(1) Cartelism

In some sectoral settings, most notably in coal, clothing, construction, transportation, and some light manufacturing, employers often collectively desire to maintain wages that are above market clearing levels. Sectorally organized unions are necessary partners in the setting of such wages with centralized institutions of negotiated cartelism. Using strikes, they enforce on wayward employers the minimum standards agreed upon in collective bargaining, and reduce entry on a low-wage basis. Both sides share the monopoly rent from this non-governmental regulatory floor under product and service prices.

(2) Solidarism

Capitalists might also organize to hold wages below market clearing levels for the sake of maintaining low prices. Swedish capitalists' organization and collective action, as I will show, is exactly of this “solidaristic” nature (after Swedish parlance), suited especially to keeping Swedish products price-competitive abroad. Organized labor's participation in market-defying wage suppression, unlike employers', might seem somewhat puzzling. Nevertheless, for reasons clarified later, a cross-class alliance for negotiated solidarism was the central foundation of the Swedish political economy as its welfare state evolved. The solidaristic system and its supportive cross-class alliance help explain both the unusual organizational strength of the labor movement and puzzling questions about the timing and shaping of the Swedish welfare state.

(3) Segmentalism

Instead of cartelism and solidarism, employers can construct a decentralized, but equally well institutionalized system, defined here as segmentalism.¹⁸

An integral part of the segmentalist labor market regime is the widespread company-level provision of social benefits – “welfare capitalism” in the parlance of historians. Administrative, underwriting, and other economies of scale for group insurance and related benefits make welfare capitalism a cheap efficiency wage strategy for larger employers. The moral economy of equity, reciprocity, and social responsibility in workplace relations – the partly self- imposed normative bounds of managerial sovereignty – help institutionalize the system. Unions may thus participate in a system of negotiated segmentalism. This “unionized welfare capitalism,” dominating postwar U.S. labor relations such as autos, steel, and other important mass production sectors, coexisted with negotiated cartelism in other industries.²¹

In the following sections, I begin by discussing American segmentalism at length, and cartelism only very briefly, demonstrating how politicians responded to the interests of both in the passage of social insurance legislation. For the sake of brevity, I focus only on the Old Age Insurance portion of the Social Security Act. I then turn to Sweden, where the behavior of political reformers in a different institutional context – solidarism – will reveal the logic of a radically divergent and differently timed course of welfare politics.

Old Age Insurance

A look at the Old Age Insurance (OAI) component of the Social Security Act of 1935 (SSA) will indicate the enormous plausibility, if not certainty, that New Dealers were anticipating capitalist interests and reactions. Corporate progressives from the segmentalist camp were most prominent nationally in revealing to liberal politicians the changing nature of their interests. Most active on the social insurance front were individual businessmen like Gerard Swope from General Electric and Marion Folsom of Eastman Kodak, both of whom were brought into administration discussions of an omnibus Economic Security Act. Swope had advocated corporatist, albeit quasi-private social insurance, including disability protection, in his famous “Swope Plan” of 1931. The plan was an important inspiration for FDR's corporatist-cartelist National Industrial Recovery Act. (For more on the significance of the NIRA for the SSA, see below.) Furthermore, it was also a strong signal of interest in the regulatory benefits of social insurance. Swope's plan would have empowered sector-specific trade asso- ciations to extend their own – better – standards on non-members, and therefore “place the same social burdens on companies competing in various parts of the United States.” If necessary, federal government enforcement would be enlisted. Likewise, Marion Folsom changed his view from the 1920s that voluntary action was enough. His Depression experiences persuaded him that “the only thing to do” was have the federal government impose a solution.³¹

In deliberations on the SSA, the Roosevelt administration also relied heavily on the intellectual services of experts such as Brown and Murray Webb Latimer, both tied by careers and financial dependence to Rockefeller wealth from the oil industry. (Walter Teagle from Rockefeller's Standard Oil of New Jersey joined Swope and Folsom as advisors to Witte). Latimer was an employee of Industrial Relations Counselors, Inc. (IRC), a firm subsidized on a yearly basis by John D. Rockefeller, Jr., as G. William Domhoff makes abundantly clear. In fact, according to Witte, “almost the entire research staff of the [IRC] was placed on the payroll” of FDR's Committee on Economic Security (CES) to study the matter of unemployment insurance.³²

Rockefeller's possible behind-the-scenes encouragement is interesting to speculate on. In 1933, Raymond Fosdick, Rockefeller's attorney and close advisor (and FDR friend), reported glowingly to Rockefeller on IRC activities. In his letter, he reminded Rockefeller that his subsidy to the IRC in 1932 was seven times as great as its service revenue and consulting retainers. Among other things Fosdick praised were IRC activities in “shaping and administering [unemployment and pension] legislation” in Wisconsin and Minnesota, and pursuing contracts to do the same in two other states and Canada. All this he included in a letter designed to persuade the oil baron to part with more money for IRC efforts.³³

The oil industry offers an excellent example of segmentalists with regulatory interests in social legislation during the Depression. Slack demand and high unemployment exposed the oil industry, even the Rockefeller imperium, to disruptive competition from low-wage and low-benefit product market competitors. In written testimony to the Senate, an editor of National Petroleum News declared that “most if not all of bigger oil companies now have, and some have had for many years, various forms of pensions.” If their thousands of competitors “were forced to contribute to such protection as bigger companies are now doing, it might help to lessen some of their price cutting by bringing up their costs.”³⁴

Low-standard manufacturers in many other product markets also relied less on worker good will and low turnover for productive efficiencies than their larger competitors. Furthermore, because they reaped the benefits of the major employers' success in crippling organized labor, they enjoyed great flexibility to lower wages and shed workers to gain advantage in intense price competition. To that extent, segmentalists fouled their own nests. New entrants to competition could enlist unemployed workers at very low wages. To follow sub-standard competitors down the low road, segmentalists would have had to violate their institutionalized, company-based moral economy. Those who went ahead with deep cuts faced worker unrest; those who cut back the most were most vulnerable to a militant strain of industrial unionism.³⁵

Because their wage and benefits practices gave them market and managerial advantages (and relative peace from union organizers), segmentalists' interest lay in compulsory social insurance taxation of their competitors. This would compensate them for segmentalism's market vulnerability under combined conditions of deflation, unemployment, and union weakness.³⁶

Harvester's experience in this particular regard sheds light on a fascinating episode, about which we still know relatively little, in the Roosevelt administration's OAI deliberations. When the cabinet-level Committee on Economic Security (CES) contemplated dropping old age insurance from the bill, New Deal big business advisors strongly urged them not to. We know about this truly extraordinary intervention primarily from CES staffer Brown.³⁸

For one thing, the OAI carrot was every bit as tempting to the corporate progressives as the Townsend stick was threatening. Not only did it have a regulatory advantage for segmentalists as a whole; it also delivered a large subsidy for many of the 300 or so large employers across America, ones who had company plans, many of them very badly administered. Company pension plans, all too often, were badly or entirely unfunded. An actuary's nightmare, they were a money-hungry “juggernaut,” as a consultant from the Business Research Corporation put it, that employers had unwittingly set in motion in the 1920s. As much thought had gone into them as into ordering pencils, according to one analysis.⁴⁰

The crisis was exacerbated by intense price competition from manufacturers who had made no pension promises and employed younger workers, or were now hiring up large numbers of the desperate and undemanding unemployed. Now, with the Depression in full force, the need to retire large numbers of less-productive elderly workers without reneging on their pension promises assumed crisis proportions. One solution was to purchase retirement insurance on their employees behalf. However, the price for private insurance was, an IRC study estimated in August 1935 (the same month the SSA passed), 33 to 100 percent more than what OAI was going to charge to guarantee the exact same benefits their workers were to receive with OAI.⁴¹

In sum, the New Dealers knew that OAI offered a large bailout as well as a regulatory blessing for major segmentalists. They were not, however, the only employers with interests, understood by the reformers, in legislation that would impose a floor on labor costs. Cartelists were also injured by cut-throat low-wage/low-price competition associated with deflation and unemployment. For example, the New York Building Congress, a cross-class confederation composed of over 600 municipal contractor associations and unions, began as far back as 1930 to call for compulsory pensions. It sent a lengthy report in 1932, advocating the same, to the Senate. (Its chairman, a vice president of F.W. Dodge Corporation, also believed a compulsory unemployment insurance system was “inevitable and desirable.”) This same organization testified in favor of New York Senator Robert Wagner's labor relations bill because it would empower unions to uphold wage standards “for the benefit in the long run of the employers of that industry.” Imposing standard pension costs across the building industry would also conform nicely with negotiated cartelism, especially by enlisting the power of the state to impose new costs on substandard, nonunion contractors.⁴²

Solidarism

Understanding the Swedish labor market regime requires examining its features in light of economic theory about efficiency wages and segmentalist labor markets. Because Swedish employers in the period before and during the major social policy reforms of the 1940s and 1950s were organized in highly centralized and authoritative organizations, a number of questions arise: If large-scale employers who are inclined to favor an efficiency wage and welfare capitalist strategy are able to overcome numerous obstacles to joining together in a collaborative strategy, how would their wages and benefits differ? If they chose a centralized multiemployer approach, what would the effects on the interfirm and interindustry wage structure be? How would they deal with the problems of labor supply, turnover, and effort? What would make it possible for them to enforce the collective choice?

An interesting offshoot of the efficiency wage literature helps answer some of these questions. Beginning, crucially, with the theory's assumption of a wage-productivity nexus, models of unilateral centralized pay setting by a highly disciplined association of employers indicate that wages would be set lower than in the decentralized model. Likewise, employment would be higher. Interestingly, the same conclusions hold for models incorporating joint union influence in negotiations.⁴⁴

The equilibrium in the centralized model is one where many firms offer wages below market-clearing levels, not above, as in the segmentalist case. In other words, the marginal revenue to a firm of hiring one additional worker exceeds the wage the firm has agreed to offer under the terms of the centralized arrangement. In that sense, centralization creates for many firms a scarcity of labor. That is to say, there will be a shortage of workers willing to take work at the low, market-defying rate. Many firms, in this case, would actually prefer to raise wages to reduce quits or shirking on the job, expand employment and production, or otherwise improve productivity and profits. This approach, however, would only work if others voluntarily abstain or can be prevented from following suit. Because “poaching,” or cheating on the arrangement by raising wages would be profitable, centralized wage setting presupposes some mechanism for monitoring adherence and punishing violations.

In short, the centralized equilibrium sees “involuntary” underpricing of labor by the individual employer, and a scarcity of labor, rather than the segmentalist's voluntary overpricing and excess labor. Because segmentalist behavior is voluntary, no external agency is necessary to enforce it. But in solidarism, monitoring and policing the actions of individual firms must succeed to make it possible to speak of any kind of equilibrium. This is an artificial equilibrium, however, in an economic sense, for individual firms have a strong profit-motivated interest to behave in ways that would upset it. Workers will gladly rock the boat by offering their services at higher wages.⁴⁶

For reasons that cannot detain us here, analysis of solidarism indicates that in addition to setting wages across much of the labor market below equilibrium, employers will also collectively pursue a number of other policies designed to maintain solidarism's viability. The most striking of these is to facilitate compliance with wage restraint and reduce turnover by coordinating and enforcing pay policies in the direction of pay compression through interfirm and intersectoral pay standardization. In part, this is simply the most viable administrative rule for maintaining the legitimacy, supported by norms of fair competition over labor, of a rationing system (allocated wages rather than coupons or stamps) under conditions of scarcity. Wage compression in the form of ceilings becomes more important than the wage floors associated with negotiated cartelism; given the scarcity phenomenon, wage floors are largely superfluous.⁴⁷

Most important for the current analysis, employers, acting collectively, are likely to regulate the use of company-based social benefits. This tendency is part and parcel of the solidaristic policy of pay compression. For several reasons, a solidaristic employers' association is likely to choose suppression and even elimination of company welfare over centralized regulation. Heterogeneity in capital and labor force attributes, along with other market constraints, will make agreement difficult to reach on common standards due to funding and benefit costs that vary considerably across firms. Labor-intensive, dangerous, or unhealthy firms or sectors, for example, will have dramatically higher insurance costs than others who will be reluctant to cross-subsidize them to even out the costs. Given the difficulty of achieving agreement around standardized schemes, complicated by the presence of pre-existing company based benefits, suppression of benefit growth and expansion may be the employer association's most feasible option. Complete elimination also recommends itself, but, as Sweden suggests, political intervention with compulsory alternatives may be necessary.⁴⁸

Of course, unions have to be part of the overall picture. Solidarism, in theory, can be pursued unilaterally by employers. Given its emphasis on wage repression, this might be the only thing one would expect to see; however, a cross-class alliance with organized labor at a highly centralized level is also a possibility, as indicated above, and can serve as the necessary cement of solidarism. It might, in fact, be only in the crucible of conflict with unions, and the evolution of centralized relations within and between employer and union camps, that the cohesive norms and disciplinary mechanisms of solidarism arise among employers.

Because organized labor's common concerns for pay equity harmonize well with organized employers' self-interest in pay standardization, solidaristic employers have something to offer unions in exchange for rigid wage discipline. Also, labor's ideological concerns about equity across the entire labor force, not just for those currently employed, give it reason to favor low levels of unemployment, and therefore restrain workers' demands. Finally, centralized governance of labor markets can win the approval and participation of organized labor because it promises to strengthen the insecure hierarchical control of union leaders over their fractious membership and sub-units. Union leaders can promote the solidaristic policies as their own, propagate egalitarian norms to justify them, and likewise participate in their design and implementation.⁴⁹

Wage Standardization

That sectoral multiemployer bargaining brought solidarism, not negotiated cartelism, is manifestly apparent in the substance of the centralized agreements forged.⁵⁴

The solidaristic normal wage system spread to other sectors. Paper, paper pulp, and construction opted for normal wage schemes in 1921. Engineering employers came under intense criticism from other sectors, textiles in particular, for failure its failure to cap wages. Labor scarcity in textiles was partly a consequence of the drift of manpower away from textile producing areas and across sectoral lines into engineering, with its higher and upwardly drifting wages. For technical reasons peculiar to the highly heterogeneous sector, engineering rejected normal wages; instead they tried to deal with the wage drift problem in 1923 with a complex surrogate. With normal wages or an alternative in place, firms that exceeded their ceilings were exposed to the censure and sanctions of their associations – just as workers taking wages below negotiated minima in American joint cartelism suffered the censure of their union comrades.

Solidaristic Suppression of Welfare Capitalism

Early on, eliminating company-based social benefits stood high on Swedish employers' collective agenda – unlike in the United States, where all major employer associations dominated by segmentalists energetically preached the virtues of “welfare work.” Although it was not spelled out in writing, the 1905 VF-Metall agreement occasioned the wholesale dumping of physician, credit, and accident insurance benefits by one major shipbuilding firm. In 1939, looking back, a prominent shipbuilder and employer statesman recalled, with gratification, VF's success in 1905 as a clean break with paternalism. At one of Sweden's largest machine making works, social benefits as a proportion of total remuneration dropped after peaking in between 1905 and 1909, as the engineering agreement sunk its institutional roots. Centralized bargaining, according to the director of Sweden's largest building materials firm, allowed him to stop being a “benevolent patriarch.” It was better, he thought, that workers “help themselves” with strong unions and the wages they negotiated.⁵⁶

The first sector to take strong action against corporate welfarism was steel.⁵⁷

The textile agreement of 1914 required individual employers to eliminate in-kind remuneration, including worker housing. Medical services and accident benefits were retained for the time being, though standardized across the industry. In 1921, paper employers collectively bought off company welfare benefits with standardized wages. Elimination of social benefits had been a desideratum since the organization's inception in 1907. In 1922, engineering employers took unilateral action, during a temporary hiatus in centralized bargaining, ordering member firms to eliminate all in-kind company benefits, such as free medicine, and free or subsidized housing and other “necessities.” In 1932, with backing from LO, the Swedish Labor Confederation (Landsorganisationen), pulp employers imposed an agreement that included “an old demand” – substantial elimination of such social benefits as rent subsidies and free company housing.⁵⁸

Postwar Conditions and Benefit Drift

From its inception, the immediate and inevitable consequence of wage restraint, especially for the most dynamic capitalist enterprises, was labor scarcity. For them, wage restraint delivered high profits and excess demand for labor. To avoid costly lockouts, and achieve a more stable and equitable wage structure – not to mention a secure, high-prestige role for union officials – unions participated in setting these below market wages. Inevitably, labor poaching and wage and benefit drift followed.

For example, less than a year and a half after the 1905 engineering agreement, manufacturers in engineering complained bitterly about fellow employers, and little about Metall. We are witnessing, as one put it, “the most cutthroat and worst sort of competition that exists, when . . . people drive up wages by outbidding each other.” Because the shops were “so piled up with work . . . there are too few skilled men.” The best remedy, he said, was to take in more apprentices, because “by competing with each other and driving up wages we cannot create more men.”⁵⁹

With Social Democrats in power in the 1930s, LO did more to hold back wage militancy, especially in the construction sector, than to fan its flames. Revival of economic growth in the late 1930s started to bring similar “wage drift.” Rising piece work earnings contributed because firms studiously neglected to reset piece rates with the advance of labor productivity. Also, employers held each others' “disloyal recruitment” responsible. Frustration turned into alarm as these symptoms of labor scarcity grew during the war years and beyond. Again, the labor confederation continued to help restrain wages.

The years 1945 and 1946 were extreme, but the symptoms of labor market disequilibrium arising from wage restraint and intrasectoral leveling in the face of strong growth were not atypical of what happened in ensuing years.⁶⁰

Unemployment among union members reached rock bottom at about 2 percent in 1946. SAF calculated that Swedish industry could easily have absorbed another 100,000 to 120,000 workers at repressed wage levels; labor mobility and turnover was “abnormally high.” Roughly 5 percent of all workers changed jobs on a monthly basis, a disturbing number of “job hoppers” (hoppjerker). The Social Democratic government tried with prolonged wartime building controls to check upward wage pressures from the high-wage building sector. Top employers had called for these state measures, while also resolving among themselves not to “let wage competition run free.”⁶²

In the late 1940s, mutual recriminations about disloyal wage practices, sometimes called “black market wages” (svartabörslöner), rose in volume.⁶³

People's Pensions

As tensions flared inside SAF over the leadership's desire to suppress segmentalist cheating with company benefits, the Social Democratic government passed its first major universalistic social insurance reform in 1946. This legislation brought a large increase in the uniform, flat-rate “people's pension” (folkpension) last revised in 1937. The fiscally conservative Social Democratic government had initially favored a cheaper, more means-tested system concentrating resources on the more needy. By contrast, ironically, the Conservative Party came out openly against means testing and favored the more expensive universalistic solution.⁶⁸

The Conservatives' position was probably influenced by SAF. From the very beginning SAF “totally rejected” a cheaper incomes-tested scheme. Means-testing would, SAF feared, reduce incentives for pension recipients “to improve their situation through work and savings.” In other words, labor scarcity would be exacerbated if workers' earned incomes and assets disqualified them from receiving full government pensions at age 67. SAF executive director Fritjof Söderbäck insisted that aging workers with “the ability and will to offer their manpower to the disposition of the national economy, should not . . . suffer a substantial reduction in their people's pension.”⁶⁹

The labor confederation agreed with the employers. On the great similarity in SAF's and LO's views about the size and structuring of pensions, an editorial in Dagens Nyheter, entitled “Between Brothers,” speculated that some sort of secret agreement had been reached on the matter. This solid cross-class unity must have contributed to the government's change of mind. The reform, to take effect in 1948, passed with broad, multiparty support. Although not exactly a “triumph for the right,” as the Conservatives' Sydsvenska Dagbladet put it, it was at least an occasion of “national unity,” as the Liberals' Dagens Nyheter proclaimed.⁷⁰

SAF did not simply choose expensive universalism over cheaper means testing as the lesser of two inevitable legislative evils. SAF documents show unequivocal welcome for government action. Private provision of pensions only confounded SAF's efforts to govern the labor market on a solidaristic basis. Shortly before passage, turmoil over private benefits was intense and interest in state measures heightened. SAF's Motor Vehicle Transport Employers' Association warmly endorsed the universalistic scheme under consideration, hoping “that the government pensions would assume such dimensions and form that the worker pension issue [will be] completely settled and that requests for further pension benefits from employers can be rejected.” The association had been grappling with internal conflict over the issue, because unorganized firms were raising benefits to compete over labor with the public sector as well as with more restrained SAF members.⁷¹

As tight labor market conditions intensified throughout 1947, SAF members praised the new legislation for the relief it provided from pressure on firms to use pensions as a “drawing card” to attract labor.⁷²

Comprehensive Health Insurance and Sick Pay

A similar story can be told about the passage of health legislation in 1953.⁷³

Employers' hopes were fully dashed because of labor scarcity and the unregulated growth of company benefits, especially sick pay, in the 1940s. By the early 1950s, almost all workers employed by SAF members received some mix of free or subsidized medical services, hospital benefits, sick pay, or employer assumption of payments into the subsidized insurance fund system.⁷⁵

SAF's chief expert on social insurance sat on the commission that drafted the legislation. He also tried to persuade employers critical of various details of the complex legislation to overlook them in light of its broader advantages.⁷⁸

For a mix of complex reasons, SAF employers strongly favored sick pay scaled according to a worker's wages, not a flat, and therefore cheaper Beveridge-style system, which Social Democratic Minister of Social Affairs Gustav Möller fervently advocated. Möller expected privately provided or negotiated supplements to fill them out. However, SAF wanted employers out of the business altogether. Initially, LO was uncertain and divided on the issue. By 1953, however, the labor confederation reached consensus with a more resolute SAF on the sick pay question and away from Möller's position. This alliance no doubt helped swing the government toward the more comprehensive, compulsory system. Once again, as with the pension reform of 1946, a cross-class coalition of the employer and labor confederations prevailed over Social Democrats' initial plans for cheaper and less comprehensive reforms.⁸⁰

Active Labor Market Policy

No major Social Democratic reform initiative of the period is so closely identified with the innovativeness if not power of the left in Sweden than its famous “active labor market policy” (ALMP). ALMP's expenditures and innovative development took off after the mid-1950s, when Social Democratic governments began lavishing growing public resources on programs such as the retraining and geographic relocation of workers made redundant by rapid technological and market changes. Workers in depopulating areas even got help with selling their homes, many of which were acquired as a company benefit; ALMP thus helped SAF uproot some remnants of welfare capitalism. Other policies remained important, such as temporary job creation to keep people off passive measures like unemployment compensation, and in healthy, active contact with the labor market. However, active supply-side programs, including retraining and relocation, grew the fastest.⁸¹

The labor scarcity of the 1940s and 1950s, generated by solidarism in combination with robust demand and complementary macroeconomic policy, was the underlying cause of ALMP. (Restrained solidaristic wage policy made it possible to pursue relatively non-inflationary full employment policies; ALMP was designed to make it even easier to reconcile price stability and full employment.) Employer leaders complained of continuing “overfull employment” and a “disturbing degree of competition over labor among companies.” Wage drift threatened “bankruptcy for the bargaining system.” In shipbuilding, for example, major companies routinely evaded solidaristic controls by contracting out work to firms paying wages above what VF and SAF allowed in order to get profitable work done. Even the LO leadership joined in criticism of the practice. One shipbuilder complained of “a lack of balance between demand and supply in the labor market.” The situation was so untenable, he said, that there were only two logical choices available: let wages loose or ration labor.⁸²

In other words, administered pricing of labor inevitably spawned the idea of administrative rationing. ALMP partially served that purpose. As a labor allocation device, it serviced the labor supply needs of a system of employers' own making. In a vivid first-hand account of efforts to promote active labor market policy from the late 1940s through the 1960s, LO economist Gösta Rehn never mentions employer opposition. According to his writings in the 1950s, the main objective was to set up “permanently increased efforts to stimulate the adaptation of manpower to industry's needs.”⁸³

Rehn also modestly dismissed the idea that he and other LO economists' invented the idea of mobility measures, for “such demands come from outside” the trade union movement. By this, he could only have meant from employers.⁸⁴

It is thus necessary to reexamine Esping-Andersen's view that ALMP “was only possible due to the extraordinary labor market powers of the union movement.”⁸⁷

And flow they did, increasing to about 2 percent of GNP and 6 percent of total government budgets in the 1970s. Active supply-side programs like retraining and relocation grew the most rapidly, reaching nearly 40 percent of total spending by labor market authorities in 1982 from only a few percent before 1960.⁸⁹

In conclusion, the labor market program harmonized so well with employer objectives in solidaristic governance of labor markets that one SAF negotiator praised active labor market policy as “employers' ally.” Another SAF official characterized bourgeois party attempts to discredit the Social Democrats' labor market policy as “entirely against our interests.”⁹¹

Strategic Talk and Real Interests

Hacker and Pierson try to clear the way for their argument about the shifting balance of power against capital with three important assertions. First, that the New Dealers could not have been sensitive to capitalist power and interests because almost all organizational, and therefore publically vocal manifestations of capitalist preferences were negative regarding the SSA, especially from multisectoral organizations such as the National Organization of Manufacturers and the U. S. Chamber of Commerce. Second, that those few prominent capitalists who expressed sympathy for reform did so only to be invited into deliberations about the legislation, and thus to be in a position to make it as benign as possible. Employers were “driven by fear” – that is, of less attractive alternatives. Third, they argue that capitalists and their IRC experts brought into deliberations did not get their way on certain details of the legislation that occasioned controversy.⁹⁶

On the first point – regarding widespread capitalist opposition – Hacker and Pierson bring no new evidence to the debate that Theda Skocpol and her collaborators have not already offered. Nor do they present and confront the strong and detailed criticism that these old assertions about the unbroken wall of business opposition to the New Deal have already come under.⁹⁷

The most interesting and important of these three arguments is the second one. Hacker and Pierson forcefully argue that individual capitalists who spoke in favor of reform were strategically exaggerating their support for government action and, therefore, hiding their real preferences for exclusively private, voluntary action. Capitalists did so in the hopes that they would then be in a better position to weaken the legislation.

Confidence in this claim can be rather easily shaken. Threats of legislation were not on the horizon in 1927, when a Pennsylvania state commission found 38 industrialists “in favor of compulsory old age insurance,” even though, not surprisingly, they differed on how to construct it. As mentioned earlier, the New York Building Congress started calling for compulsory pensions in 1930.⁹⁸

In 1932, with no legislation looming, Henry Harriman, president of the U.S. Chamber of Commerce, said that, if businesses were not able to control themselves, compulsion might be necessary to regulate competition. In that spirit, he wrote, “we must place upon business the responsibility for setting up reserves to tide over periods of unemployment and to assure men against economic want and suffering because of sickness, accident, or old age.” By “accident,” Harriman was referring to long-term disability, and there was no legislative activity brewing on that.⁹⁹

With their emphasis on strategic talk, Hacker and Pierson claim to have improved the interpretation of evidence by taking into account the role of “anticipated reactions” in the power politics of social policy. Supposedly, corporate progressives like Swope anticipated a worse outcome from speaking the truth than wriggling like chameleons into policy making circles. Their only direct evidence for this, a “smoking gun” as it were, is a speculative statement by Robert Lund, Chairman of NAM, speaking on unemployment insurance. Lund, claiming to speak for other businessmen in FDR's mostly progressive Business Advisory Council, chided them for signaling interest to FDR, although they did not, in fact, “believe in” unemployment insurance.¹⁰⁰

Lund's protestations must, however, be taken with a large grain of salt. His recent takeover of the NAM leadership was paid for with contributions from people like the iron-fisted Tom Girdler of Republic Steel, responsible for Chicago's “Memorial Day Massacre” of picketing and picnicking workers two years later. How candid and nonstrategic, we might ask, were other capitalists in giving their opinions to the leader of a radicalized organization like the NAM (whose members were actually only a small share of American manufacturers), that might give them, as corporate progressive Edward Filene put it, a “reputation for radicalism that hampers them for further influence in business circles”?¹⁰¹

Reasons for reticence and strategic disinformation between capitalists probably varied quite widely. Many would have feared what Filene learned first-hand, that the openly liberal businessman risks a lonely life and “brings down upon his head the criticism of important groups of his fellow business men.” Hard-line circles spoke of Swope, for one, as a “dangerous man for industry” for his political activity and declaring the NAM unnecessary. His wife, “a former pal of Jane Adams [sic] of the Hull House, is active with him in their parlor pink activities,” one top East Coast union buster complained.¹⁰²

Compare Lund's perceptions with those of J. Douglas Brown, justifiably respected in corporate circles as one of the country's most sophisticated experts on company benefits. In a letter to Edwin Witte, Brown recounted that, during the February 1935 annual conference of the American Management Association, he had a chance to present the old age insurance plan as currently formulated “to a large number of industrial relations executives.” Their reaction, he said, was “very favorable.”¹⁰³

Attention to strategic political talk in power analysis is all for the better. But Hacker and Pierson's application of the insight is one-sided and, therefore, imputes great clarity and certainty where none exists. Not willing to be deceived by businessmen taking progressive stands, Hacker and Pierson are willing, on the other hand, to take at face value all outspoken opposition from others. In doing so they neglect to explore the distinct possibility, posited by Sanford Jacoby, who argues that capitalists in the National Industrial Conference Board, for example, were pursuing a “calculated strategy” in exaggerating their opposition to the SSA, not their support. Jacoby argues they wanted to punish FDR for other things.¹⁰⁴

Because Hacker and Pierson neglect to apply strategic analysis to anyone but corporate liberals, they implicitly and unwittingly endorse the equivalency premise, which states that the welfare state threatens capitalists' interests. Thus, if capitalists express support, we must investigate strategic motives, not direct interests; when they oppose it, we can comfortably take them at their word. Along with historical and comparative evidence against the premise, there is also an a priori reason, in fact, to think that exaggeration of opposition in some cases might have been the better strategy. In other words, all businessmen, even those who looked rather favorably upon legislation, would have been best to express opposition to make the New Dealers extra cautious in crafting it. (This would apply especially for businessmen who had no time or desire to be brought into the deliberative and legislative crafting process – that is, the organizational leaders cited by Skocpol, and now Hacker and Pierson, as sincere opponents.) It would be a perfectly rational strategem in any bargaining situation where moving in the direction of the other party is a favored outcome – whose plausibility the interest analysis confirms – but only up to and not beyond a certain point.

If there is any uncertainty about how far beyond that point the deciding party would go if real preferences were revealed, then vocal opposition might be rational. This would certainly be the case when the deciding party is moderate and shares very similar preferences with their putative opponents but also must listen to more radical constituents and advisors. The offering of feigned opposition can give moderates a welcome scapegoat and pretext for moderation.

Scholarly power analysis should, therefore, skeptically dissect strategic utterances of all sides from many angles. It should, by the same token, attribute to the real practitioners of power analysis – politicians, that is – the same and probably superior ability to interpret strategic talk. Evidence suggests that the New Dealers, in fact, perceived pervasive division, uncertainty, and even mutual fear among businessmen and their organizations, moved and dominated as they were by ideologies and ideologues. In the case of unemployment insurance, as prominent social insurance reformer Isaac Rubinow put it, “Individual employers are found to be much more ready to express their acceptance of [unemployment insurance] proposals in private.” When asked for open endorsements, however, “They prefer ’to have their name kept out of this'.”¹⁰⁵

FDR thought, for good reason, that major American business organizations such as the U.S. Chamber of Commerce misrepresented business interests in opposing the SSA. He shared this belief with “Brains Truster” Raymond Moley, a good friend of big business and the man who coined the term “New Deal.” Moley argued that businessmen were as susceptible to demagogy as anyone else. Kodak's CES advisor Marion Folsom viewed the NAM, also a strong opponent of the SSA, as an organization with low quality leadership and inept staff.¹⁰⁶

Domhoff brings to light suggestive evidence that Rockefeller himself would have been hesitant to attach his controversial name to New Deal legislation, preferring to delegate influence to other people, and thus avoid besmirching the enterprise in the eyes of reformers on the left. In a 1930 letter to Rockefeller, his son declared it would be “fatal” to have J. Douglas Brown, later a key CES advisor, known as a “Rockefeller man.” Brown's Princeton salary was being paid directly by Rockefeller; therefore John D. Rockefeller III thought the family's endowment to the university's Industrial Relations Section should be increased so that the salary come out of university funds.¹⁰⁷

In conclusion, while Hacker and Pierson only examine capitalists' strategic behavior, and only one-sidedly, they neglect to analyze the strategic thinking of political reformers. In particular, they neglect to consider politicians' strategic anticipation of capitalists' reactions in the future, based on their knowledge from the past about business interests and behavior before and after reforms, including possibly, strategic exaggeration of opposition. Strategic behavior also included personnel choices. Roosevelt and Frances Perkins chose many Wisconsinites to work for the CES – the same ones, for example, who had calculated that their state's unemployment insurance plan, according Nelson's meticulous study, would “win the approval of the progressive employers.” Initially they failed to generate open approval, but in the end, the Wisconsin employers became enthusiastic promoters.¹⁰⁸

Anticipated Support after Passage

As alluded to above, there is good evidence from earlier reform episodes that politicians became familiar with a pattern of initial business opposition and later support after passage. Hacker and Pierson do not acknowledge and challenge this evidence for a thorough treatment of the strategic interaction issue.¹¹⁰

Hacker and Pierson find evidence for their doubt in a book written by Edwin Witte the year after the SSA's passage. Witte wrote that the CES paid “relatively little attention” to existing company pension plans in fashioning OAI. The implication they find here is that the administration was strategically indifferent. They neglect to mention, however, what Witte wrote next, in virtually the same breath. Relatively little effort was spent because J. Douglas Brown and Murray Latimer took the trouble involved to talk to “some of the men in charge of personnel relations in several large companies.” In the process, the two experts confirmed that the employers “would not object” to the plan as currently conceived, even if it did not include “special provisions for employers with existing industrial pension plans.”¹¹²

Recall Brown's letter to Witte, mentioned earlier, which hardly documents strategic indifference. First Brown found among a “large number” of executives at the February 1935 American Management Association conference a “very favorable attitude” to the OAI plan. This is not surprising, considering that it was going to cost them 33 to 100 percent less than a private plan for the same benefits to their employees – mostly because the benefits for their many workers about to retire (“unearned annuities”) would far exceed what employers could afford in a private system. Brown also talked with “a number of key men” about letting companies maintain their own plans in order to skip paying social security taxes. What he discovered was that they “understood entirely the fact that contracting out is of decidedly doubtful advantage to their companies.” One of the companies was AT&T, whose actuary, Richter, Witte chose to assist Brown and Latimer in the CES.¹¹³

No doubt Brown's and Richter's findings had been included in a February 1 “memorandum to clients” of the IRC, which stated that “the combined cost to companies of the revised company plan and the national plan would presumably be less than the cost of their present plans.”¹¹⁴

This brings us to the Clark opt-out amendment, whose inclusion in the social security bill held it up in Congress but finally disappeared at FDR's insistence – and possibly (Brown wondered) because corporate executives helped break the logjam.¹¹⁷

Even Kodak's Folsom, whose initial advocacy of the Clark Amendment had only been “lukewarm” (as he put it later), came around quickly to the view that it was worse than unnecessary, it would also be a regulatory mistake. As he explained in a 1936 issue of the Harvard Business Review, in practice, only companies with a young workforce would have found the Clark exemption advantageous. Many of these would be new entrants or existing nonsegmentalist competitors. Their “selection against the Federal plan” would make it more expensive for large segmentalists with low turnover, and therefore with older unproductive workers now ready to retire, to fund the system adequately. Social Security taxes would then have to be higher. In other words, it would give back the regulatory advantage over competitors that the segmentalists had just gained. Thus, Folsom, one of the country's leading corporate progressives, opposed giving other capitalists a free market choice.¹²⁰

In conclusion, we have documentary proof that the CES knew of the overall benefits of OAI for big employers as early as February 1935, just after the administration bill had been submitted to Congress, and six months before SSA passage. We can not yet be sure if the facts were known during the drafting process between September 1934 and January 1935. Nothing suggests that they could not have been, for the CES experts, including Latimer, Brown, and Richter, had the facts they needed at their fingertips. More important, they also had a strong motive to see what the fallout would be for the kinds of employers that had paid them in the careers that got them where they were. We know they did go to great trouble to assemble the facts for external enlightenment once opponents shoved the Clark amendment in the way before passage and, then, made it clear that they would reintroduce it the following year. Then it became necessary to clarify for big corporate employers that the amendment was against their interests. That was the IRC's job, which it completed in August, at the time of passage, in a report showing that for, major employers, “the dollar for dollar benefits of the federal system cannot be matched by any private carrier.”¹²¹

It is fascinating, but not surprising, that less than three days after President Roosevelt signed the SSA, the IRC announced to clients that it was “now engaged in the formulation of several types of private plans which will supplement the pension benefits provided under the federal scheme and more adequately cover employees in the higher salary brackets.”¹²²

This is exactly the kind of “integration” that Hacker and Pierson portray as, at best, only a dimly perceived possibility until well after the SSA's passage.¹²³

Institutional Determination of Interests

Another intriguing part of the Hacker-Pierson argument concerns the institutionally conditioned power shift associated with the Depression, and it deserves extended discussion. True, policy making did shift to the federal level, and this had enormous consequences. Hacker and Pierson believe so because capital mobility could no longer check majoritarian reformist pressures as it had once done at the state level. Unfortunately for Hacker and Pierson, their implicit ceteris paribus does not hold. For capitalists' interests were different at the federal than state level. Furthermore, their interests in federal-level regulation were heightened by the Depression. Hacker and Pierson do not examine these possibilities, despite the broad consensus among historical institutionalists that interests, not just power, are conditioned by institutions.

Before the Depression, as Hacker and Pierson point out, capitalists stood to lose from state-level regulations and taxation that might give out-of-state competitors an advantage. (In the case of the country's first important social insurance reforms, workmen's compensation, there was actually support for state legislation, but supporters thought it would bring savings and therefore competitive advantage, not extra costs.) However at the federal level, the story is profoundly different. Although there is plenty of dispute on the details, all important scholars of business in American politics find bustling business activity in the design of pre-1930s federal regulation. Often business interests sought national regulation of other sectors, as shippers, merchants, farmers, and other businessmen did for railroads. The movement, backed also by some railroads, culminated in the Hepburn Act of 1906, which strengthened the Interstate Commerce Commission's power to control railroad shipping rates. At times, the leadership of business organizations did not always accurately and unambiguously reflect membership opinion – a phenomenon that, we have seen, shows up again during the 1930s. For example, a survey of the NAM membership revealed that a majority rejected the NAM leadership's position, and supported strengthening the ICC. Broad-based business support within the NAM, even including some in the food industry unhappy with state level legislation, also overruled the association's leadership on the question of food and drug legislation. The NAM's Pure Food Committee, siding with the majority against the leadership, claimed a large share of credit for the 1906 food and drug legislation.¹²⁴

As railroad and food regulation indicates, capitalist interest in federal regulation extended even to the matter of controlling one's own sector, not just other sectors. Historian Robert Himmelberg shows that the decade of the 1920s witnessed a long tug-of-war between the sectoral trade association movement seeking exemption or revision of antitrust laws and opponents such as Herbert Hoover. As commerce secretary and president, Hoover strongly favored industry self-regulation to uplift standards, promote efficiency and innovation, and stabilize competition, but consistently rejected large numbers of national trade groups' argument that administrative or legislative relaxation of the Sherman Anti-Trust Act was needed to make self-regulation effective. But even Hoover agreed on occasion with businessmen about the need for some compulsory national-level regulation, especially in such hypercompetitive and struggling resource industries as bituminous coal and oil, where cut-throat competition and therefore chiseling on prices and wages was rampant. He even favored negotiated cartelism, and therefore a strong national labor union, as an institutional mechanism for regulating bruising competition in coal mining.¹²⁵

By catapulting regulatory issues from the state to federal level, the Depression focused unprecedented levels of business attention on their regulatory desires. It also intensified those desires. According to Ellis Hawley's authoritative account, the idea of self-government of industry “moved rapidly to the fore” as the economy crashed. Thus “businessmen turned increasingly to the idea of government-supported cartels . . . [and] production quotas, price agreements, entry controls, and cost-accounting formulas.” Businessmen believed, Hawley writes, that these would facilitate calculations of the “right price” whose federal imposition “would wholly and permanently eliminate unfair price cutting” and “promote an unequaled degree of stability in all phases of industrial and economic affairs.”¹²⁶

At the outset, passage of the National Industrial Recovery Act in 1933 handed over to capitalists an almost unhampered right to devise codes of fair competition, impose price floors, and restrict production. Thus the National Recovery Administration (NRA) proceeded more or less exactly as the trade association movement wanted. Businessmen and their association lawyers were heavily engaged in the drafting of the legislation and then served as “code authorities” (especially industry-level deputy administrators of the NRA). The NAM's model code fairly “bristled with provisions for controlling prices and output,” as Hawley puts it, and a number of trade association spokesmen “made it plain that they hoped to imitate the practices of European cartels.” Business support for the NRA start to wane somewhat only after farmers, consumers, labor groups, Senators (especially progressive Republicans Borah and Nye), and a variety of government agencies, including the FTC, the Department of Agriculture, and the PWA, began to countermobilize and force the NIRA to override business input.¹²⁸

So business interest in regulation at the federal level was an enduring and now, in the Depression, intensifying phenomenon. In 1919, both the NAM and the U.S. Chamber of Commerce had called for revision of antitrust policy to allow collective self-regulation through trade associations. Throughout the 1920s, the business press railed against the destructive forces of competition – but only the “excessive” portion of it, of course; at the beginning of the Depression, large numbers of businessmen trooped to Washington to support regulation. While the NIRA was in force, business cries of indignation ultimately blocked executive efforts to scotch the widespread use of open-pricing arrangements in industry codes, a technique that proved highly effective in fixing prices. When NIRA extension looked in doubt in May 1935, 1,500 businessmen rallied in Washington for its continuation. After the Supreme Court sent NRA bureaucrats packing, “there was still considerable support for publicly sanctioned cartelization,” according to Hawley.¹²⁹

Thus in passing OAI , as the NIRA indicates, the New Dealers, acting at the federal level, were responding to intensified national regulatory interests, not a new balance of power against capital. They were also acting out of optimism that a cross-class alliance of interests would make the legislation politically robust. The same optimism had animated Robert Wagner, who piloted the NIRA through Congress, and persuaded a reluctant Franklin Roosevelt to give the thing a try. Labor did not support the NIRA only for its paragraph 7(a), which affirmed workers' right to organize. As Hawley notes, prominent labor leaders John L. Lewis and Sidney Hillman thought it illogical to expect “a chaotic and overly competitive industry, one that was almost chronically depressed, to pay decent wages.” American Federation of Labor Vice President Matthew Woll favored antitrust revision as a way to guarantee higher standards for workers, and arranged for AFL periodicals to spread the idea.

A cross-class alignment of interests is also evident in the fact that the National Civic Federation “brought leaders in both business and labor together and to agree upon some program of constructive legislation which will help to ameliorate the conditions that are bound to follow through enforcement of the Anti-Trust laws.” Speaking here was NCF antitrust committee member Wheeler P. Bloodgood, an influential corporate and trade association attorney. Even NAM general counsel James Emery touted cross-class interests in relaxing anticartel enforcement, in efforts to get passage of something similar to the NRA, saying that the “inability to make cooperative agreements with respect to hours of production” had resulted in “cutthroat competition, under-payment of wages, and demoralization of industry.” We therefore do not have to doubt the truth, though perhaps the compassion, in NAM president Robert Lund's statement that industrial codes would offer “better protection of labor.” When the U.S. Chamber of Commerce trumpeted the NIRA as a “magna charta of industry and labor,” it was probably as much for shared regulatory motives as for the compromise inclusion of 7(a).¹³⁰

But the NRA was a train wreck from the standpoint of economic policy and politics. (Lack of state or administrative capacity, identified by Skocpol and Finegold, was not the problem.¹³¹

Finally, the gathering anti-NRA backlash forced the Roosevelt administration to assert more bureaucratic authority over and through code authorities, where businessmen and their lawyers initially had free reign. The wresting away of the NRA apparatus from businessmen meant a looming, potentially uncontrollable attack on capital's carefully protected entrepreneurial sovereignty. This growing threat, of course, helped erode capitalist support for the NRA across the board. Ultimately, many of the politically important firms that stood to gain the most, often very large segmentalists that were part of the initial coalition for the NRA, began to suffer profound doubts. It had been those same doubts that made even Gerard Swope hesitate to support anything but arms-length government control over sectoral self-governance. Even big oil firms, some of the most vocal enthusiasts, began to lose interest, having discovered that evolving state-level regulation, coordinated across the small number of oil states, was proving perfectly adequate without the risks of bureaucratic intervention brought on by political and economic factors irrelevant to oil.¹³³

By comparison, the SSA was far superior politically and economically to the NRA, even as it performed a similar regulatory function – hence its longevity. By using taxation instead of bureaucratic rule making for regulation, the SSA minimized threats to managerial and entrepreneurial sovereignty. (The remaining threat – fiscal indiscipline and rising taxes – was real, but less dangerous.) In the distributional and regulatory realm there was room for a robust cross-class alliance; across the control battlefield is where capital drew the line. Anxiety about capitalist prerogatives made otherwise NRA-friendly businessmen reluctant to resuscitate anything like the NRA after the Supreme Court put it out of everyone's misery; it did not mean they abandoned hope for national-level regulation.¹³⁴

In sum, enforcing uniform social insurance costs across the board would, as people like Swope, Harriman and many others acknowledged, including FDR advisor and NRA enthusiast Adolf Berle, had an anticipated effect similar to other regulatory controls. All three had contemplated including social insurance provisions in their early thinking about cartelist legislation. But it would not come with the same political disadvantages. It would not arouse consumer, agrarian, and other counterconstituencies to the same extent, if at all. It would not stir up the same internecine conflict within the business community and across code authorities that tormented the NRA.

Arms-length, indirect regulation through social insurance was also superior economically as well as politically. As Wagner recognized, it would not eliminate competition, only displace it from the chiseling of wages and sweating of workers onto a search for innovative efficiencies in the use of well-treated labor.¹³⁵