The Adoption of Workers' Compensation in the United States, 1900-1930

Price V. Fishback and Shawn Everett Kantor, University of Arizona and NBER

Nore: Please contact the authors for a copy of the paper upon which this abstract is based. This research has been supported by National Science Foundation Grant No. SBR-9223058. Responsibility for the interpretations presented here remains ours.

The adoption of workers' compensation during the Progressive Era altered the political- economic landscape of the United States in significant ways. Workers' compensation was the first widespread social insurance program, the first major tort reform, and the first major employer mandate of the 20th century. In the paper upon which this abstract is based, we examine the political economy of the origins of this major institutional change.

Workers' compensation was probably the most universally accepted of the economic, social, and political reforms proposed during the Progressive Era. Many scholars have recognized that one of the key reasons for the general success of the workers' compensation movement was that it was supported by large numbers of employers and workers (see Berkowitz and McQuaid 1988; Dodd 1936; Somers and Somers 1954). While we agree with the conclusions of these prior studies, our analyses of the gains and losses that various groups experienced as a result of the adoption of workers' compensation sheds new light on the origins of this important legislation. In this abstract, we discuss the traditional view of the adoption of the legislation and then explain why employers, workers, and insurance companies all stood to gain from workers' compensation.

If all of the major interest groups with a stake in workers' compensation expected to gain from the law's adoption, then what explains the legislation's enactment in the 1910s, and not many years earlier along with European countries? In the paper we trace out several trends at the national level that led various interest groups to switch their attention from minor reforms of the common law to a wholesale change to workers' compensation. We then test hypotheses about these trends with information on the timing of the introduction of workers' compensation across the 48 states between 1909 and 1930.

Reformers' Claims About Workers' Compensation

The introduction of workers' compensation shifted the rules governing workplace accident compensation away from the common law rules of negligence to a no-fault system. The change substantially increased both the percentage of workers receiving compensation and the amount they would have received under the common law. For example, under negligence liability, roughly 60% of the families of fatal accident victims received no compensation from the deceased's employer, while nearly all workers' families received compensation under workers' compensation. Moreover, among those families receiving positive sums, largely through settlements, the payouts were typically between 65 to 150% of a year's income under the negligence system. Workers' compensation, on the other hand, guaranteed death benefits in the range of two to four times annual income.

Social reformers argued that a switch to workers' compensation would lead to a series of additional benefits: Wasteful litigation and legal fees would be reduced because fault and the three defenses would no longer be subject to dispute. Overhead ratios for insurance would fall. Finally, by shifting the financial burden of the accident from the worker to the employer, employers would now have an incentive to monitor their more careless workers and institute accident prevention measures.

In hindsight, it is not clear that workers' compensation achieved all of these goals. Throughout the 1920s workers and reformers complained that workers' compensation benefits, although higher than under negligence liability, were still inadequate. Ceilings on weekly payments meant that many workers received only 30% of their lost earnings while injured and not the 50 to 67% figures promised when the laws were passed. In addition, total administrative costs for workplace accidents might not have necessarily fallen. The administrative costs per case may have fallen with the introduction of no-fault insurance, but the number of cases administered rose dramatically after the law was introduced. Further, since most cases under negligence liability were settled outside the courts, even the administrative costs per case might not have fallen much.

Finally, the evidence on improved accident prevention is mixed. Reformers held the naive view that there was little incentive for employers to prevent accidents under the negligence system. Yet, insurers used experience rating and inspections to reward employers with lower premiums for keeping a relatively safer workplace. Further, employers could potentially cut their wage bill by making their workplaces safer, thus reducing the risk premium they paid in wages (see Fishback and Kantor 1992). Reformers also failed to consider moral hazard problems. With workers potentially taking less care and employers taking more to avoid accidents, it is theoretically uncertain what will happen to overall accident rates. For example, Chelius (1976) found that fatal accidents in manufacturing declined as a result of the adoption of workers' compensation, while Fishback (1987) found that compensation legislation led to an increase in fatal accidents in the coal mining industry. In general, the traditional views of the gains from workers' compensation provide some insight into the law's adoption, but the gains that historians have emphasized were not as clear cut as they first appeared.

New Insights into the Gains from Workers' Compensation

By 1910 organized labor and major employer organizations fully supported the concept of workers' compensation. Social reformers saw the substantial rise in benefits under workers' compensation as a great boon for workers. But why did employers support such a major redistribution of income?

Lubove (1967) and Weinstein (1967) claim that employers supported the legislation as a means of buying labor peace, as a way to stem the tide of court rulings that increasingly favored injured workers, and as a way to reduce the costs of settling accident claims. Given the evidence presented in the next section, employers may have seen workers' compensation as a reasonable alternative to the increasingly burdensome and uncertain negligence system. None of these explanations seems completely satisfying because none seems to have offered large enough benefits to offset the overwhelming increase in post- accident compensation that employers accepted. What contemporary social reformers and modern scholars failed to recognize, however, is that increases in employer-mandated benefits often lead to wage declines that are large enough to fully pay for the increase in expected benefits. Fishback and Kantor (1995) analyzed wages from the coal mining, lumber, and construction industries in the early 20th century and found that nonunion workers essentially "bought" the more generous and more certain benefits mandated by workers' compensation laws through lower real wages. Union workers, on the other hand, experienced much smaller wage offsets.

That such a large fraction of the work force would have experienced wage offsets helps to explain employers' widespread willingness to embrace the idea of workers' compensation. What appeared to be a large-scale transfer of income from employer to worker was, in fact, largely illusory. If employers could anticipate that workers would pay for the increase in post- accident benefits, then they were more likely to favor a no-fault compensation system that was less acrimonious and more certain than negligence liability. Similarly, organized labor's diligent lobbying on behalf of workers' compensation is understandable given that union members experienced relatively small wage declines.

What is less clear is why nonunion workers, who constituted the majority of the labor force, also supported workers' compensation. After all, they could expect to fully pay for their new benefits in the form of lower wages. Workers would have had little desire to "buy" the higher accident benefits under workers' compensation if they could just as easily have used the risk premia in their old wages to purchase their own workplace accident insurance. A central question concerning the economic motivation for the adoption of workers' compensation, therefore, is the extent to which workers had access to their desired levels of private accident insurance around the turn of the century.

We argue that workers were largely unable to purchase the full amount of accident insurance that they desired (see Kantor and Fishback 1996). Although accident insurance represented the most direct way for workers to insure against occupational accident risk, the personal accident insurance business was very limited in the early 20th century. Only $18.8 million in accident premiums were collected by commercial insurance companies in 1911, compared with $564.7 million in whole life premiums and $750.9 million in industrial life premiums.

The accident insurance market might have been limited by the informational problems of insuring an individual worker's accident risk. With little information on the accident proneness of the individual, the insurance industry had to base insurance premiums on occupational averages. Such pricing would have led to adverse selection problems, as accident-prone workers would have purchased the insurance and more careful workers would not. Insurance companies could expect no help from employers in identifying accident-prone workers because negligence liability rules allowed employers to invoke the contributory negligence defense to avoid compensating careless workers. Thus, employers had less incentive to fire irresponsible workers or to impose restrictions on their behavior.

The standard means of reducing problems of adverse selection is to limit the amount of insurance the worker could buy or to establish pricing policies designed to discourage more accident- prone individuals. Accident insurers followed both practices. The Aetna Life Insurance Company imposed limits on the risks they would insure, setting death benefit maximums as low as $250 for coal miners, who faced the most dangerous working conditions in the early 20th century. Physicians, on the other hand, could insure their lives for up to $10,000 for accidental death. Further, accident insurance was noted for its high load factors. Even with the high loads, a number of companies writing accident insurance failed over the period 1917 to 1926, while the surviving stock companies suffered a slight underwriting loss (Kulp 1928, p. 576). The end result was that many workers were unable to purchase complete coverage, with some possibly shut out of the market altogether.

Workers, through unions and sometimes with the help of employers, tried to obtain insurance through union funds or establishment funds. These funds, which were nearly entirely funded by workers' contributions, expanded the range of insurance that workers could obtain. But the amount of coverage in the funds was typically small. In 1908, most death benefits in these funds were less than $100, and thus covered only burial expenses. Temporary disability benefit levels were typically fixed at between $3 and $7 per week and many funds limited the number of weeks of payment to 13 or less (U.S. Commissioner of Labor 1909).

Absent full insurance coverage, families had to rely on household mechanisms to insure against accident risk, such as saving. Saving was a relatively costly means of insurance. In an empirical analysis of saving behavior among households surveyed for the 1917-1919 Bureau of Labor Statistics Cost-of- Living study, we found that for every $1 increase in expected post-accident benefits when workers' compensation was introduced, workers reduced their annual saving by $1.50. This disproportionate reduction in precautionary saving indicates the costly nature of using saving to insure against accident risk. Thus, even though their wages might have fallen, risk-averse workers or workers rationed out of the accident insurance market might have benefited from workers' compensation because the laws provided them with insurance coverage against workplace accident risk that was difficult to obtain privately under negligence liability.

Insurance companies stood to gain from the passage of workers' compensation, as long as the states did not outlaw private insurance and establish compulsory state funds. Because of the adverse selection problems associated with selling individual accident insurance, insurers stood to gain if the law compelled employers to insure their entire payrolls. The rise in post-accident payments from employers to workers under workers' compensation meant that employers would purchase substantially larger amounts of insurance than they did under negligence liability. In fact, premiums collected by commercial insurance companies for workers' compensation insurance rose from zero in 1911 to $114 million in 1920, despite the presence of compulsory state funds in 6 states and competing state funds in 11 more. The $114 million rise more than offset a $41.5 million shortfall in employers' liability premiums between the actual level of $86 million in 1920 and a predicted level of $129.5 million based on the annual growth rate from 1905 to 1911, the years prior to the introduction of workers' compensation laws. Workers' compensation legislation clearly expanded the effective opportunities for writing insurance, which explains the insurance industry's general support for the legislation.

If employers, workers, and insurers stood to gain from workers' compensation, why was a governmental solution necessary? It seems reasonable to expect that firms would design ex ante contracts in which the worker waived his right to a negligence suit in the event of an accident in return for a guaranteed set of benefits, thereby making fault no longer an issue. The courts, however, did not recognize such contracts whereby a worker waived his right to sue before the hypothetical accident occurred.

In essence, workers' compensation laws allowed a form of contract between worker and employer that had been disallowed by common law decisions and statutes. The pre-accident nature of the workers' compensation contract was extremely important in expanding the amount of insurance that employers could offer to workers. Prior to workers' compensation, the courts allowed firms to offer relief contracts in which the worker received some compensation and then waived future negligence claims. Because this type of post-injury waiver still gave the worker the option of filing a negligence suit, relatively few employers established accident benefit funds where they were the primary contributor. In only 140 of 461 establishment funds examined by the U.S. Commissioner of Labor in 1908 (pp. 339, 538-553) did employers make contributions to the funds and in most cases their contributions were less than 1/3 of the levels contributed by workers. The employer gained little from contributing to the funds because he expanded the number of workers to whom he was paying benefits, while not removing the uncertainty of the negligence system. Injured workers who would have received nothing under the common law could now claim the guaranteed benefits from the employer's fund, while the workers with strong cases still could choose to go for the large court awards or settlements. By offering the guarantee of post-accident benefits in exchange for the worker's right to sue, employers were able to eliminate the uncertainties of enormous court awards. The resulting compromise provided workers with benefits that, on average, were substantially higher than those they expected to receive under negligence liability.

The Timing of Adoption

If the major groups stood to gain from workers' compensation, why was the introduction of workers' compensation delayed until the 1910s? There were several trends occurring during the early 1900s that led to increased attention to the issue of compensation for workplace accidents.

First, there appears to have been a rise in the share of employment in more dangerous industries between 1900 and 1910, as well as a possible increase in accident risk within industries. There was clearly more publicity about accident risk, as reported accidents increased sharply in most state and national reports. Second, employers in many states experienced an expansion in their liability through employer liability laws or through state court decisions. The greater uncertainty associated with the changes in liability led to a sharp increase in the number of cases reaching state supreme courts. The changing liability climate also led to substantial increases in the premiums paid for employer liability insurance.

Circa 1909, major employer groups began calling for a shift to a workers' compensation type system. Meanwhile membership in labor unions was expanding rapidly. By 1909 unions appeared dissatisfied with the results of the employer liability laws, and in that year the American Federation of Labor moved from advocating employer liability laws to strong support of workers' compensation.

We investigated the impact of the changes in liability rules, interest groups, and Progressive Era politics further by using probit analysis to examine the timing of the adoption of workers' compensation laws across the 48 states between 1909 and 1930. The results of the analysis largely support the qualitative descriptions above. The changing climate of negligence liability clearly influenced the adoption of workers' compensation, as states with employer liability laws and higher employer liability insurance spending adopted workers' compensation relatively quickly.

The strength of various interest groups also affected the timing of adoption. States with an industrial mix that was more unionized and states with higher percentages of workers in manufacturing adopted earlier. Specific subgroups within manufacturing fostered adoption, as the presence of large firms and firms with high value added per worker enhanced the speed of adoption. Once a group of states adopted the law, there was a strong contagion effect, as the number of states that already adopted increased, the probability of adopting in other states increased substantially.

Finally, there is the question of the impact that political reform groups had on the adoption of workers' compensation. The regression results show that political power shifts in the state legislatures helped to speed adoption, but the size of the effects were small relative to the effects of the other variables in the equation. Strong political reform movements were not required to pass workers' compensation because it received widespread support. This statement is not true, however, for specific features of workers' compensation, like the issue of state versus private insurance. Political reform groups played a strong role in enacting monopoly state insurance in a handful of states. Insurers and unions fought bitterly over the issue of state insurance, and unions were strongly helped in achieving their goal of state insurance when political reformers gained strength in state legislatures (see Fishback and Kantor 1994).

Concluding Remarks

Workers' compensation was one of the most universally accepted progressive era reform because it was supported by employers, workers, and insurers. Employers anticipated reducing the frictions of the common law negligence system without having to fully pay for the enhanced benefits provided to workers. Workers, who paid for the benefits through wage offsets, were better insured against accident risk than before and insurers expected to expand their business.

The timing of adoption of workers' compensation in the 1910s was determined by a combination of changes in the liability climate during the early 1900s and the pressures of a variety of interest groups. Workplace accident risk in the 1900s was increasing, while the courts and state legislatures were undermining employers' defenses against injured workers' claims for damages. Litigation soared, liability insurance premiums increased, and membership in labor unions rose rapidly. After the United States Congress enacted workers' compensation for federal workers in 1908, the momentum for reform across the country was strong among both employer groups and unions. By 1910 the major manufacturing trade organizations and the AF of L supported workers' compensation. In 1910 New York enacted the first serious workers' compensation law and in 1911 nine more states adopted, and by 1920 all but five states had established workers' compensation.

Of social reformers' radical proposals for alleviating the financial stresses associated with working in a modern, industrializing economy, workers' compensation received the least objection. While unemployment, health, and old-age insurance were championed by elite reformers, workers' compensation received broad support from a wide variety of disparate interests. Only personal injury lawyers, the group that stood to lose the most under a no-fault compensation plan, objected. If the general idea of shifting the tort rules governing workplace accident compensation did not generate much opposition, the actual details of the laws did. Our view of workers' compensation is that the main interest groups with a stake in the legislation all stood to benefit and the real battles were fought over how the laws would be written. How these fights were resolved politically determined how the benefits of reform were distributed among the various interest groups.


Berkowitz, Edward, and Kim McQuaid, Creating the Welfare State: The Political Economy of Twentieth-Century Reform (New York: Praeger, 1988).

Chelius, James R., "Liability for Industrial Accidents: A Comparison of Negligence and Strict Liability Systems," Journal of Legal Studies, 5 (June 1976), pp. 293-309.

Dodd, Walter F., Administration of Workmen's Compensation (New York: The Commonwealth Fund, 1936).

Fishback, Price V., "Liability Rules and Accident Prevention in the Workplace: Empirical Evidence from the Early Twentieth Century," Journal of Legal Studies, 16 (June 1987), pp. 305-328.

Fishback, Price V., and Shawn Everett Kantor, "'Square Deal' or Raw Deal? Market Compensation for Workplace Disamenities, 1884-1903," Journal of Economic History, 52 (December 1992), pp. 826-48.

Fishback, Price V., and Shawn Everett Kantor, "A Prelude to the Welfare State: Compulsory State Insurance and Workers' Compensation in Minnesota, Ohio, and Washington, 1911-1919," National Bureau of Economic Research Historical Working Paper No. 64, December 1994.

Fishback, Price V., and Shawn Everett Kantor, "Did Workers Pay for the Passage of Workers' Compensation Laws?" Quarterly Journal of Economics, 110 (August 1995), pp. 713-42.

Kantor, Shawn Everett, and Price V. Fishback, "Precautionary Saving, Insurance, and the Origins of Workers' Compensation," Journal of Political Economy, forthcoming April 1996.

Kulp, C. A., Casualty Insurance (New York: Ronald Press Company, 1928).

Lubove, Roy, "Workmen's Compensation and the Prerogatives of Voluntarism," Labor History, 8 (Fall 1967), pp. 254-279.

Somers, Herman Miles, and Anne Ramsay Somers, Workmen's Compensation: Prevention, Insurance, and Rehabilitation of Occupational Disability (New York: John Wiley & Sons, 1954).

U.S. Commissioner of Labor, Twenty-Third Annual Report: Workmen's Insurance and Benefit Funds in the United States (Washington, DC: GPO, 1909).

Weinstein, James, "Big Business and the Origins of Workmen's Compensation," Labor History, 8 (Spring 1967), pp. 156-174.