Roughly twenty years after passage of Germany's—the world's first—national system of social insurance, Henry Axel Bueck, the chief executive of its most powerful capitalist organization, the Central Association of German Industrialists, remarked that the accident, health, and pension laws of the 1880s had failed to turn the working class away from the socialist labor movement. Bueck had correctly predicted that Kaiser Wilhelm I and chancellor Otto von Bismarck were in for disappointment in thinking that the laws could “conciliate the masses.” Nevertheless, Bueck supported the legislation then and through the following two decades. Looking back, he said,

Recent research focusing on German capitalists and early social insurance comes to the conclusion that, because of their managerial interests—not a desire to suppress socialism—the leadership of German industry supported the building and maintenance of the world's first welfare state. In the case of industrial accident insurance, industrialists were ahead even of Bismarck. As one leader said, with some exaggeration, “Who was it after all that introduced the whole accident insurance idea? It was we and without us the whole legislation would not have gotten underway!”²

In the building of the Imperial Germany's welfare state, a decline of capitalist power, or shift in the balance of power between capital and labor—however one conceptualizes or measures these things—cannot possibly explain much about its timing and content. By political science's conventional measures of “labor power,” the German working class had very little: Its political party activities and trade unions were outlawed. Only nine socialists had seats in the Reichstag. With the first election following the lapse of the antisocialist laws in 1891, the Social Democratic Party still only received about 25 percent of the vote.

For Sweden, a shift in the balance of power against capitalists also fails to explain the development of the world's most advanced welfare state, which, by the 1960s, clearly overshadowed Germany's. My work on Sweden, partially presented in “Varieties of Capitalist Interests,” demonstrates that the late evolution of compulsory social legislation resonated strongly with the interests of Swedish capitalists, in strong cross-class alignment with those of organized labor, and that reformers acted on knowledge of this fact.⁴

Comparative-historical findings like these—counterintuitive to many American and Swedish experts—calls for critical scrutiny of all theory that attributes welfare state development to politicians' declining regard for capitalist interests and therefore a decline of capitalist power. Jacob Hacker and Paul Pierson, both experts in the study of American social policy, have recently presented just such a theory; Pierson's comparative work also, uncritically I think, incorporates the assumption that growing labor power against capital drives growth phases of welfare state development.⁵

I. POWER AND THE AMERICAN WELFARE STATE

I will begin by repeating Hacker's and Pierson's “central methodological claim,” which is that more historical material on the New Deal “will not settle fundamental questions about causal dynamics without proper attention to basic issues in the assessment and study of power.” They make this rather unobjectionable claim in the context of an assertion that, in my SAPD article, I do no more than “[throw] up a dense new thicket of historical findings.” Now Br'er Fox and Br'er Rabbit might disagree on the value of a dense thicket, but my position would be that we need more of everything, good evidence and good theory. To be sure, Hacker and Pierson would probably agree on this, too.

Where we really differ is in characterizing what I write. With an analysis of the National Recovery Administration, I rely entirely on good secondary sources, and present really nothing new in the way of historical knowledge. What is new is the presentation of the NRA in a theoretical context: to show that with the depression and shifting of the political agenda from the state to the federal level, major employers' interests changed in such a way that they increasingly favored national level regulation, including regulation of social protections. Therefore, I do exactly what Hacker and Pierson advocate, even as they criticize me for not doing it: examine “fundamental questions about causal dynamics” with a sophisticated analysis of power as well as interests.

In fact, my argument about both the political fallout of the Depression and the shift of the institutional locus of social politics to the federal level is, by far, the most important one challenging Hacker's and Pierson's theory. Why? Because it focuses on the same two factors on which they build their own argument. By zeroing in on the two pillars of their argument, I hoped to shatter any apparent certainty they try to generate about demonstrating power shifts and resulting legislative change without first empirically investigating continuity or shifts in interests. Surely, at least, this challenge merited a response, in that it confronts head-on a “basic issue in the assessment and study of power.” Curiously, they beat a wide path around this briar patch.

On this question of interests, Hacker and Pierson ask at one point what basis I have for my “confidently expressed thesis about the intense desire of New Dealers to accommodate large employers' anticipated interests.” The answer seems obvious, but I am glad to answer. Money is one basis—I make clear reference to it in my theoretical discussion about regulation and rent seeking. Democrats, then as now, heavily depended on contributions from corporations and their well-paid executives. The New Dealers understood that angering the vast majority of the politically most important capitalists with radical legislation would redirect the flow of money in favor of the Republicans. Hence, no doubt, Edwin Witte's wish to avoid “trouble hereafter.”

I want to repeat these arguments because Hacker and Pierson attach great importance to the “massive political victories of Democrats” in the Depression-era 1930s, which apparently “shifted the locus of social policymaking to the federal level,” thus bringing a “massive decline of business power.” On that note, however, I have to question whether the big 1930s shift in the composition in the U.S. Congress and presidency to the advantage of the Democratic Party, highly dependent as it was on capitalist money, really could have dealt capitalists a severe blow to their power. Recall also that roughly 75 percent of Republicans in the Senate and House voted for versions of the social security bill sent to conference. If Democrats had not been dependent on capitalist money, and Republicans had resoundingly voted against the Social Security Act (SSA), then Hacker and Pierson would have had something solid to work with.

Thus, as soon as we think about the role of money in American politics, we have to be very clear and precise, and perhaps get into a thick empirical thicket or two, before drawing strong conclusions about power shifting between capital and labor. Because Hacker and Pierson bring up electoral politics and legislative shifts, it is their task to guide us analytically through the complex historical evidence about the New Dealers' dependence on capitalist money, and therefore capitalist power in the electoral process.

Finally, to clear up any misunderstanding, I am wholeheartedly in favor of power analysis; I am against, however, what has passed as such for a long time in the comparative literature on the welfare state. I actually help reintroduce the issue of class power (following William Domhoff, but in a distinct way), now almost hegemonically crowded out by Theda Skocpol's brand of institutionalism from the study of the American welfare state. Power analysis should pay close, theoretically informed attention to changing interests in order to make supportable statements about how the “shifting balance of class power” shapes the welfare state. Still, I remain very skeptical if any such statements are possible. The study of cross-class alliances is a far more valid form of power analysis than one treating classes as homogeneous entities. In any case, power analysis should be comparative, especially if it is dealing with gross and abstract categories “labor” and “capital.” At this point, I remain entirely doubtful about the possibility that a shifting balance of power between capital and labor can explain anything about welfare state development only in America (or only in interwar America), but not in other countries and times.

II. STRATEGY AND DETAILS

Hacker and Pierson make a number of contentions in their rejoinder that are worthy of comment, though space allowed will not accommodate all of them. Some have to do with interpretations, some with use of historical facts, in the analysis of reform strategy and legislative outcomes.

Strategic Context

In my article, I commend Hacker and Pierson for their insistence on analyzing strategic contexts to draw better conclusions about power and interests from historical evidence. But I show that they are too narrow in their application of the insight. Here, I already mentioned one strategic context they ignore—capitalist money in partisan politics. I need not repeat the others.

So I confess to be at a loss in trying to understand why Hacker and Pierson now state that I give insufficient attention to “the strategic context within which actors operate and vigilance against common pitfalls of using the historical record to judge power and interests.” I am further perplexed to read, when they actually do acknowledge one of my points on the issue, that “Swenson suggests, rather implausibly, that employers had reason to understate their support even as the [social security] bill neared passage.” Why would that be implausible? If there is a chance that the legislators might tidy up a few details to suit you before passage—like the car salesman who might, for example, reduce the price or lengthen the warranty—why show enthusiasm instead of resistance? Also, leaders of business organizations have to be careful, thus often ambiguous, given that their constituents have diverse interests and ideological inclinations. Here, I simply repeat my view that they need to address my criticisms of their evidence, like Skocpol's before them, supposedly demonstrating a seamless wall of sincere opposition.

Some of my criticism of Hacker and Pierson rests on an interpretation of their argument saying that the few prominent capitalists who expressed open support for reform and cooperated with FDR “did so only to be invited into deliberations about the legislation, and thus to be in the position to make it as benign as possible.”⁷

I chose my formulation for two reasons. First, Hacker and Pierson said in their initial Politics and Society challenge to my argument, which initiated this debate, that “Swenson never truly grapples with the strong evidence that what business support there was for . . . the SSA was itself a strategic accommodation, driven by fear of less attractive alternatives” (my emphasis). Second, the only “strong evidence” they adduce for this, and make a very great but I think exaggerated deal out of, is a quote from National Association of Manufacturers (NAM) leader Robert Lund criticizing fellow industrialists for “attempting to . . . modify or change or meet half-way, the situation” and “meeting half-way what we know is fallacious.” I remain surprised, therefore, that Hacker and Pierson so vehemently object to my characterization of their claim. If, as Hacker and Pierson believe, these business figures were fundamentally opposed to legislation, why else would they have wanted to disguise their opposition other than to avoid being considered personae non grata and thus unwelcome in discussions that might “change the situation”?

Finally, on strategy, Hacker's and Pierson's response criticizes literature that, in their view, overstates the ability of “crafty corporate leaders” to shape key features of social legislation. They no doubt mean, in particular, G. William Domhoff's historical research. In their Politics and Society article, they explicitly denigrate my use of his work, with all its “conspiratorial claims.” While their characterization of the literature in general, and Domhoff's in particular, needs to be challenged, this is not the appropriate venue. I would like to emphasize one more time, however, that, in my argument, what matters is the craftiness of politicians, not corporate executives. There is no doubt that FDR was crafty; according to Barbara Armstrong's fascinating oral history, he was decisive in selecting the diverse kind of people involved.⁸

Details and Sausage Making

Hacker and Pierson seem to think that it is important for my argument that the Old Age Insurance (OAI) part of the SSA was progressive corporate capitalists' “first choice” among a full range of alternatives, including no pension legislation. My argument actually does not require that, and as a matter of fact, rejects it. The New Dealers were pursuing their own ends, while also attempting to please other interest groups. The outcome was bound to be a bit messy. Recall that, in Germany, both capitalists and Bismarck thought that the legislative process, like sausage making, was an unappetizing sight, but the proof was in the eating.

What matters for my cross-class argument is that the New Dealers regarded OAI as very good for big American employers, and better than no legislation at all. Admittedly this is impossible to prove. Yet so is Hacker's and Pierson's view that capitalists' favorite outcome, as perceived by the reformers, was the extreme one of no legislation at all. They think, it seems, that the vast majority of employers in America, including the corporate progressives, call them “actor” A, preferred A₁ (no legislation) over A₂ (legislation with the Clark opt-out amendment) and A₃ (the SSA with improvements, but not the Clark amendment), on a policy continuum of A₁, A₂, A₃, B₃, B₂, B₁. In my view, the employers that the New Dealers were trying to accommodate (not the vast majority), preferred A₃; the New Dealers themselves preferred B₃ (the SSA with more accommodations to other interests and views). What resulted was something between A₃ and B₃. And adjustments over time, under Roosevelt, moved things more in the direction of the capitalists' A₃ than anywhere else—for example, the important shift toward pay-as-you-go financing and favorable tax treatment of supplementary company pensions. Cross-class alliance making was an ongoing process.

Here Hacker and Pierson should probably agree with me that the historical record about businessmen's position taking on its own will not resolve this controversy. That is why we need to introduce theoretically informed knowledge about interests, which is a prerequisite for good strategic and power analysis. Knowledge of interests, not just positions strategically adopted, is an ingredient that crafty legislators, if not all political scientists, often go to considerable lengths to procure.

The details of New Deal legislation that disappointed corporate progressives (deviation from A₃) were relatively insignificant, and far too much can be made of these small defeats—just as one can make too much of aspects of legislation that annoyed Swedish or German capitalists. No one would ever argue that politicians could please all of any group of capitalists all the time, and even the pleased ones try not to look too pleased. Likewise, repeatedly insisting that many or even a majority of capitalists wanted to “modify” the SSA, according to the 1939 Fortune survey, does not justify calling their position “opposition.” Any employer today could quickly come up with a way to modify her or his favorable tax treatment of employee health benefits, but that would not amount to “opposition” against the existing rules. That “modify” could mean “make voluntary”—that is, move to A₂ by adopting the Clark amendment—as Hacker and Pierson think, lacks credibility, because the evidence definitely shows that the business community's leaders, including the insurance industry that once pushed for it, had totally rejected the idea by 1939.

In sum, the instrumentally powerful actors here were very well-informed politicians. In the American case, they handed capitalists whom they wanted to please a system of compulsory, contributory social insurance, financed by payroll taxes. The historical evidence, bolstered by theoretical analysis of interests, indicates that the New Dealers had good reason to believe that the capitalists did not want means testing, financing out of general revenues, or a Clark opt-out amendment. Segmentalists, the politicians knew, would like compulsory pensions big enough to take a significant load off their backs, and modest enough that they could afford tailor-made, tax-favored company pensions to supplement them.

Company Pensions Before and After the SSA

Hacker and Pierson believe that my argument is weak because only a relatively small number of employers had pension plans before 1935. First, the theory behind my interest analysis indicates that all employers structurally able to pursue a segmentalist strategy stood to benefit in product market competition from the regulatory aspect of imposed social security charges whether or not they already offered pensions. Second, the usual data, according to which there were no fewer than 307 plans in the United States and Canada outside railroads, do not include “informal plans” of which there were many (at least 148 in 1925), often at medium or small, but also some large employers.¹⁰

Even employers with informal plans had every reason to be worried about having overextended themselves with pension promises, if only informal ones, given that “good employers” taking responsibility for their aging employees was becoming the norm—propagated by none other than capitalist organizations themselves. OAI promised assistance to employers with formal or informal plans in fulfilling promises to their workers at lower costs than they had been anticipating, while imposing unanticipated costs on competitors who had no plans at all. Second, even if only forty or so employers employed 50 percent of employees covered by formal company plans, these were huge employers. Their economic and therefore political significance vastly outweighed their numbers. The dangerously unfunded pension liability of badly designed company pension plans was in the neighborhood of $2 billion—or around 2 percent of the U.S. gross national product. The New Deal was a big deal for big employers.¹¹

For various reasons, Hacker and Pierson think that real corporate support for OAI came only in the 1940s and beyond. They make this claim to demonstrate a lack of realism in my argument about New Dealers' anticipations in the early 1930s. In fact, they assert that “the clearest example of manifest business support of OAI” came during debate over the extension of benefits in 1950. But Hacker and Pierson offer no evidence or argument for thinking that the percentage supporting the 1950 extensions actually exceeded the percentage supporting keeping the SSA in the 1939 survey. Furthermore, they should have mentioned the 1943 U.S. Chamber of Commerce survey. According to Kodak's Marion Folsom, the survey generated the qualified majority necessary to reverse the Chamber's hostile but certainly very divided position of 1935. Two-thirds or more voted in favor of every important feature of the SSA.¹²

Finally, Hacker and Pierson try to downplay the importance of premeditation or foreknowledge on the part of New Dealers in taking a big expense (unfunded liabilities) off large companies' hands with OAI. It would, they knew, allow companies to either drop their existing plans so that OAI could honor their retirement promises to blue-collar workers at a lower cost to the companies, or else allow them to spruce up their benefits in new ways at little or no extra new expense. On this, I will only add one more fascinating twig to the thicket of historical facts: AT&T's pension plan, before the passage of the SSA, contained a provision “to decrease pension payments by any amounts payable by law under a public benefit plan.”¹³

Richter quite probably, we can comfortably surmise, helped spread the word in the CES that, like AT&T, other big employers could do the same, and save money. How long it took companies to “integrate” plans, an issue Hacker and Pierson take up, is almost entirely irrelevant. What matters is that politicians anticipated that capitalists would be happy, wanted that result, and had no interest in exposing the reality for all to know. Adjusting your behavior in anticipation of another group's reactions is evidence of their power at work, however hard it is to observe and historically substantiate.

III. CONCLUSION

Hacker and Pierson have a bit of fun taking me to task for saying that the evidence and theory I present “probably” explain the SSA better than a sudden loss of capitalist power. Here, they persuade me that I was too careless in my use of the word “probably,” and that I should have left it out. There are definitely too many problems with Hacker's and Pierson's argument about the Depression and the shift from state to national politics—for example, too much evidence from the NRA experience that capitalists' interests changed—to conclude that capitalists' loss of power can explain the passage of the SSA. Finally, there is too much comparative evidence about capitalist interests in welfare state development in other countries to believe that a shifting balance of class power will ever explain much about the American welfare state.