immigration affects trajectories of nonprofit development. Demand-side arguments about nonprofit density posit that a more diverse population will produce a bigger nonprofit sector by boosting demand for culturally-specific programs, best provided for by nongovernmental organizations. When we examine our 25 metros, however, we see only a weak negative correlation of .04 between nonprofit spending and foreign-born population in 2010. But a comparison of our two regional models reveals something different. The correlation between foreign-born and nonprofit spending is positive at .2 for the metros in the North and Mideast and a negative .4 correlation for the South and Mountain West. One interpretation is that areas with strong nonprofits increase their capacity as immigration increases, while areas with weak nonprofits cannot respond to a rise in immigration in a comparable way. The divergence between the regions provides support for a supply model of nonprofit growth advanced by Frumkin.^{Footnote 85} The preexisting supply of organizations and personnel in the North and Midwest makes it easier to expand nonprofit capacity to respond to immigrants and can ultimately lead to a virtuous cycle of immigrant incorporation.^{Footnote 86} By contrast, in the South and Mountain West, the weakness of these local resources means that the rise in immigration does not engender a response of comparable scale in the nonprofit sector.

The impact of the timing of urbanization is evident in divergent patterns of employment in the for-profit social services. To be sure, the growth of for-profit provision affected all areas of the country. Between 2002 and 2012, the for-profit sector share of private social services employment grew by roughly similar percentages in all regions of the country. However, the balance between for-profit providers and nonprofit providers varied sharply across the regions, reflecting their very different starting points. In the Northern states, the for-profit sector grew from 9.7 percent of all private employment in the social services in 2002 to 18.5 percent in 2012. In the South, for-profit employment grew from 14 percent to 37 percent in the same time period. By 2012, for-profits accounted for 41 percent of the private employment in social assistance in the Mountain West metros.^{Footnote 87}

These trends suggest that if social spending in the South and the Mountain West were to “catch up” to the other areas of the country, the additional spending would do little to expand the nonprofit sector. In the absence of a strong nonprofit sector, for-profit providers face few barriers to entry. And for public officials, for-profits may present the simplest and most efficient choice for providing services as their populations and needs grow. But many of the metros with large for-profit sectors, such as Riverside, Las Vegas, and Phoenix also have very small public sectors. The public-sector social-services workforce in these metros is relatively small, with 2, 9, and 17 employees per 100, 000 persons, respectively compared to an average of 44 employees per 100,000 persons in our metros.^{Footnote 88} With weak public capacities, these metros are poorly equipped to conduct the monitoring of for-profits needed to guard against rent seeking and cherry picking clients.

Philanthropic assistance offers a possible source of external support to spark change in nonprofit strength. A handful of large national philanthropic organizations have played a major role in building the nationwide nonprofit infrastructure for affordable housing and have simultaneously bolstered a wide range of social services.^{Footnote 89} Foundation funding is widely valued because it supports innovation and organizational networking that public dollars rarely support. Yet a recent review of 169,000 community- and economic-development grants by the largest foundations between 2008 and 2013 found that “the density of nonprofit organizations and the presence of large local foundations” were the best predictors of grant receipt.^{Footnote 90} As this study and other similar findings indicate, foundations contribute to the virtuous cycle in the North and Midwest where nonprofits are historically strong but they find fewer toeholds for involvement in places with historically weak nonprofit sectors.^{Footnote 91}

Over one hundred years ago, demographic and religious forces created the unique alchemy that lay behind the civic-public model of assistance in the North and Midwest. The very different history of the South and the Mountain West fostered a private-religious organizational configuration, which provides weaker support to low-income residents. These regional differences have largely endured and taken on a new dimension with the growth of for-profit services. Patterns of change over the past two decades suggest that the institutional arrangements that support the two models will be difficult to alter. Within this overall pattern, higher income metros enjoy an advantage when it comes to creating new nonprofit capacities. America’s delegated state puts poor residents in declining Midwestern metros, especially those with large black populations, at risk of reduced services. Simultaneously, though for different reasons, it leaves less well-off metros in the South and Mountain West with a frail set of institutions to serve the poor.

By clarifying the enduring effects of two distinct developmental patterns, the historical-institutional approach to subnational variation helps to reconcile the disparate threads in contemporary studies of nonstate organizations in American social policy. These studies tend to fall into one of two streams: efforts to document bottom-up organizational creativity or the impact of top-down policies of retrenchment. In emphasizing organizational innovation, the former research stream views the emergence of the third sector in largely positive terms. Research has documented the growing capacity of Community Development Corporations (CDCs) since the 1970s and the system of philanthropically supported finance that has made them the main vehicle for building affordable housing.^{Footnote 92} Likewise, studies have praised the growth of large human service organizations capable of delivering a diverse array of social programs.^{Footnote 93} The second research stream, in focusing on retrenchment, tells a very different story. In place of ever-growing organizational capacity, this work highlight the holes in social provision, the subterranean politics that makes it difficult to monitor private rent seeking, and cuts in public funding that weaken nonprofits or lead them to act more like for-profits.^{Footnote 94}

Analyses that celebrate organizational creation tend to treat places where these innovations have not thrived as exceptions that will eventually learn to innovate. By contrast, studies that emphasize retrenchment highlight the power of federal and state policy to undermine local arrangements. We argue, instead, that there may be a distinct institutional logic to the “underperforming” places that puts them on fundamentally different trajectories. Similarly, economically strong regions with long-established nonprofit sectors may have more ability to withstand federal retrenchment policies.