1 - The Welfare Workforce

Summary

Why does the supply of mental health care vary across countries? Moreover, why would the state supply services to those who cannot demand them? This chapter introduces how a comparative, political-economic, and historical perspective can explain mental health care outcomes, as well as how studying mental health can inform comparative political economy. It then turns to the theoretical argument, explaining why and how public sector managers and workers – the “strange bedfellows” of the “welfare workforce” – shape the supply of public social services. The chapter closes with a sketch of the book’s research design and how it structures the following chapters

December 28, 2020, was “Demolition Day” at Allentown State Hospital. Nearly ten months into the Covid-19 pandemic, the decision to raze a health care facility might appear counterintuitive. After all, the calls to protect and extend the medical supply had sounded loud and clear. Such calls were especially relevant to a psychiatric service such as this one, where patients faced a high risk of Covid transmission within the hospital but even worse outcomes if discharged. Increasing capacity in order to maintain social distancing among patients was crucial. But in effect the demolition of Allentown State Hospital was the last of a long line of policy changes in mental health care. The facility had been nonoperational for over a decade, following similar closures at the nearby Mayview State Hospital in 2008, the Harrisburg State Hospital in 2006, and a hundred other state mental hospitals across America (SAMHSA 2011; Witkin et al. 1996). Like many of its peers, Allentown State Hospital was over 100 years old and eligible for inclusion in the US National Register of Historic Places. Simply maintaining its 217-acre campus, though, cost the state about $2 million (USD) each year. The hospital’s fate, in the end, was lamentable but cost-effective, and not unlike that of most other American public mental health care facilities.

Meanwhile, across the Atlantic, another mental hospital was faring very differently in the pandemic. Just a few months later, in March 2021, the new director of the Centre de Postcure et de Réadaptation de Billiers announced the launch of three new projects: an in-home care service, an expanded vocational training program, and a short-term housing unit to accommodate the expanded clientele. This was on top of the facility’s existing services. Located in a small French village on Brittany’s western coast, the campus already housed 145 beds, an arts center, and sheltered employment workshops in four areas: agriculture, horticulture, technical fields such as carpentry and mechanics, and dining and hospitality (the site’s restaurant and guest accommodation serve as the training ground). The facility also provides each patient with individualized attention from a large team of professionals, including a psychiatrist, a nurse, a psychologist, a neuropsychologist, an occupational therapist, several social workers, several special needs teachers, and even a sports coach. Many of these services are required by law, as this area is home to one of France’s approximately 1,200 psychiatric “sectors,” which stipulate at least one such care team for about every 60,000–70,000 people (Chevreul et al. 2015; Coldefy et al. 2009).¹ Evidently, mental health care in France did not face the same fate as that of its American counterpart.

Despite the two countries’ dissimilar political reputations when it comes to social policy, the explanation for these alternative fates is not obvious. During the pandemic, the US government expanded some social policies, while the French government sought ways to support markets and the private sector. Historically, too, the general health system in France has been similar to that of the United States, insofar as they share a strong tradition of private practice and payment. Moreover, the demand for mental health care is not higher in Billiers than in Allentown. In fact, the opposite may be the case: The population of Allentown is well over 100 times that of the Billiers. If the prevalence of schizophrenia is around half a percent (estimates vary, WHO 2022), Allentown would have over 500 residents in need of long-term psychiatric attention, but Billiers would have just five. Nor is Billiers, and its surrounding region, more financially capable of supplying mental health care. As is so often the case with these services, their high costs are absorbed by the government, where tightening budgets are becoming less likely to accommodate them. Only a few years earlier in 2015, the Breton health authorities had sought to redress an annual structural deficit of €300,000 (about $360,000) by halving care.²

There is another important similarity between Allentown State Hospital and the Centre de Postcure: their respective roles as employers in deindustrializing local economies. The Billy Joel anthem – “Well we’re living here in Allentown / And they’re closing all the factories down” – might as well apply to Billiers (Joel 1982). Situated between the ports of Saint-Nazaire and Lorient, Billiers can no longer depend on the shipbuilding, cargo, and naval industries for employment the way it once did; just as Allentown can no longer depend on the coal, steel, textile, and cement industries of the past. In both places, the one-time reliance on industrial production has shifted dramatically to services, especially social services. Jobs at the local psychiatric hospital are some of the few well-paid and protected jobs still left in town.

The difference, as I will demonstrate in this book, lies in the political response of hospital employees to the respective cuts. Even though, prior to its closing, Allentown State Hospital employed more than triple the staff of the Centre de Postcure (379 full-time employees compared to 110), resistance was more limited in Allentown. To be sure, the closure upset the Allentown employees. The state had transferred a quarter of the staff to another facility and left the rest looking for work or otherwise forced to retire. Unions representing the workers, such as the American Federation of State, County, and Municipal Employees and the Service Employees International Union, argued that the state had made this decision without their consent, failing to account for both their members’ livelihoods and the fate of the patients who would be sent elsewhere.

Picket signs in Billiers – “Health and Jobs in Danger” – raised similar concerns at the prospect of budget cuts in 2015, but with the opposite result. The Confédération française démocratique du travail, France’s largest labor union for public employees, organized a sweeping mobilization; 5,000 people signed a petition opposing the cuts. With their supporters, hospital staff marched four kilometers from the Centre de Postcure to the neighboring town of Muzillac and then on to the prefecture in Vannes. Alongside them were 17 local mayors. Elected officials had voted unanimously to support the mobilization, even closing a town hall and post office on the same day as an endorsement. The group of workers, unlike their counterparts in the United States, achieved their objectives. “This new future will help to preserve jobs,” the local press concluded, “a major advantage of these projects.”³

The contrasting trajectories of mental health care in Billiers and Allentown help to answer a puzzling question of political economy: Why would the government provide social services to clients who cannot demand them? Welfare states have minimal incentives to do so.⁴ Not only are the costs of supplying such services extraordinarily high, their beneficiaries rarely request them. Marginalized groups, such as people with chronic and severe mental illness, possess neither the political clout nor the purchasing power to command robust government attention. Other examples abound. Consider the paradoxical, though varied, choice to supply care homes to the frail elderly, health services to noncitizen immigrants, and even schools to nonvoting-age children. In each of these cases, disenfranchised clients cannot compel the welfare state to deliver these expensive services, but it does, sometimes.

The cross-national variation in mental health care, the empirical subject of this book, throws that puzzling theoretical question into stark, pragmatic relief. What occurred in Allentown and Billiers is in fact typical of national mental health policy patterns in the United States and France, respectively. The supply of US psychiatric care is so limited that only a small portion of the 50 million Americans who have a mental illness obtain treatment for it (Kohn et al. 2018; NIMH 2021). France, meanwhile, supplies more than double the care of its counterpart, facing fewer problems of shortage. But a mid 20th-century observer could not have predicted that the supply of public mental health services in France would eventually triple that of the United States. If anything, the opposite might have been more likely, as supply in France had degenerated and declined during the Second World War.

What is more, the cross-national differences in mental health care do not necessarily map onto the general trends in social welfare provision across countries. The large public welfare states of Norway and Sweden, for example, diverge on mental health care. More generally, these cross-national differences exist despite each of these countries’ efforts to “de-institutionalize” patients. A process that began after the Second World War but which peaked in the 1970s and 1980s, deinstitutionalization sought to depopulate mental asylums and, social reformers promised, replace them with non-hospital, “community-based” alternatives. A range of factors contributed to this process, including postwar political enthusiasm for social reform, academic critiques of large mental institutions, and the rise of a film industry that spread these critiques into popular culture.⁵ This international impetus, though, did not produce identical outcomes in domestic policy. To understand why, this book presents a comparative and historical analysis of mental health care development prior to and during deinstitutionalization (from the mid 19th to late 20th centuries) in the four countries mentioned, selected for their ability to inform broader trends in mental health services provision across the affluent democracies (where the deinstitutionalization movement first took root).⁶ I draw on rich and often unexplored archival sources to explain the political origins of contemporary mental health care systems.

This book points to the central role of workers employed by the welfare state – what I call the “welfare workforce” – in shaping the supply of social services. These workers, an important segment of postindustrial labor markets and the linchpin of the contemporary trade union movement, can become the most ardent defenders of their source of employment and, by extension, of services for the vulnerable. Most often employed in the public sector, they encompass a wide range of occupations and can include, for example, nurses, teachers, caregivers, facility support staff, and, importantly, managers. Absent powerful beneficiaries, the scope of public social services depends on the political influence of these workers’ trade unions.

Yet they do not always succeed; their success can depend on whether managers participate in this advocacy, a unique source of political power for the public sector workforce. Like private sector managers, public sector managers can access the levers that change production capacities and alter compensation; however, they often happen to benefit from the same pay and protections as their subordinates. In short, public sector managers can make policy decisions that jointly benefit themselves and their employees. A coalition between these two groups within the welfare workforce can expand employment and service provision. Its absence, though, can reduce them. I find that a public labor–management coalition is more likely when public sector managers organize independently of private sector managers. This arrangement amplifies the political voice of public sector managers and facilitates their alliance with their employees. Such was the case in the development of French and Norwegian mental health policy, in contrast to that of the United States and Sweden, where the absence of a public labor–management coalition enabled the reduction of public mental health services.

To elaborate on this argument and the objectives of this book, this introductory chapter unfolds in three parts. First, it explains the importance of studying the field of mental health with the tools of comparative political economy; doing so can illuminate both areas. Herein the phrase “political economy” refers to the relationship between governments and markets, where the adjective “comparative” refers to how this relationship varies across countries. Although I use the term “mental health” to refer to all conditions requiring psychiatric or psychological care, I focus in particular on the most challenging and severe conditions that consume most of the public mental health system’s resources (e.g., schizophrenia, chronic depression, severe bipolar disorder, and other severe mental illnesses). This first section also presents the theoretical puzzle and empirical questions of this book in greater detail. Second, it turns to the argument, providing more information about its logical coherence. Third, the chapter closes by articulating the research design and how it structures this book.

Why Study the Comparative Political Economy of Mental Health?

Comparative political economists have spent very little time thinking about mental health policy, and mental health specialists have spent very little time thinking about comparative political economy. This has produced deficiencies in both areas. More fundamentally, the absence of engagement across these two fields is a disservice to the millions of people who experience a psychiatric condition. Neuropsychiatric conditions account for almost 30 percent of the global burden of disease in the affluent democracies, and as data collection improves, a growing percentage worldwide (WHO 2011). Despite these high needs, it can be difficult to finance, deliver, and sustain comprehensive and ample psychiatric services.

Often, in the minds of both the public and numerous scholarly observers, the challenges of mental health care delivery are the result of cultural, not political or economic, factors. Stigma, in particular, is often blamed (The Lancet 2016). But where does this stigma come from? It is partly the result of psychiatry’s institutional heritage. Popular perceptions of insane asylums equated those buildings with social deviance, insofar as they relegated people with mental illness to a life of permanent seclusion. Eventually, so the story goes, a changing society opted to embrace the mentally ill, cease their confinement, and close the asylums. Yet the public could not completely decouple mental illness from its historical stigmatization, whose cultural legacy continues to plague policy today.

In one of the only political-economic studies of post-asylum mental health care, Andrew Scull (1984) questions this cultural narrative by tracing the structural development of psychiatric deinstitutionalization. Studying the examples of the United States and England, he demonstrates how the economic prosperity of the postwar period reduced the population of long-term residents in psychiatric hospitals, in part because the expanding welfare state created new social programs that made life outside the hospital more feasible for people with mental illness than it had been in the past. Rather than seek custodial and medical care in charitable institutions, patients could rely on increased access to disability benefits, health insurance, and public housing for support. What ultimately closed the asylums, Scull argues, was the advent of economic crisis in the 1970s. The welfare state had to make choices about which programs to keep and which to cut. The expensive mental institutions, whose resident patients had largely left, were first on the chopping block. Although mental health reformers had hoped to replace these hospitals with outpatient and non-hospital alternatives, these countries lacked the political and financial will to do so. The result was a failed policy transition, as well as the continued relegation of the mentally ill to the unattended, stigmatized corners of society.

Scull’s compelling analysis has contributed to the centrality of the Anglo-American, and especially the US, experience in the structural interpretations of psychiatric deinstitutionalization (see also Brown 1985; Dear and Wolch 1987; Jones 1993; Rochefort 1993). The term “de-institutionalization” has in fact become the anglicism imported by non-English speaking scholars to describe that process in their home countries, even though alternative language might be more context-appropriate (e.g., the French “déshopitalisation,” Coldefy 2010; see also Chapireau 2021). The international literature often presumes that this process will result in institutional closures, as was the case with American state and county mental hospitals. It also assumes that countries will struggle to provide outpatient and non-hospital services, as was the case with the US community mental health center program (WHO 2014a).

Ironically, comparative and historical political-economic research was foundational to the study of the welfare state even while mental health has thus far been excluded from this work. Perhaps that Anglo-American focus is one of the reasons why. Not only has much welfare state research focused on the (non-Anglo) European cases, the existing (US-inflected) interpretation of psychiatric deinstitutionalization in many ways conforms with how scholars of the American politics subfield understand the development of US social policy. While at one point the United States did provide some substantial welfare benefits, they soon became associated with unequal 19th-century systems of patronage, perceived as a threat to the country’s racialized economy and geography, or otherwise encountered too many political hurdles in the country’s veto-ridden federal, congressional, presidential, and judicial system (Lieberman 1998; Mettler 1998; Skocpol 1992; Weir et al. 1988). Asylums, in short, were bound to close in the United States. Even if they did make it past the 19th and early 20th centuries, these hospitals would face steep opposition from the politicians attempting to retrench and/or privatize social benefits in the late 20th century (Hacker 2002; King 1987; Pierson 1994). More generally, the United States’ limited welfare state prefers to distribute benefits to (white) recipients whose hard work has rendered them “deserving” (Gilens 1999; Quadagno 1994), a designation the destitute mentally ill could not hope to achieve.

For comparative political economists, the presumed lack of financial importance of the mental health care sector has put it in a secondary position relative to the big-budget items of the welfare state, such as pensions, education, and general health care. Nor is mental health care clearly tied to the labor market and its needs, precisely the area that the welfare state seeks to redress. Moreover, the Anglo-American experience of psychiatric deinstitutionalization also conforms with how scholars of comparative politics understand the national types of social policy (Goodwin 1997). That the United States and England were unable to maintain and expand mental health care after deinstitutionalization may not be very surprising, given that these countries belong to the “liberal” category of welfare states (Esping-Andersen 1990) and the “liberal” variety of capitalism (Hall and Soskice 2001). In these contexts, the state is much less likely to intervene in the market and provide social benefits to workers than it is in Scandinavia and continental Europe, the home of the “social democratic” and “corporatist” categories of welfare states and the “coordinated” variety of capitalism.

And yet some countries do devote significant public resources to the nonworking mentally ill. They also manage to provide extensive outpatient – and inpatient – care. These cross-national variations, furthermore, do not map onto the typologies of welfare states or capitalism that anchor comparative political economy. Figure 1.1 plots the supply of psychiatric care across 16 affluent democracies. These countries were the first to experience deinstitutionalization, since their early industrialization in the 19th century prompted the rise of the asylum, while their economic prosperity in the 20th century prompted its decline. Contrary to the presumed logic of deinstitutionalization described above, the countries that maintained more hospital care (or “institutional” care, which Figure 1.1 measures in terms of psychiatric beds) also tended to expand more non-hospital care (or “community” care, which Figure 1.1 measures in terms of outpatient and nonresidential facilities). Those countries also tend to spend more public resources on mental health care. Although the anglophone cases tend to fall at the lower end of the supply spectrum, they are not alone there. Social democratic and corporatist welfare states, such as Sweden, Denmark, Italy, and Austria, also provide low levels of inpatient and outpatient mental health care. Meanwhile, other countries with coordinated market economies provide starkly different distributions of care. In short, the supply of public mental health services varies across the affluent democracies, in ways not expected by either mental health scholars or comparative political economists.

Figure 1.1 Scatterplot of psychiatric beds and community care facilities per 100,000 in 16 high-income democracies, with percentage of health budget allocated to mental health (as available) and line of best fit

Source: WHO (2011)

These variations do not map onto the expected national typologies of welfare states and capitalisms in part because those models presumed a very different economy, one based on industrial and, especially, manufacturing labor. Although scholars have observed that these models may not fully represent the approaches of contemporary governments to the postindustrial, liberalized, service-based, and knowledge economy (e.g., Baccaro and Howell 2017; Pierson 2001), less has been made of the fact that services now occupy a staggering segment of the welfare state itself. Models of welfare states center on the primary role of social transfers – such as pensions, unemployment insurance, and health insurance – in compensating formal, industrial, and largely male workers. Heroic efforts have certainly been made to characterize the scale and scope of the rising social service sectors in the welfare state, areas that include the development of secondary and tertiary educational institutions, advanced and more complex health care services, and long-term care services for children, the elderly, and the disabled (Alber 1995; Castles 2009; Jensen 2008, 2011). Since scholars must develop specialized expertise to understand each of these social service sectors, they are often studied in isolation.⁷ Attempts to examine them together or to generalize across sectors are less common.

Much can be gained by analyzing the comparative politics of social service provision in the postindustrial economy as a comprehensive trend. In one important study, Making Markets in the Welfare State (2011), Jane Gingrich examines why policy-makers have recently adopted different approaches to marketizing social services, finding that the interests of service providers play a crucial role in the outcomes chosen by politicians. In cases of large-scale or universal social services, such as the entire education or health care sector, Gingrich argues that the preferences of a broad client group can motivate politicians to introduce market mechanisms to improve service quality, delivery, or distribution. Whether and how politicians act, though, depends both on their partisan orientation and on how the preexisting structure of the social service shapes provider interests. According to Gingrich, the interaction of partisan and provider preferences, in turn, shapes the type of market reform pursued.

Building on where Gingrich left off, this study develops the political economy that sustains services for the disenfranchised, a group of pressing concern in an era of sharp political, social, and economic inequalities. In fact, even the studies of social services in developing countries can presume some client demand, insofar as politicians deploy those services as instruments of patronage and vote-buying (e.g., Auerbach 2019; Stokes 2013; Thachil 2014; Wilfahrt 2022). But unlike the mobilized constituencies invested in protecting large-scale and universal social services, services such as mental health care, elderly care homes, or schooling for poor children lack this political pressure.⁸ The same issue arises for social services delivered to noncitizens, such as undocumented immigrants, or in democratically unincorporated regions, such as the overseas territories of many of the affluent democracies.

Mental health care, then, is a window into the political economy of social service provision to disenfranchised and destitute populations, who lack the electoral power to motivate politicians to reform or expand those services. Although mental health has recently entered popular conversations about personal well-being, sustained mobilization tends to be limited to the efforts of a few families. This is a political constituency that is often geographically dispersed (and less compelling to elected officials), usually middle class or affluent (and able to share some of the costs of care with government), and primarily concerned with intellectual and developmental conditions (not those of poorer adults with severe and chronic illness). These important efforts notwithstanding, beneficiary advocacy is much less common in mental health than in some other social policy areas, and with good reason.

Consider the experience of schizophrenia. Prevalence rates cut across national boundaries in ways that are broadly comparable (Charlson et al. 2018). This severe mental illness can inhibit the basic activities of daily life, such as doing household chores or paying the bills, and so the act of political organization becomes even less likely. Moreover, the condition tends to present itself in early adulthood, often after completing compulsory education but before obtaining occupational security. This unfortunate timing reduces the incentives for either the educational system or the labor market to address the condition. Nor does this timing enable people with schizophrenia to access work-based social benefits or accrue the necessary disposable income to demand services with the help of their unions, wallets, or, for that matter, political parties. Where schizophrenia and other forms of serious mental illness contain socially constructed components, furthermore, they do so in ways that align with a society’s existing axes of marginality and political impoverishment (see, for example, Metzl 2009). In exacerbating these vulnerabilities, mental illness is an extreme example of weak demand, thereby shifting and focusing political analysis squarely on the supply side.

Beyond its ability to explain the distribution of welfare to the disenfranchised, a study of mental health care can also inform studies of public services more generally for two reasons. First is its “public” component; that is, its heavy dependence on government financing. For the reasons discussed above, very few clients can afford to pay for costly, labor-intensive psychiatric services on their own. Consider how the unadjusted costs of psychotherapy could amount to a patient’s monthly rent or more. If a licensed professional therapist earns $100–$200 per hour of treatment, weekly meetings (the usual clinical recommendation) would amount to $400–$800. Government subsidies of those labor costs hence reduce those high out-of-pocket costs for clients.

Mental health services that cannot rely on public funding reduce their size and scope, relying instead on patients with resources to sustain less-intensive services. The provision of mental health care, as a result, can be especially unequal, since recipients are either unable to afford costly long-term treatments (and therefore depend on public generosity) or are sufficiently moneyed to cover the substantial out-of-pocket health and social care costs associated with mental illness (see Perera 2019). Providers tend to serve one clientele or the other, with higher needs among the former group. The lines between public and private care have become increasingly blurred, particularly as welfare states continue to delegate provision to the private market and introduce market reforms to public services and policies (Gingrich 2011; Mettler 2011; Morgan and Campbell 2011); but even services that delegate mental health provision to private actors still tend to rely on public financing to support it.

Second is the “services” component of mental health care. In addition to depending heavily on public financing, mental health care includes components from all three core social services: education, health, and care. Since its inception in the 19th century, mental health care has stood out for its combination of medical and custodial care, financed, if not outright delivered, by government (see Ansell and Lindvall 2020). Over time, educational and vocational services have become part of the equation too, as society began to perceive the mentally ill as both capable, even desirous, of employment and schooling. Today, mental health care can involve a wide range of professionals, such as doctors, nurses, therapists, social workers, caregivers, vocational instructors, and special needs teachers, as well as the administrative and maintenance staff required to operate a variety of facilities.

Like other social services, mental health care has experienced significant pressures to reduce its labor-intensive costs. These strains became acute during the deinstitutionalization period, as Scull (1984) observed. The two most threatening factors to peak during this time were those that beset all services: automation and cost-control. When the first anti-psychotic medication hit medical markets in the 1950s, it threatened to replace much of the work conducted by the staff of mental health institutions. As an automatic sedative, it reduced the amount of time that staff spent constraining patients who were experiencing difficult episodes, as well as that spent engaging them in other activities to avoid relapses. By the time the economic crisis hit in the 1970s, governments (urged by pharmaceutical companies) had realized that patients could take these medications on their own, outside of the expensive hospital system and with the help of the expanded “outdoor relief” of the postwar welfare state. The state pressure to reduce costs and close down hospitals became a ubiquitous force across the affluent democracies. Moreover, these pressures increased over the subsequent decades, as attempts to privatize, marketize, or otherwise retrench public services accelerated. The fact that some countries retained and even expanded their supply of public mental health care suggests that politics can help to explain why some services resist these pressures.

In sum, a study of the comparative political economy of mental health, in historical perspective, is instructive for several reasons. First, mental health specialists lack an explanation for why the supply of mental health care varies across countries, which is attributable neither to cultural factors (as a political economy perspective has already shown) or to a single, unifying set of conditions (to date, the dominant way of understanding the structural dynamics of mental health care provision). Second, the cross-national differences in supply suggest that sometimes welfare states deliver social services to clients who are unable to demand them. These differences bring to light a puzzle whose solution has become even more necessary in recent decades, as inequality increases and services (especially public social services) expand in the affluent democracies. Not only is demand for mental health care especially weak, the challenges in its provision are emblematic of all public services. Despite their heavy dependence on government financing and significant cost pressures, public mental health services have managed to survive, and even thrive, in some countries. The theoretical contribution of this book is to explain why, as detailed in the remainder of this chapter.

The Welfare Workforce and Supply-Side Policy Feedback: A Theory of Alliances

The politics of the “welfare workforce” can explain why the supply of mental health services diverges across countries and, indeed more generally, why the supply of public social services for the disenfranchised often varies as well. When services lack robust demand, their maintenance and expansion can hang on those who depend on them for their employment. Public service employees have become a significant segment of the labor force in affluent democracies, on average one of every five workers is now employed by the state, most often in the health, education, and care sectors (OECD 2021). As employees of the welfare state, these workers have a strong interest in maintaining and expanding its public social service infrastructure. Otherwise they might lose their jobs, which are among the best compensated and protected in postindustrial, and increasingly precarious, economies.

Although the welfare workforce can play a role across the gamut of social services, it is especially influential for services with weak client demand. As Gingrich (2011) has demonstrated, the distribution of social services with broad, universal support depends on the relationships between constituents, partisan politicians, and providers. My research builds on that work, exploring the provision of social services that lack such electoral support. Without mobilized beneficiaries and, by extension, the politicians that represent them, providers become the key political stakeholder. But providers do not always succeed in maintaining and expanding the welfare state.

Figure 1.2 illustrates the theory of “supply-side policy feedback,” which I have developed to explain when, why, and how welfare state employees can alter the course of social policy. This term draws on the scholarly literature in American political science on “policy feedback,” which demonstrates how the structure of public policy conditions elite and public demand for it, in turn reinforcing that structure (Campbell 2003; Mettler and Soss 2004; Patashnik 2008; Weir and Skocpol 1985).⁹ The politics of services for the disenfranchised, though, can depend more on the politics of supply than those of demand. In supply-side policy feedback, the structure of public policy conditions suppliers’ demand for it, producing a similar, self-reinforcing pattern. Unlike theories of policy feedback that focus on elite and mass politics, this theory emphasizes the meso-level: the role of workers who depend on those policies for their employment (the welfare workforce).

Figure 1.2 Supply-side policy feedback model: Effects of public sector worker alliances on the supply of public social services for disenfranchised populations (basic diagram of theoretical argument)

The supply-side policy feedback process begins when services face sharp cost-containment pressure, pushing the wages of employees downward. Cost pressures are formidable and frequent in these service sectors. For example, related or automated services can become serious competitors. Nor are the growing service sectors able to compete with wage increases in the smaller but more productive sectors, such as manufacturing (Baumol and Bowen 1966). Public services are also subject to political deregulation and government cost-containment initiatives.¹⁰ Each of these factors incentivize service providers to reduce costs, especially in their most expensive budget line: labor and wages.

In response to these cost pressures and downward wage pressures, the employees of these services mobilize, so long as public policies, laws, and regulations permit them to do so. Where organized labor can strike, bargain, and advocate for its interests, workers can mobilize to preserve their employment, resist layoffs, and advocate for wage increases. Institutional context, whether national or occupational, is important. Compare, for instance, teachers and caregivers. Differences in the organizational rights and capacities afforded to each group, as well as the public or private sector location of their employment, have granted them very different degrees of political voice. While protests by public sector teachers’ unions have become a fixture of the postindustrial economy, the same cannot be said for private caregivers. Whether and how workers mobilize, then, is the second step of the supply-side policy feedback process.

The third step can be the most consequential for the fate of public services. Here workers can develop the political alliances necessary to procure protection and compensation. Sometimes, politicians can become important allies of public service employees (Anzia and Moe 2016). A particular district’s or party’s dependence on public employment, for instance, can incentivize elected officials to respond to their demands. But note that this theory departs from others developed on public sector employees by focusing on their role as workers, not voters. In cases where clients are disenfranchised and the electoral relevance of the policy area is low, public sector trade unions seeking to protect the employment of their members might turn to actors other than politicians for support, such as interest groups, professional associations, and even sympathetic social movements.

Of particular importance to this theory, therefore, is the presence or absence of a political alliance between public service managers and public service workers. To be sure, it can be difficult to define “managers” and “workers” in the public sector. A distal and complex principal–agent relationship characterizes public sector labor relations, especially given the role of elected officials as potential principals (Moe 2006). State-employed managers, this theory underscores, are another set of principals. I hence define public managers as those government-employed supervisors whose positions do not depend on electoral politics and political appointments. Public managers nonetheless may make decisions about hiring and firing, workplace protections, and production capacities (like their counterparts in the private sector); but they do not “own” the means of production (unlike those counterparts). Rather, public managers remain dependent on government funds and employment contracts for their wages. They can include the heads of government schools, the directors of public hospitals, or the administrators of public childcare services. Public workers are their supervisees, who also depend on government funds and employment contracts for their wages.¹¹

Unlike labor–management relations in the private sector, in the public sector these relations are not necessarily antagonistic. In the private sector, managers are often responsible for maximizing revenue while minimizing labor costs, incentives that tend to put them at odds with rank-and-file workers, who would, on average, prefer higher compensation and strong job protections. But because many of the incentives private sector managers face are often relaxed in the public sector – managers, for example, are rarely concerned with profit margins – managers and workers are often much less at odds. For example, managers’ employment contracts and benefit packages are often very close, if not identical, to those of their own employees. Perhaps more importantly, the two ultimately have a shared interest in protecting their source of employment, especially when proposals that introduce high levels of insecurity, such as those to cut funding or privatize the existing institution, emerge. Cross-class coalitions can therefore be more common in the public sector than the private sector because workers and managers have stronger incentives to align.

When the state threatens to reduce the supply of public services, employees mobilize using the usual tools of labor influence, such as strike action. In doing so, they generate pressure that urges managers to respond. One way to raise funds is to charge clients more for services, but this approach is unreliable if clients are poor. Instead, and following Niskanen’s (1971) classic formulation, public managers can become “budget maximizers” who can turn to the state to obtain secure and generous compensation for public services. Departing from Niskanen, though, I propose that what motivates managers to do so is the strong pressure of their employees, for labor costs are often the largest line item of most public budgets.¹² Public managers are unlikely to advocate without this employee pressure, especially since managers tend to have more opportunities to exit to private industry when their public sector careers are no longer viable.

This alliance shapes eventual policy choices through two main mechanisms. The first is “brokerage” – that is, when an individual or group connects previously disconnected social sites.¹³ Acting as brokers between the public sector workforce and policy-makers, public sector managers can draw on the various tools at their disposal to shape public policy choices. Furthermore, the complex principal–agent relationships that characterize the public sector mean that “managers” can work at multiple levels of government and, by extension, control a broad range of rent-extracting tools. These efforts can reap considerable rewards for the coalition and the protection of public employment as a whole.

The second mechanism works through the “adaptive expectations” of policy-makers.¹⁴ Simply put, policy-makers agree to increase funds in order to avoid escalating retribution from the powerful coalition. These sanctions can take multiple forms. For example, sustained labor unrest against cutbacks may pose a significant public disturbance, for which policy-makers may want to avoid any associated blame. Or politicians might worry that public sector workers will penalize them at the ballot box in their next term (as documented, for instance, by Anzia and Moe 2016). Electoral considerations can reenter the distributive equation here; but they might not. Exacerbating these potential penalties are those techniques deployable by public managers. Stifling the progress of an important policy-making commission or refusing to administer a public service, for instance, can cause headaches for both elected and unelected officials. Policy-makers hence retract attempts to cut back on public funding after realizing the political costs of doing so.

Many factors can influence whether the public labor–management coalition can form, though in this book I primarily focus on one: the organizational structure of public managers. As I have argued elsewhere, how a group aggregates the interests of its members can shape its public policy agenda (Perera 2022).¹⁵ Where a united group of public sector managers can express their interests independently of private sector managers, they are more likely to form a coalition with public sector workers. Organizations that independently represent public managers have clear economic priorities, advocating only on behalf of the public sector. Organizations that represent both public and private managers, however, must contend with the competing positions of each camp. An organization with a mixed attitude hence is unlikely to become a strong promoter of public services. Moreover, it is important that public managers speak with a single voice. A singular focus on the public sector also makes public managers facing a hostile bargaining environment both more willing and more effective, especially when urged to do so by employee allies. It also encourages cohesion between different types of public managers (e.g., at varying levels of government or service areas). Where the political representation of public managers is fragmented, they are likely to express differing policy agendas. Those conflicts can make it difficult to solidify an alliance with workers. Public managers who organize together and independently of their private sector counterparts, therefore, are better equipped to satisfy workers’ demands to raise revenues, form a political coalition, and help to increase the supply of public services for the beneficiaries who cannot demand them.

This public labor–management coalition is therefore a unique source of political power for state employees and can have long-lasting effects on social service provision. When stable and regular, these otherwise strange bedfellows can procure generous revenues and protections from government, reinforcing the political strength of the welfare workforce, spurring additional rounds of policy feedback, and expanding public social services. The absence of a durable coalition has the opposite effect: Without the political alliance with management, public sector workers can have difficulty procuring support from government, which results in less revenue, fewer protections, and the retrenchment of public services. These patterns produce what policy feedback scholars call “positive” and “negative” effects, respectively (see Jacobs and Weaver 2015; Pierson 2000b; Weaver 2010).

These self-reinforcing feedback effects notwithstanding, the public labor–management coalition holds independent influence on policy outcomes. Without it, the initial cycle cannot launch. Although institutional factors may influence the propensity of the welfare workforce to form alliances, they do not necessarily determine them (nor do they independently predict the policy outcome, an alternative hypothesis discussed in Chapter 2). Although state actors may attempt to manipulate alliance formation, they are more likely to do so in subsequent feedback cycles, and usually only once the coalition has demonstrated its potency. Reversing the direction of feedback is nonetheless possible, particularly if the coalition is not yet very powerful nor very weak. More generally, the presence or absence of a public labor–management coalition prompts alternative path-dependent processes in which the alliance gradually gains or loses political power over time. These two supply-side policy feedback cycles have shaped the trajectory of mental health service provision across the West.

Comparing Mental Health Care over Time and across Countries

Explaining why the supply of mental health care varies across countries requires analyzing the policies that structure those differences and how they developed over time. Structural outcomes such as public policies have multiple and complex causes. To isolate the most salient and determinative factors for mental health policy, I compare the late 20th-century process of psychiatric deinstitutionalization in two countries with similar initial conditions but different contemporary outcomes: the United States and France. To assess whether these factors can explain variation elsewhere, I then trace them in Sweden and Norway, a concise comparison that also allows me to rule out a few unresolved alternative hypotheses from the first comparison.

Explaining highly aggregated outcomes such as mental health policies requires thorough attention to both their individual parts and their systematic whole. The analytic approach of this book is therefore comparative, historical, and political-economic, examining how the relationship between politics and markets evolves over time and across governments. It takes into account how economic and political incentives combine to shape structural outcomes, such as those of an entire public policy area or economic sector. A historical perspective allows exploring the complete range of potential causal factors, including the chain of events that shaped their evolution. A comparative perspective, in turn, helps to narrow the list of potential causes by evaluating patterns in their presence or absence across cases. In this way, a cause that may seem important to one case is subject to testing and confirmation in another. This approach hence guides the analysis toward identifying the most important factors driving complex outcomes in contemporary political economies.

To help identify cases for comparison, the study begins by systematically charting the variation in public mental health care provision across affluent democracies and over time (this is developed in more detail in Chapter 2). To date, most of the research on mental health care provision has tended to examine individual countries, without a clear sense of how it compares to that of other countries (for an exception, see Goodwin 1997). Much of this predicament is due to the inherent difficulty of comparing indicators across national contexts, as they are often collected and defined in very different ways. A recent data collection initiative of the World Health Organization, however, aims to redress this problem. I use this standardized data to analyze contemporary differences in mental health care supply in the affluent democracies, the countries that first pursued deinstitutionalization reforms and set global expectations for public policy in this area (Figure 1.1). I also explore whether and how the historical evolution of mental health care supply may have shaped these differences, using the novel data set described in the Appendix. That effort reviewed the national statistical yearbooks of the 16 countries where deinstitutionalization first took root, collected data on mental health care services over the full course of deinstitutionalization, as well as tracked the initial conditions of supply, how it changed during that international wave, and with what results for 21st-century mental health systems (1935–2000). The data shows that all countries experienced deinstitutionalization but to widely varying degrees.

Drawing on the guidelines for case selection in comparative politics, I select two major countries where deinstitutionalization produced contrasting outcomes, despite the presence of important similarities prior to its onset. For this purpose, I select the United States, as noted the most influential case in the scholarly literature on deinstitutionalization, and France, of which relatively little has been written (especially in English). The two countries shared similar mental health and social care systems prior to deinstitutionalization. Policy-makers in both countries adopted similar blueprints to reform mental health care in the postwar period, notably with better prospects for implementation in the United States than in France. But the two systems eventually diverged.¹⁶ I trace three supply-side policy feedback loops in each case, from the 1960s to the 1980s, documenting how the presence or absence of cross-class coalitions in the public sector shaped the increase of services in France and their decrease in the United States.

Although this paired historical comparison is able to eliminate a wide range of general and case-specific alternative hypotheses, it is nonetheless important to assess the validity of these findings in other countries. Because of the complex and context-specific knowledge required to analyze mental health care systems and their development, I have undertaken a second paired comparison of two smaller countries (instead of a “large-N” statistical analysis of far more governments, where this complexity could be lost). For this purpose, I selected Sweden and Norway, two expansive Nordic welfare states with uncharacteristically different approaches to mental health care. As Ansell and Lindvall (2020) note, Sweden, like the United States, had high rates of psychiatric institutionalization in the 1930s (over 0.3 percent, Ansell and Lindvall 2020), but rates fell dramatically later in the 20th century. This trend stands in contrast to what occurred in France and Norway. Moreover, and mirroring the contrasting patterns in the United States and France, Sweden’s lower mental health care supply has received more attention in the international literature on deinstitutionalization than Norway’s much higher supply. I trace one supply-side policy feedback loop in each case, since their divergence occurred primarily in the last decade of the 20th century (when economic crisis prompted their governments to speed up deinstitutionalization, which like that of its counterparts in the United States and France had already begun during the postwar period of welfare state expansion).

Studying these four cases in comparative, historical, and political-economic perspective is not only substantively instructive; it also departs from conventional approaches to political science research. First, this perspective is rare in the study of American public policy and politics. The organization and development of the political science discipline has foreclosed opportunities to apply this lens to the United States, instead fostering research on the country’s peculiar political institutions and the opinions and behavior of its citizens. These avenues have generated important contributions to US politics, and indeed (if unintentionally) elsewhere, where Americanist research is often influential. Much can be gained, however, by incorporating a comparative and historical political-economic lens, insofar as it can help identify possible oversights in the field to date. Pioneer research by a few Americanists – notably those who have tended to write on the intersection of social policy and American political development – have already demonstrated the utility of comparative and historical political-economic analysis for the study of the United States (e.g., King 1995; Pierson 1994; Sheingate 2001; Weir and Skocpol 1985). The study of US health politics, in particular, has benefited from a tradition of comparative research (e.g., Giaimo 2002; Gusmano et al. 2010; Jacobs 1993; Maioni 1998; Tuohy 2018). Such work has been foundational to the emerging sub-disciplinary field of American political economy (Hacker et al. 2021). This book builds on and advances this under-utilized lens in American politics research.

Incorporating non-European countries in this study (through a close case analysis of the United States and the inclusion of Australia, Canada, and New Zealand in quantitative, cross-national comparisons) also expands the conventional comparative-historical scholarship on social policy and political economy, where the experience of Western Europe has been the dominant focus. Although the landmark typologies of welfare states and capitalist varieties include the affluent democracies of Northern America, East Asia, and Oceania (Esping-Andersen 1990; Hall and Soskice 2001), only a few scholars have devoted substantial attention to those countries (e.g., Estévez-Abe 2008; Morgan 2006; Obinger et al. 2005; Pontusson 2005; Swenson 2002). This attention deficit is the result of both the regionalist structure of the political science discipline and the fact that most of the cases of interest are located in a single region. Here, too, an expanded case comparison can help to identify oversights. The political-economic development of the non-European cases often occurred later and under very different conditions (e.g., settler colonialism, internal racial division, Eastern business models) compared to Europe. Moreover, the peculiarities of the French case do not fit neatly into the existing typologies either (Schmidt 2003). France’s presidential and republican (anti-monarchical) history, for example, breaks from the political patterns found in the parliamentary constitutional monarchies of its neighbors. Rich insights can emerge, therefore, from this French–American comparison, an unconventional one in political science (though for some examples of books on the politics of social policy that use this approach, see Morgan 2006; Toloudis 2012). Moreover, the inclusion of Sweden and Norway, whose levels of mental health care supply are respectively much lower and much higher than scholars might expect, offers a more nuanced take than the stylized facts about these countries.

The subsequent chapters present how supply-side policy feedback and public labor–management coalitions developed mental health policy, first with a detailed French–American comparison and then with a more concise Norwegian–Swedish comparison. As Chapter 3 describes, in the 19th century public managers in both the United States and France organized independently, but over time public managers in America allied with their private sector counterparts (in this case, academic neurologists and psychiatrists in private clinics) while French public managers maintained their separation and unity. This seemingly small organizational difference nonetheless had major implications for public mental health care workers during the period of psychiatric deinstitutionalization. American public sector workers had difficulty gaining the support of their managers, whose representatives privileged attention to private psychiatry, while in France the coalition between public workers and managers resisted attempts to retrench services during this time. Chapters 4 and 5 discuss the American and French postwar and late 20th-century pathways to deinstitutionalization, respectively, assessing relevant case-specific alternative explanations along the way, such as the role of the centralized state and generous welfare coverage in France or the role of federal-state politics and racialized, more limited, and privatized welfare coverage in the United States. Chapter 6 then assesses the validity of these findings in other countries by corroborating them in the abbreviated study of Norway and Sweden in the 1990s, ruling out the major remaining alternative hypotheses such as the general strength of the public sector trade union movement or the influence of a state-oriented welfare state. None of these alternative hypotheses hold significant muster against the central hypothesis of this book: that an alliance between labor and management in the public sector, enabled by the independent and unified organization of public sector managers, granted the welfare workforce political influence over the maintenance and expansion of the service that employed them.

The four cases also work together to assess the argument in other ways. First, they allow me to leverage a range of historical moments to assess whether public mental health policy depends more on the support of public workers, public managers, or indeed their combined potency. Table 1.1 presents examples of what occurs under each of these scenarios, drawing on examples from throughout this book. This study finds that neither public workers alone nor public managers alone can resist retrenchment efforts. Second, the four cases can also test whether the argument travels across the main types of western welfare states and varieties of capitalism: the “liberal” anglophone countries of Europe, North America, and Oceania (exemplified here by the United States), as well as in the “coordinated” market economies of “conservative” continental Western Europe (here, France) and “social democratic” Nordic Europe (Norway and Sweden) (see Esping-Andersen 1990; Hall and Soskice 2001).

Table 1.1 Independent and combined effects of the presence or absence of a public labor–management coalition on the expansion of public mental health services in four countries over time

		Public managers join coalition
		Yes	No
Public workers join coalition	Yes	France (post-1968), Norway (1990s): significant expansion	United States (post-1970s): no significant expansion
	No	France (pre-1968), United States (pre-1920s): no significant expansion	United States (1920s–1970s), Sweden (1990s, most reforms): no significant expansion

The empirical material for this study relies on a wide range of primary and secondary historical sources (the documents relevant to mental health policy-making prior to and during psychiatric deinstitutionalization), few of which have received much academic scrutiny to date. Prior to this study, many of the sources on mental health policy produced by professional and, especially, labor organizations (such as trade press, newsletters, and memoirs written by their leaders) had been under-utilized and under-analyzed. For example, Chapter 4 draws on the American Federation of State, County, and Municipal Workers’ three expansive collections on psychiatric deinstitutionalization (45 linear feet, per RL PPAD),¹⁷ an important source hitherto unexplored, to the best of my knowledge, by other scholars. In fact, across all four countries, extensive archival research brings a wealth of reports, administrative memoranda, meeting minutes, and newspaper clippings to scholarly attention. Bolstering these primary historical sources are out-of-print secondary sources consulted at national, industry, and medical research libraries. For supplementary information and when possible, I conducted interviews with key policy-makers and political actors. This book is distinctive, moreover, in that it discusses the material of several countries in English; to date very little has been written about the non-English speaking countries in English (especially France and Norway). The primary sources consulted are listed in the Bibliography and cited at a more detailed level in the chapters.

New sources bring new information to the study of mental health policy. In fact, they often draw out peripheral points made in other histories and bring them to the center of the analysis, producing an account that is historiographically distinct from others. Of particular note is this project’s treatment of elite opinions. While Whiggish histories have tended to emphasize the idealistic intentions of mental health policy reformers (e.g., Barton 1987), critical histories, instead, note how medical superintendents and other powerful actors deployed mental health care and its institutions as instruments of social control (e.g., Rothman 1980).

I examine written documentation linking elites to politics to highlight how their political and institutional environment shaped their economic priorities. Although I pay particularly close attention to official organizational statements, following Swenson (2018) I also assess the completeness of these sources’ claims by triangulating them, wherever possible, against those made elsewhere, such as commentary in flagship journals and trade press, testimonies at legislative hearings, and points of discussion in governmental meetings. Readers can assess my interpretations of these sources by consulting the detailed bibliographic information, and often the fully quoted material, in the text. Source triangulation therefore allows me to develop a comprehensive understanding of various actors’ demands, as well as whether and how their organizational representatives conveyed these demands to policy-makers. Taken as a whole, the sources consulted for this book shift attention away from elites’ moral and therapeutic inclinations and toward the political, administrative, economic, and institutional incentives that may shape them.

Moving beyond the contributions and findings of this book, Chapter 7 concludes by advancing its implications for postindustrial welfare capitalism more broadly. Canonical theories explain the development and variation in welfare states by pointing to the role of left politicians and unionized private sector workers (Esping-Andersen 1990; Korpi 1983; Stephens 1979). This book, though, demonstrates that the rise of service employment, especially in the public sector, has shifted both the products and the politics of the welfare state. Today, social service administrators and unionized public service workers have decentered the influence of those conventional actors. A new political logic is motivating social policy. The conclusion explores that new logic, and its consequences for people with mental illnesses and other beneficiaries of social services.