Trends in Accident Compensation Before and After the 1972 Burton Commission

Source: William A. Sundstrom. Injury Rates in Manufacturing, Mining, and Railroads: 1922–1970, Table Ba4742-4749, in Historical Statistics of the United States (Millennial ed. Susan L. Carter et al. eds., 2006).

Sources: Based on data collected and measures created by Samuel Allen. See generally Allen, supra note 45. The real wage is the national weekly average manufacturing wage. Real Manufacturing Earnings and Real Expected Benefits were deflated by the Consumer Price Index (2000=100).

One of the major problems faced by workers was a relatively slow rise in their benefit payments for lost earnings. Most states had weekly maximum benefits, and the benefits were stated in the statutes in “nominal” values that were not adjusted for inflation. Until the 1960s, no states included clauses to adjust for inflation, and only a handful of states did so in the 1960s. There is always inertia in changing laws. In the case of workers’ compensation, the employers, workers, and insurers all had a stake, and so it often took time to come to a new compromise value. As a result, the states often would not change their weekly maximums for several years. After 1933, inflation became a ­problem because it eroded the purchasing power of the weekly maximum benefits listed in the statutes. In Delaware, for example, the weekly maximum was $13.50 in 1933 dollars in 1933 and did not change until it was raised to $18 in 1939; because of inflation, the purchasing power of the $13.50 benefit fell by nearly eight percent between 1933 and 1939. Delaware changed its benefits in 1939, 1942, 1946, 1950, 1956, and 1972. Each delay contributed to an erosion of benefits until the next increase was adopted. The worst erosion occurred between 1956 and 1972, when the weekly maximum benefit stayed fixed at $60. When Delaware first introduced the $60 maximum in 1956, it was higher than the average weekly wage in the state. By 1959, Delaware had started to limit benefits to below two-thirds of the average weekly wage. By 1971, the weekly maximum limited benefits to only thirty-nine percent of the average weekly wage. The purchasing power of the benefits was thirty-three percent lower than it had been when the $60 maximum was first introduced in 1956. Roughly half of the states had time spans when the weekly maximum did not change for at least ten years, often during the period covering the 1930s and 1940s. Louisiana, South Carolina, and Texas each had two spans of ten years or longer.

One of the consequences of this process was that benefits tended to stagnate relative to wages. Figure 2 shows indexes (1970=100) for national averages for a measure of real workers’ compensation benefits, real earnings for manufacturing workers, and the ratio of benefits to earnings. The measures are turned into “real” inflation-adjusted values using the Consumer Price Index (2000=100). The construction of the benefits measure is described in Appendix A. The long-term trend for both real benefits and real earnings was generally positive between 1930 and 1970. The ratio of benefits to earnings started out relatively high in 1930 and rose through 1933 because of a strong deflation. As inflation developed and many states kept their nominal weekly maximum benefits the same, the ratio dropped after 1933 and then dropped rapidly during the inflation of World War I. The ratio then fluctuated quite a bit. The states’ patterns of waiting multiple years to change nominal weekly maximums while wages continued to rise led to the drops in the benefit ratio. The ratio would then spike upward again when the states adjusted the weekly maximums. The problems with the benefit levels not keeping up with increases in the cost of living for long periods of time were one of the major complaints about workers’ compensation that led to several of the recommendations made the Burton Commission.

III. The Burton Commission

The Burton Commission was formed as part of the Occupational Safety and Health Act of 1970. As Michael Duff describes in his article in this issue, the Commission made nineteen essential recommendations for the workers’ compensation system. The recommendations that directly related to benefits included a call for full coverage for work-­related diseases. Three of the recommendations called for maximum weekly benefits for three injury categories—total temporary disability, permanent total disability, and deaths—to be at least two-thirds of the state’s average weekly wage with a suggestion that by July 1, 1975, the maximum be at least 100% of the state’s weekly wage. The Commission called for total disability benefits to be paid for the duration of the worker’s disability or for life with no limits as to dollar amount or time. It also recommended that death benefits be paid until the spouse remarried, with an additional lump sum payment of two years of benefits at the time of marriage; further, the children should receive benefits until age eighteen or to twenty-five if they remained a dependent and were attending school. Finally, the Commission called for no statutory limit on time or dollar amounts for medical care or physical rehabilitation services.

In our view, the Commission was extremely successful at obtaining adoption of its recommendations relative to most federal commissions, which often tend to be ignored. In 1972, the average number of recommendations the states were following was 6.8 out of 19. By 1980, the average number followed had nearly doubled to 12.1. The adoption of the recommendations then stalled after 1980, such that the average number was 12.8 by 2004. Our sense is that the most powerful recommendations adopted were the calls for maximum weekly benefits to be tied to the state’s weekly wage. This “indexing” of the weekly maximum benefit allowed for automatic adjustments associated with inflation of wages and prices, and thus avoided the problems with inertia and interest group negotiations that kept the states from updating weekly maximums before 1970. The first state to index was Connecticut in 1960. By 1970, ten states were indexing. By 1980, eight years after the Burton Commission Report was published, over forty states were indexing. As of 2000, all states except Alaska, Arizona, California, Georgia, Indiana, Minnesota, and New York were indexing. Arizona and New York then adopted indexing in 2007, and it was then reintroduced by California in 2007 and Minnesota in 2013. In contrast to workers’ compensation, state unemployment insurance (UI) benefit maximums have been indexed in only about thirty states, and UI benefits have fallen behind wage and price growth to a much greater extent.

You can see the importance of indexing in the national averages in Figure 2. The index of real expected workers’ compensation benefits and the index of the ratio of benefits to manufacturing earnings jumped from 100 to 160 in the decade between 1970 and 1980. This increase means that real workers’ compensation benefits were sixty percent higher in 1980 than they were in 1970. Note that the real earnings indexed stayed the same—around 100—throughout that decade; thus workers’ compensation benefits were replacing sixty percent more of wages lost due to injury in 1980 than they had been in 1970. Both measures rose to around 180% of the 1970 measure over the next two decades, while a few more states joined in indexing. Given that the central goal of workers’ compensation is to provide aid to injured workers, the adoption of the Burton Commission’s recommendations to have weekly maximums move with the state’s weekly wage by themselves led to substantial enhancements in the programs.

IV. The Twenty-First Century

The trend toward following the Burton Commission recommendations began moving in the opposite direction around 2004, as eleven states reduced the number of recommendations they were following after that year. The backtracking was stimulated by employer complaints about the acceleration in workers’ compensation costs per injury during the 1990s. There were three main reasons for the rise. First, workers’ compensation medical costs were rising as fast or faster than the rapid five to six percent annual rise in general health care inflation between 1985 and 2000. As a result, the medical share of cash and medical costs rose from twenty-nine percent of workers’ compensation benefits in 1980 to forty-five percent circa 2000 and to fifty percent by 2010. Second, the move to indexing weekly maximum cash benefits meant that the nominal statutory benefits paid for each injury rose automatically at an average of around three percent per year. Third, increasing complaints about the procedural rules tended to give workers the benefit of the doubt in determining eligibility for benefits. Under the liberal construction rule, any ambiguity in workers’ compensation statutes or evidentiary uncertainty was to be resolved in favor of workers. Legal professionals claimed that abusive applications of the liberal construction rule exacerbated workers’ moral hazard and drove the state workers’ compensation system near to bankruptcy.

In trying to limit costs, employers were unlikely to be successful at getting others to agree to stop indexing of weekly maximum benefits. Once established, it is often difficult to change major features of programs because of the large number of stakeholders in those features. It would have been particularly difficult to stop the indexing because it would likely lead to a significant reduction in the growth in benefits. Employers found more success in following the health insurance strategies of setting medical fees and establishing physician networks, while challenging procedural rules related to burdens of proof that were more subtle in their impact and largely influenced only the disputed injury cases. By 2016, thirty-nine of the states had passed laws setting up medical fee schedules, nine had established physician networks, and over thirty had passed laws changing procedural rules.

A. What Caused the Downward Trends in Benefits Per Covered Workers

Workers’ compensation practitioners, scholars, and the press have expressed concerns that the Grand Bargain of workers’ compensation has been eroding since the early 2000s. This view has been heavily influenced by two types of information. The first is the decline in the inflation-adjusted national averages of cash and medical benefits per covered worker, as shown in Figure 3. The second is an index of workers’ compensation laws for each year between 2002 and 2014 created by Yue Qiu and Michael Grabell in 2015. In an online article at the ProPublica website that was based on the index, Grabell and Berkes concluded that many states had “slashed workers’ compensation benefits,” and were in a “race to the bottom.”

Figure 3 shows the distribution of funds for workers’ compensation divided by the number of people who are covered by workers’ compensation. The wage replacement benefits (nonmedical) are adjusted for inflation using the Consumer Price Index (CPI), and the medical benefits are adjusted for the specific CPI for medical treatments, which has risen faster than the overall CPI for several decades.

The benefits per covered worker experienced quite a reversal after the mid 1990s. Before the period shown on Figure 3, wage replacement benefits per covered worker had roughly doubled between 1940 and 1972, the year of the Burton Commission, and then doubled again through the mid 1990s after the states adopted many of the Commission’s recommendations. Similarly, real medical expenditures per covered worker rose about ten percent between 1940 and 1971 and then doubled between 1971 and 1992.

Since the mid-1990s, however, Figure 3 shows that both types of benefits per covered worker have declined. Between 1997 and 2016, the years when we could get detailed state-by-year data, the wage replacement benefits per covered worker have fallen thirty percent, and the medical benefits have fallen twenty-three percent. Why?

Sources and Notes: These are national totals based on state-by-year level data of cash benefit payments, medical benefit payments and the number of covered workers. The information comes from annual National Association of Social Insurance reports titled Workers’ Compensation Benefits, Coverage, and Costs from 1997 through 2016. The reports can be downloaded from Workers’ Compensation Research, Nat’l Acad. Soc. Ins., https://www.nasi.org/research/workers-compensation. To obtain measures in 2010 U.S. dollars, we deflate the cash benefits by the national Consumer Price Index for urban areas and medical benefits by the national Consumer Price Index for medical care. The Consumer Price Index data comes from the U.S. Bureau of Labor Statistics and can be downloaded from Consumer Price Index (CPI)  Databases, U.S. Bureau of Lab. Stats., https://www.bls.gov/cpi/data.htm.

We sought to answer this question in a 2021 study. The wage replacement benefits per covered worker (BCW) can be broken roughly into parts.

1) BCW =
Benefits When Injured * Number of Injuries
Number of Covered Workers.

The Injury Rate is defined as the Number of Injuries divided by the Number of Covered Workers. After substituting the Injury Rate for the number of injuries per covered worker, equation 1 can be rewritten as

2) BCW = Benefits When Injured * Injury Rate.

Benefits When Injured are determined by three major factors: statutory benefits, administrative procedures, and other factors.

3) Benefits When Injured = Statutory Benefits * Admin * Other.

The Statutory Benefits are determined by the benefit parameters described in the state statute. They are filtered through the administrative process, which we will designate as Admin, which takes a value of one if the administrative process awards the statutory benefits to the injured worker, less than one if the actual benefits paid are less than the statutory benefits, and greater than one if actual benefits are higher. The term Other takes into account other factors or mismeasurement that may influence benefit per covered workers (BCW) in ways that we do not know.

After substituting for the Benefits When Injured in equation 2 with the formula in equation 3 and rearranging terms, the benefits per covered worker equation can be rewritten as

4) BCW = Admin * Statutory Benefits * Injury Rate * Other.

Each of these four features might have contributed to the decline in benefits per covered worker.

In our study of this issue, we collected information from 1997 to 2016 for each state on the statutory benefits and the injury rate, as well as several factors that influenced the administrative process. The results of the study show that the statutory wage replacement benefits rose during this period; therefore, the changes in statutory benefits could not have been the cause of the decline in actual benefits per covered workers seen in Figure 3 because they moved in the opposite direction. Over the same period, injury rates fell substantially, which has been a major benefit for workers, even though it has caused a decline in benefits per covered worker in Figure 3. At the same time, the people who were worried about changes in the administrative process were correct because there is evidence that changes in the administration of the law have served to reduce benefits per covered worker in a way harmful to workers.

To measure the statutory benefits, we developed an expected benefits index like the ones in Figure 2 and the ones developed by Samuel Allen for 1940 through 2000 and by one of us, Price Fishback, and Shawn Kantor for the years 1911 to 1940. This summary measure of benefits captures how each of the legislative changes influence the benefits being paid for each type of injury. In the process, it directly incorporates how the benefits are influenced by the annual indexing of weekly maximums to wages and prices. Figure 4 shows that the national average of the benefits in 2010 dollars rose between 1997 and 2016 by about eighteen percent. This would be expected because nearly all of the states are indexing their benefits and thus are at least keeping pace with wage and price inflation. The rise reflects some additional improvements in other parameters in the state statutes, including increases in the maximum total payments for death benefits, increases in the number of weeks the benefits are paid for certain types of accidents, enhanced payments for funeral expenses, and a few other factors.

One of the most salutary labor market trends over the past twenty years has been the substantial drop in the number of injuries and deaths. The national average of the workplace fatality rate, as shown in Figure 5, fell by about one-third between 1997 and 2016. This decline is one of the reasons for the declines in benefits per covered worker seen in Figure 4. Fatalities are rare relative to other types of injuries. The national average for lost days due to illness or workplace injury fell from 3.3 per 100 full-time workers in 1997 to 2.8 in 2001, a decline of 15.2%. At that time, OSHA changed their record-keeping requirements. The subsequent series shows that lost workdays per 100 full-time workers fell from 2.8 in 1994 to about 1 in 2014, a decline of over sixty percent. Therefore, a major reason for the decline in benefits per worker was the result of a reduction of the risk of accidents faced by workers.

Source: Calculated from dataset of expected benefits by state and year from 1997 to 2016 from state statutes for the paper by Andy Yuan & Price Fishback. The Rising Burdens of Proofs and the Grand Bargain of Workers’ Compensation Laws (Nat’l Bureau of Econ. Rsch. Working Paper No. 26980 (2021)).

Sources and Notes: The fatal accident rate is defined as the number of fatal accidents per 100,000 workers in private industry. The measure is calculated with the following two strands of state-level data from the Bureau of Labor Statistics: seasonally adjusted employment and total counts of reported occupational fatalities.

B. Changes in Administrative Procedure and the Decline in Benefits Per Covered Worker

Thus far, we have shown that the trends in two of the four components of determining benefits per covered worker are beneficial to workers. It is certainly possible, however, that changes in the administrative processes might have had a negative impact on how the workers’ compensation law was applied to provide benefits to workers. Workers’ compensation practitioners, scholars, and the press have expressed concerns that changes in state procedural laws might have lowered the probability of receiving any compensation and/or prevented workers from getting full statutory compensation for injuries. A Department of Labor report suggested that such changes “ had enormous, though perhaps more hidden, impact on injured workers’ access to benefits.”

To examine these types of administrative procedure changes, we wrote a paper in 2022 in which we started with the Qi and Grabell collection of workers’ compensation laws from 2004 to 2012 and expanded the information to include the situation in place in 1997 and to cover the period 1997 through 2016. We built a panel data set with the wage replacement benefits per covered worker and the medical benefits per covered worker as dependent variables. While controlling for the expected benefits index and fatal accident rate in each state, we developed several types of panel data analyses of the impacts of the introduction of five types of laws. Three of the laws can be categorized as burden of proof legislation: (1) laws that prohibit the use of the liberal construction rule; (2) laws that presume intoxication was the proximate cause of injuries to an intoxicated worker; and (3) laws that apportion benefits based on a worker’s preexisting conditions. Two were legal changes that primarily focus on reducing medical costs: (1) laws establishing medical fee schedules; and (2) laws establishing statewide networks of physicians to treat work-related injuries.

Among the burden of proof laws, the “liberal construction rule,” as mentioned above, interprets the workers’ compensation statute by presuming that workers prevail by default if there are ambiguities related to workers’ eligibility for compensation benefits. The countervailing doctrine is the “strict construction rule,” which requires adjudicators to resolve disputes solely based on the evidentiary merits of the claims and opposes presuming either party as the default winner. On the one hand, the most commonly seen justification favoring the liberal construction rule is the idea that workers’ compensation laws are remedial and designed to aid injured workers and their families. On the other hand, states prohibiting the liberal construction rule argue that workers’ compensation is not remedial legislation. Instead, they claim that the goal of workers’ compensation laws is to strike a balance between the workers’ interests in obtaining benefits and the employers’ interests in maintaining reasonably low costs of compensating work injuries.

Prior to 1997, the beginning of our sample period, seven states listed in Table 1 prohibited the use of the liberal construction doctrine in resolving workers’ compensation disputes. But in subsequent years, several other states adopted liberal construction bans: West Virginia in 2003, Louisiana and Mississippi in 2012, and Tennessee in 2013.

In most of the states, intoxication is a statutory defense that might be invoked by the employer to reduce or eliminate compensation for an intoxicated worker injured on the job. A worker’s intoxication could significantly undermine the claim that the injuries suffered were caused by the underlying risk of the employment. If the law presumes intoxication as the proximate cause of work-related injuries, the injured worker bears the burden to show that their injuries are independent from the intoxication. If such a presumption is not allowed, the employer bears the burden of establishing that intoxication was the proximate cause of the injury to avoid compensation liability. Prior to 1997, seven states, as shown in in Table 1, allowed the presumption of workers’ intoxication as the proximate cause of injuries. Between 1997 and 2016, Missouri (2006), Illinois (2011), Mississippi (2012), and Nevada (2015) added new language allowing such a presumption.

Preexisting conditions can potentially prevent injured workers from recovering the full amount of compensation benefits. Workers with preexisting health conditions are typically eligible for compensation benefits only if the job-related injuries significantly aggravate the intensity of the preexisting conditions. Injured workers are not eligible for compensation benefits if their symptoms are a “mere natural progression” of their preexisting conditions. Once “significant aggravation” of preexisting conditions is established, states have adopted one of two compensation regimes. About half of the states pay workers the full amount, which raises the financial burden on employers. To mitigate the financial burdens and to provide more employment opportunities to workers with preexisting conditions, some states have established “second injury funds,” which partially cover compensation benefits payments to workers with preexisting conditions.

Table 1

Timing of States’ Passages of Revisions of Workers Compensation Laws

Source: This summary is based on material collected for 2002 to 2014 by Yue and Grabell, supra note 63, and reconfirmed and supplemented with our own research, using the Westlaw database. 

The rest of the states pay workers the fraction of benefits that matches the fraction of the condition that was solely induced by the workplace accident. These states usually only apportion permanent disability benefits. Workers can usually recover the full amount of temporary disability benefits, regardless of the status of their preexisting conditions. Before 1997, there were twenty-one states in Table 1 with laws that apportioned benefits when workers had preexisting conditions. Between 1997 and 2016, California (2004), Kansas (2011), and Missouri (2006) joined the group.

The rapid increases in medical costs over the last forty years have directly affected the medical costs of treating workplace injuries under workers’ compensation. As a result, states have explored a variety of ways to control costs while trying to maintain appropriate medical treatment of injuries. As seen in Table 1, the two most common forms of medical legislation are laws that alter medical fee schedules and laws that establish statewide networks of physicians to treat work-related injuries.

Adoption of a medical fee schedule puts a limit on the workers’ compensation payment to medical care providers for providing injured workers with various medical services. The fee schedule is typically adopted by the state agency in charge of administering the workers’ compensation program. Prior to 1997, twenty-four states, as noted in Table 1, had adopted fee schedule systems for medical payments. Between 1997 and 2016, an additional fifteen states adopted fee schedule provisions.

Employers and workers have competing self-interests in determining providers of medical care. Employers want to treat workers’ injuries in a way that minimizes the medical costs. In contrast, if the medical services are covered, workers tend to have their injuries treated with the “best-quality” medical care and are less concerned about the costs. State laws have developed various systems to account for the legitimate interests of both employers and workers.

In some states, workers play a major role in determining medical care, although their choices are usually subject to some restrictions. In other states, medical care is mainly determined by employers. Physician networks normally consist of doctors specialized in treating occupational diseases or general fields of medicine. To join the network, medical care providers typically need to apply and get approved by the state agency in charge of administering the workers’ compensation programs. Employers may choose to join the network for treating workers’ work-related injuries, but this is not a mandate. If the employers are in the network, the physician choices of injured workers are constrained to choose a physician in the approved network. Prior to 1997, four states in Table 1 had authorized and regulated physician networks. Between 1997 and 2016, four more states passed the laws.

Our econometric analysis in the 2022 paper found that liberal construction bans accounted for about 5% of the national drop in wage replacement benefits per covered worker between 1997 and 2016, apportionment laws accounted for about 2.6%, and physician network laws accounted for about 1.5% of the drop. Of the drop in the national average of medical payments per covered worker, the liberal construction bans accounted for 2.9%, intoxication laws accounted for 1%, apportionment laws accounted for 0.5%, and the physician network laws accounted for 6.1%. The laws were associated with drops in benefits per covered worker of as much as 20% in the states where they were adopted, but there were only a handful of states that adopted each law, which muted their impact on the overall national drops in wage replacement and medical benefits per covered worker.

Conclusion

The original workers’ compensation laws were adopted by state governments between 1911 and 1948. A significant share of people among employers, workers, and reformers anticipated benefits from adopting the laws. The shift to strict liability meant that workers injured on the job were essentially guaranteed to receive compensation. When the new laws were adopted, the average amounts of compensation for each type of injury appears to have been higher than the amounts paid to recipients of benefits under the prior negligence liability regime.

Between the early days of workers’ compensation and the early 1970s, some states expanded the coverage to include a limited number of workplace diseases and included a larger share of the workforce. A major problem faced by workers was that states adopted nominal weekly maximum benefits in their statutes and then often updated the values only after several years. Those weekly maximums were often binding and led to wage replacement benefits of less than the fifty percent to two-thirds of earnings that the state was claiming to provide. Between updates, the purchasing power of the benefits was almost always eroded by inflation.

Seeing this problem and rising trends in manufacturing accident rates and a flat trend in mining accident rates, the federal government adopted the Occupational Safety and Health Act of 1970 and created the Burton Commission to examine workers’ compensation programs in the states and to recommend reforms. Of the nineteen essential reforms they proposed, many had to do with benefit levels. Some of the most powerful recommendations called for the states to tie their weekly maximums to state average wages, and nearly all states eventually adopted this practice. This indexing of benefits sharply reduced the problems caused by delays in raising nominal benefits before 1970. The ratio of expected benefits to earnings rose sharply as the states adopted indexing and some of the other recommendations of the Commission.

Since the mid-1990s, many workers’ compensation experts have suggested that the Grand Bargain of workers’ compensation has been eroding. At first blush, the thirty percent decline in wage replacement benefits per covered worker and the twenty-three percent decline in medical benefits per covered worker between 1997 and 2016 seem consistent with that claim. When the components of the calculations of the benefits per covered workers are examined, however, the situation does not look nearly so dire. The national average of real expected benefits based on the statutes actually rose by roughly eighteen percent, which was a benefit to workers. A main reason for the drop in benefits per covered worker were drops of around thirty-three percent in fatality rates and over forty percent in lost workdays due to injury or illness, which are also beneficial to workers.

Workers may have to worry, however, given changes in the administrative procedures that have contributed to lower benefits per covered worker. A number of states have adopted bans of liberal construction and increased use of apportionment of benefits that shift the burden of proof in disputed cases to workers. A number of states have also adopted physician network laws that have contributed to declines in benefits per worker. These changes account for roughly nine percent of the decline in national wage replacement benefits per covered worker and a roughly similar percentage of the decline in medical benefits per covered worker.

Appendix A

Calculating the Expected Benefit Index in Figures 2 and 4.

The weekly maximum is an important component of the benefit structure in each state but not the only one. The benefit structure is also influenced by how many weeks of benefits are offered for each type of injury and fatality, waiting periods before benefits are paid for temporary disabilities, payments for funeral expenses, and several other features. To develop a summary measure of the benefit structures across the various types of accidents, Samuel K. Allen in 2015, Shawn Kantor and Price Fishback in 1995, and we in 2022 have created an “expected benefits index” for each state and year since the introduction of workers’ compensation. The state laws generally stipulate different payment schemes for various accident outcomes. We divided injuries into four types: death, permanent total disability, permanent partial disability, and temporary total disability. With a few exceptions, the benefits payment schedules specify a weekly payment duration and a weekly payment amount that is a percentage of usual weekly earnings up to a weekly maximum and not below a weekly minimum. Weekly cash benefits are determined by multiplying the worker’s weekly wage by a replacement percentage that is specified in statutes, usually two-thirds.

For each accident type, they calculated the present value of the stream of weekly payments using a discount rate of five percent and the national average weekly wage for that year.³⁷ They chose the national average weekly wage to ensure that the variation in the payments across states was driven only by statutory differences in the state laws. They then calculated real present values in 2010 dollars using the Consumer Price Index. To get the summary measure across accident types, they then multiplied the real present values for each type of accident by the probability that type of accident would occur and then added up the values for the four types of accidents.

This summary measure is an expected benefit index that takes into account the probabilities of getting injured and the benefits paid when injured. In making comparisons across time and place, the accident rates are held constant, so that the comparisons focus on the difference in benefits.