of the two tracks of the welfare state: at any given time during its existence, four hundred thousand African Americans and another four hundred thousand women who would otherwise have been excluded from UI and Old Age Insurance gained eligibility through their WPA paychecks and the payroll contributions deducted from them.23
Direct job creation called upon work as a cultural, social, and political symbol, used to empower as well as to oppress. Social investigations of the kind studied by Mary Furner, Barry Supple, Alice O’Connor, and other students of social scientific knowledge and progressivism make this clear.24 They turned to research conducted in the 1930s by E. White Bakke, the eminent British social scientist and expert William Beveridge,25 the WPA’s Research Division, the U.S. Labor Department, as well as an array of public and private researchers to show that work had an important symbolic meaning for the poor and the working class. In industrial societies, work provided (and still provides) a huge source of social and psychological meaning. It defined the neighborhood one lived in; it circumscribed the friends and workmates who would occupy much of the day; it offered a sense of contributing to society, of being needed.
Given the influence of immigration in industrial labor forces in the United States especially (but to a lesser extent in Europe as well), where a person worked could also be a signifier of nationality, language, and religion, and vice versa, as professions and sectors were ethnically divided. As a consequence, unemployed workers and their families, even those on public relief, believed in the idea that work entitled one to a superior set of earned rights and they wanted access to it. As one wife of a WPA worker put it, “We’re not on relief any more—my husband works for the government!”26
Early public opinion surveys, conducted by Witte and the nascent Gallup organization, of the unemployed and of relief clients discovered that the most downtrodden were much more adamant about finding a job than they were about welfare for all. While scholars favoring social citizenship or universal human rights (or a guaranteed minimum income) might argue that this reflects the hegemony of American individualism, there is no escaping the evidence of millions of unemployed workers themselves who proclaimed, “[W]e don’t want relief, give us work.” The sentiment reflects the conviction shared by many (including Marx) that the creative potential of employment is missing from social citizenship. As Beveridge argued in 1944, “A person who cannot sell his labour is in effect told that he is of no use … a personal catastrophe. The difference remains even if an adequate income is provided, by insurance or otherwise, during unemployment … the feeling of not being wanted demoralizes.”27
In the context of full employment and the right to a job, work requirements lose something of their punitive or exclusionary force. Instead they become a constitutive element in a horizontal social contract between the employed, who agree to pay the taxes required to employ the jobless, who in turn agree to produce public goods and services to enlarge the commonwealth and beautify the public square. On this reading, economic citizenship becomes less a characteristic of laissez-faire systems, and more a characteristic of solidaristic systems—hence the reason job policy is a vibrant characteristic of social democracies.
Viewed through the lens of job policy, the distinction between social and economic citizenship breaks down. This in turn carries significant implications for the broader debate about American exceptionalism and the welfare state. Go/sta Esping-Anderson’s model of Anglo-American, Continental, and Social Democratic welfare state regimes relies on these distinctions.28 Yet if both the United States and Sweden were developing job policy simultaneously in the 1930s at a time when other nations within the Anglo sphere of influence (the United Kingdom, Australia and New Zealand, Canada, and so forth) and Continental nations such as France and Germany were not, we have a dilemma on our hands. How explanatory is the Anglo-American model of welfare states? How exceptional and separate was the United States? The model does not hold. Likewise, if at times Britain, France, and Germany balked at mass direct job creation (even in the midst of an economic crisis like the Great Depression) then the United States’ frustrated attempts to institutionalize job policy mean that the American social policy is not as far as some would have it from European models.
Direct Job Creation and the Welfare State
Social Security occupies a prominent place in policy history. Sociologists and political scientists have been interested in why the United States is a welfare laggard and why the country adopted certain structures of eligibility and their underlying “theories” of worthiness and not others. The American political development (APD) school has made particularly important theoretical contributions to our understanding of the ways institutional changes and structural features shape policy outcomes historically. In this model, initial policy decisions create a political dynamic that acts to reinforce them over time. This “policy feedback effect” entrenches one approach and excludes others, creating a “path dependency” in institutional development.
Theda Skocpol, Margaret Weir, Ann Orloff, Jill Quadagno, Jacob Hacker, and others in this tradition point to Social Security as a classic case of path dependency. Once a policy so controversial that after enactment it was targeted for repeal by the opposition party, it became a third rail of American politics, so strong that it was easily able to weather antistatist periods in American politics. APD studies have tended to focus almost exclusively on social insurance and welfare programs. They tend to treat other policy initiatives considered by the CES as examples of policy options that were ruled out as a consequence of path dependency.
Direct job creation throws a wrench into this model of the origins of Social Security. To begin with, path-dependent studies rarely involve what we might call “dual origins” of public policies. To the extent that direct job creation has been studied by the APD school, it has been through the work of Edwin Amenta, Drew Halfmann, and others, who portrayed this type of job program as part of a policy agenda that was separate from the Social Security Act:
When people think of the origins of American social policy, they usually think of the 1935 Social Security Act…. [T]hese programs were somewhat marginal to New Deal social policy because they dealt with special categories of “unemployable” citizens…. Scholars miss the fact that most Roosevelt Administration policymakers did not see themselves as designing a two-track welfare state; rather, the WPA was a means-tested program that gave relatively high benefits…. These programs constituted an incipient “work and relief” state favoring the unemployed.29
I take issue with this perspective. As I will explain in greater detail, Social Security did in fact include “a commitment to work programs” as well as programs aimed at “unemployable” citizens. This raises the question of whether direct job creation was part of a “work and relief” state or rather a “work-insurance” state; the latter is what I believe the CES intended. Path dependency would seem to have limited utility as an overall approach to the development of the welfare state if we take the CES as the starting point. Social insurance and direct job creation shared very similar origins: they came out of the same committee, served much of the same clientele, used much of the same rhetoric of justification and similar economic theories, and yet, in the long run, met very different fates. This may well point to a more contested and contingent process by which different public policies win political support. Social insurance was, after all, not the most popular element of the Social Security Act—there was a long lag before benefits were paid out, payroll taxes were not popular (especially at the higher levels required for a prefunded Old Age Insurance program), and the program was challenged all the way up to the Supreme Court. And yet, unlike the WPA, which had a much stronger institutional base in many respects from an APD perspective, social insurance lasted much longer.
Shaping Historical Memory
One of the reasons why direct job creation has not generally been celebrated as an achievement of the CES is that a conscious attempt was made to write that type of program out of the historical record. The CES is seen today as the organ that created social insurance and social insurance alone.
In 1936, Witte wrote an account of the development of the Social Security Act that was used for many years as the official in-house history of the Social Security Board.30 While the memoir’s purpose was to serve as an eyewitness