Most Americans between the ages of 22 and 65 spend 40 to 50 percent of waking hours at work. Every year millions of Americans suffer injuries and thousands experience deaths in our workplaces. Yet little effort has been made to estimate either the extent of these injuries, deaths, and diseases or their cost to the economy. Thus, important questions about workplace safety and the economic resources expended due to workplace health problems remain unanswered. In this study, we address these questions by presenting estimates of the incidence, prevalence, and costs of workplace-related injuries, illnesses, and deaths for the entire civilian workforce of the United States in 1992. We also consider controversies surrounding cost methodologies, estimate how these costs are distributed across occupations, consider who pays the costs, and address some policy issues.
Our major findings are as follows.
- Roughly 6,371 job-related injury deaths, 13.3 million nonfatal injuries,
60,300 disease deaths, and 1,184,000 illnesses occurred in the
U.S. workplace in 1992 (see table 1.1).
- The total direct and indirect costs associated with these injuries and
illnesses were estimated to be $155.5 billion, or nearly 3 percent
of gross domestic product (GDP).
- Direct costs included medical expenses for hospitals, physicians, and
drugs, as well as health insurance administration costs, and
were estimated to be $51.8 billion.
- The indirect costs included loss of wages, costs of fringe benefits, and
loss of home production (e.g., child care provided by parent and
home repairs), as well as employer retraining and workplace
disruption costs, and were estimated to be $103.7 billion.
- Injuries generated roughly 85 percent whereas diseases generated 15
percent of all costs.
- These costs are large when compared to those for other diseases. The
costs are roughly five times the costs for AIDS, three times the
costs for Alzheimer's disease, more than the costs of arthritis,
nearly as great as the costs for cancer, and roughly 82 percent of
the costs of all circulatory (heart and stroke) diseases.
- Workers' compensation covered roughly 27 percent of all costs.
Taxpayers paid approximately 18 percent of these costs through
contributions to Medicare, Medicaid, and Social Security.
- Costs were borne by injured workers and their families, by all other
workers through lower wages, by firms through lower profits, and
by consumers through higher prices.
- Our study appears to be the first to use national data to produce
estimates on costs for occupational injuries and illnesses. Prior
studies have underestimated costs by ignoring nondisabling
injuries, deaths, and workplace violence, by taking inadequate
account of diseases, and, most importantly, by relying on only one
or two sources of data.
- The Annual Survey of the Bureau of Labor Statistics (BLS) provides the
most reliable and comprehensive data on nonfatal injuries.
However, it misses roughly 53 percent of job-related injuries. This
omission, in part, is due to the exclusion of government
employees and the self-employed and also, in part, due to illegal
underreporting by private firms.
- Contrary to the Annual Survey data, we find small firms have
exceptionally high injury rates.
- Occupations contributing the most to costs included truck drivers,
laborers, janitors, nursing orderlies, assemblers, and carpenters.
On a per capita basis, lumberjacks, laborers, millwrights, prison
guards, and meatcutters contributed the most to costs.
- Occupations at highest risk for carpal tunnel syndrome include dental
hygienists, meatcutters, sewing machine operators, and
assemblers. Among well-paid professions, dentists face the highest
- Any of the major sources of data, such as the Bureau of Labor Statistics,
National Institute for Occupational Safety and Health, workers'
compensation systems, or National Health Interview Survey, by
themselves underestimate the numbers of injuries and illnesses.
- Greater efforts need to be directed toward gathering data on job-related
injuries and illnesses. The United States needs a comprehensive
data bank for fatal and nonfatal injuries and all illnesses. Future
researchers should not have to investigate the over 20 sources of
primary data and 300 sources of secondary data that we
TABLE 1.1 Number and Costs of Injuries and Illnesses in 1992
Costs (in $billions)
Source: Current study.
aMay not sum due to rounding.
bThe number of deaths and morbidity for illnesses cannot be summed precisely.
These costs are great, but the reason for their size is no mystery. Roughly 120 million of us worked in 1992. Every job carries some risks (Leigh 1995a). Many of us are exposed to job-related safety risks of traffic accidents, falls, murder, electrocution, fire, being struck by objects, explosion, heat, cold, animal attacks, and airplane crashes, as well as health risks from radiation, asbestos, silica, benzene, coal dust, tuberculosis, secondhand smoke, carbon monoxide, pesticides, benzidine, arsenic, lead, chromium, and stress.
The estimates are the result of an exhaustive compilation of data from a variety of sources. Chapters 2 through 6 present a detailed account of our methodology and estimates. In developing the estimates, we most frequently selected conservative rather than generous assumptions. The assumptions with greatest consequences are listed in appendix B for chapter 10. Here we mention four. First, with 7.4 percent of the workforce unemployed, 1992 was a high unemployment year. When fewer people are employed, fewer job-related injuries and diseases occur. Second, we did not account for health effects of occupational injuries and illnesses on the relatives of victims, or, more importantly, for the cost of caregivers' time and energy (Arno, Levine, and Memmott 1999). After a serious injury or disease, someone in the family frequently provides care. Third, we restricted job-related circulatory disease deaths to people under 65 years old. It could be argued that jobs have a cumulative effect on circulatory disease that becomes evident only during retirement. Finally, our Human Capital method of estimating costs ignored costs of pain and suffering. These costs would add at least an additional $350 billion to our overall $155.5 billion estimate. ...
II. Number of Injuries
Major general findings are listed in the following.
- We estimate that 6,371 deaths and 13.34 million new nonfatal injuries
occurred in 1992.
- Disabling injuries accounted for 5.326 million of these injuries, and
nondisabling injuries accounted for 8.011 million. Disabling means
that the injury resulted in at least one day of work loss, whereas
nondisabling means no full days of work loss.
- Within the disabling category, there are several subcategories. We relied
on the workers' compensation (WC) categories: Permanent Total
(PT), Permanent Partial (PP), and Temporary Total and Partial
(TTP). We estimated 12,124 PTs, 741,000 PPs, and 1,947,000
- No one source of data is sufficient to estimate deaths or nonfatal injuries.
The National Safety Council omitted violent acts. The Rand study
by Hensler et al. (1991) omitted deaths. The National Traumatic
Occupational Fatality Study relied solely on death certificates. The
Census of Fatal Occupational Injuries (CFOI) may have resulted in
an undercount because of the strict two source requirement. The
BLS's Annual Survey underestimated injuries from small firms. All
other sources had additional problems.
- Econometric time-series models using the National Health Interview
Survey (NHIS) data as well as NHIS data on black/white injury
rates suggest that the NHIS data may not be as reliable as is
- Workers' compensation records underestimate the number of injuries by
The most important findings involving socioeconomic and geographic characteristics are listed in the following.
The most important findings pertaining to types of injuries are listed in the following.
- Disabling injuries are strongly correlated with job experience. New
employees, regardless of age, experience a high and
disproportionate number of injuries.
- Men are more likely than women to sustain a work injury. This is
especially true for an injury resulting in death The nonfatal injury
ratio for men to women is nearly 2:1, whereas the fatal injury ratio
is about 11:1.
- Blacks and Hispanics experience greater injury rates than non-Hispanic
- In 1992, the CFOI and the NHIS underestimate injuries experienced by
- The self-employed, persons employed in small firms, and persons over
age 65 are at high risk for sustaining an injury death.
- Laborers, truck drivers, and taxi drivers generate among the highest death
rates of all occupations.
- Mining, farming, and construction are the industries with the highest rates
of fatal and nonfatal injuries.
- Murder is the most likely cause of death for business executives and sales
- Operators and laborers generate the greatest numbers of deaths and
nonfatal injuries among all broad occupation groups.
- Laborers, truck drivers, nursing aides, janitors, assemblers, stock
handlers, and cashiers generate the most disabling injuries among
- Being at work is not safer than being at home. People who work are more
likely to be injured at work than at home. This is especially true for
men. Moreover, work-related injuries are more likely to result in
hospitalizations than injuries originating outside of work.
- Injuries to the back generate the highest frequency of disabling injuries.
- Recall bias on questions asking for incidents dating back 12 months may
result in a serious undercount of nondisabling injuries.
- Transportation accidents involving highway vehicles, industrial vehicles,
and aircraft boats and railroads contribute to 40 percent of injury
deaths. Transportation accidents have frequently been ignored by
the Occupational Safety and Health Administration (OSHA).
- Assaults and violent acts contribute another 20 percent of injury deaths.
These, too, have frequently been ignored by OSHA.
- Transportation accidents, assaults, and violent acts comprise a smaller
share of nonfatal injuries than fatal injuries. Assaults and violent
acts are more likely to be fatal than most other injuries at work.
The numbers of deaths and nonfatal injuries were estimated after considering five primary sources and four secondary sources. The primary sources included the BLS Census of Fatal Occupational Injuries (CFOI), the BLS Annual Survey of Occupational Injuries and Illnesses (Annual Survey), the Ultimate Reports of the National Council on Compensation Insurance (NCCI), the National Health Interview Survey (NHIS), the National Traumatic Occupational Fatalities Study (NTOF), and the BLS's Supplementary Data System. Secondary sources included studies by Hensler et al. (1991), Rossman, Miller, and Douglas (1991), Miller (1994), and the National Safety Council (1992, 1993). These data have strengths and weaknesses. The BLS's CFOI and Annual Survey data were regarded as the best data, and our estimates were ultimately derived only from them. ...
V. Costs of Injuries
- Direct costs comprise 29 percent, and indirect costs 71 percent, of total
- Within the direct cost category, medical only costs are roughly $26 billion
(68 percent), medical insurance administration costs are $5.5 billion
(14 percent), and indemnity insurance administration costs are $6.8
billion (18 percent).
- Within the indirect cost category, lost earnings summed to $67 billion (71
percent); fringe benefits, $15.7 billion (17 percent); home
production, $9.3 billion (10 percent); and workplace training,
restaffing, and disruption, $2.2 billion (2 percent).
- Fatality costs comprised only roughly 3 percent of the total. Sensitivity
analysis that would have altered interest rates for present value
calculations would not have appreciably affected our results.
- Insurance administration costs have frequently been omitted from prior
cost studies. This is a mistake. Insurance administration costs (for
both medical and indemnity insurance) are significant, comprising
32 percent of direct costs.
Estimation of the costs of injuries required multiplying the number of injuries in each category by the average costs of such injuries. Direct average costs for medical care were drawn from the National Council on Compensation Insurance Ultimate Reports. Lifetime medical costs (1992 dollars) for deaths were valued at $17,226; for Permanent Total at $113,372; for Permanent Partial at $15,342; for Temporary Total and Partial at $2,782; and for no work loss at $294. The medical expenses were drawn from workers' compensation accounts and did not require adjustment for charges versus payments since workers' compensation paid virtually 100 percent of medical bills in 1992; that is, very few co-payments or deductibles were charged to clients.
The calculation of the indirect costs was based on a variety of sources, including National Council's indemnity data and federal government data on employment, earnings, and mortality. Home production costs, as well as hiring, training, and workplace disruption costs, were priced in accord with estimates in the literature. Indirect costs for fatalities required a present value calculation. We assumed that persons who died would have earned what others of the same age and gender earned. The distribution of deaths by age and sex was estimated with information from the CFOI. These age and sex data were combined with information on wages and on probabilities of survival to age 75, as well as on the employment within those categories.
The National Council figures also provided us with indemnity benefits that were used to estimate wage loss. The indemnity benefits themselves were not added to wage losses. The indemnity benefits were adjusted assuming workers' compensation paid to clients the following rates: 40 percent of pretax wages for Permanent Total conditions; 50 percent for Permanent Partial conditions; and 60 percent for Temporary Total and Partial conditions. Fringe benefits were assumed to be 23 percent of the pretax wages for men and women combined.
Insurance administration costs were assumed to be 31 percent for workers' compensation and 15 percent for all others. ...
VII. Workers' Compensation Costs across Occupations
- The public is frequently misinformed about job hazards. Most of the high
cost per person jobs, such as production helpers, laborers, janitors,
nursing orderlies, sales workers who drive on the job, truck drivers,
polishing machine operators, kitchen machine operators,
assemblers, and others, are not generally regarded as dangerous
by the public.
- Many of the most costly occupations are not well described by U.S.
Census categories but appear to occupy the lowest status
categories, for example, laborers, miscellaneous machine
operators, freight handlers (not elsewhere classified), production
helpers, construction helpers, and miscellaneous food preparation
- The cost per person lists reinforce the view that the most hazardous jobs
enjoy the least pay. Occupations within the laborer and operative
categories receive the lowest pay of all occupation groups but
generate among the highest costs.
- Jobs that are high on both the total and per person cost lists include truck
drivers, laborers (inside and outside of construction), janitors and
cleaners, nurses aides, assemblers, carpenters, miscellaneous
food preparation occupations, timber cutters, electricians, welders,
bus drivers, police officers, and firefighters. Jobs that are high on
both lists should be candidates for greater attention from
occupational safety and health regulators and researchers.
This chapter uses exclusively workers' compensation (WC) data to rank occupations by costs. Data were drawn from a large national representative BLS data set -- the Supplementary Data System. Information was obtained on occupations and WC category of injury and illness and was then matched to information on costs. Six broad occupations were ranked by total costs. Six broad and 223 specific occupations were ranked by costs per person (average costs). Unlike cost data in all other analyses of the book, these rankings applied to 1985 and 1986, not 1992.
VIII. Who Pays?
- Using the nominal payment method, we found that injured or ill workers
and their families absorbed about 44 percent of the costs.
Medicare, Medicaid, Social Security and other government
accounts contributed 18 percent, or roughly $28.5 billion.
- Using the incidence payment method, we found employers absorbing
some noninjury costs in terms of lower profits, consumers
absorbing some in terms of higher prices, and all workers
absorbing some in terms of lower wages.
There are two methods for assessing who pays, the nominal method and the incidence method. The nominal method considers who writes the check. The incidence method uses economic theory to assess the burden. For example, the business owner writes the WC premium payment check to the insurance company. But the owner may try to pass on the cost of that premium to the consumer in terms of higher prices. There is considerable controversy surrounding how much employers, consumers, and workers pay in the incidence method, however. We therefore prefer the nominal over the incidence method for assessing the cost burden of job-related injuries and illnesses.
IX. Policy and Cost Comparisons
- One policy option would be to provide more information to workers
pertaining to the hazards of their jobs. A report card could be
prepared by the BLS that would rank and compare occupations and
industries across the United States. The report card could be
attached to every job application form.
- We suggest that a general occupational injury and illness tax be levied on
all employers to pay for the substantial amount of costs that is
currently being shifted to taxpayers and the general public. This tax
could be modeled on the Federal Black Lung Trust fund that taxes
all coal companies on a per tonnage amount to pay for the medical
costs of pneumoconioses. Taxes would vary by industry based
upon that industry's contribution to circulatory diseases, cancer,
and so on.
- We argue for more and heavier fines on firms that willfully underreport
injuries to the BLS.
- The effect generous WC benefits has had on encouraging injuries is likely
to be small.
- Small firms are treated gingerly by OSHA. They should not be since they
have the highest injury and illness rates of all firms.
The methods introduced in this chapter pertain to the economic laws of diminishing returns and increasing opportunity costs. Put simply, the last, say, 5 percent of heart disease spending could be reallocated to occupational injury and illness spending with the result being a substantial net gain in lives saved and illnesses and injuries prevented.
X. Limitations and Assumptions
- The dollar amount of fraudulent WC claims submitted by workers pales in
comparison to the amount for claims never filed and, more
importantly, the overall small amount of total costs paid by WC
systems. Moreover, fraud committed by insurance companies at
workers' expense is likely to be significant.
- We list 31 critical assumptions: 25 result in a smaller estimate than
otherwise would obtain; two result in a higher estimate; the bias on
the remaining four is unknown.
- Human Capital costs can be viewed as measuring overall health and are
strongly proportional to quality-adjusted life years (QALYs).
- Many episodes of occupational injuries also involved innocent bystanders.
For example, a single pilot death may be associated with scores of
deaths to passengers. We estimated 218 deaths and 68,000
nonfatal injuries to innocent bystanders in 1992. The total costs of
deaths and injuries to bystanders were $2.9132 billion.
Our study attempted to estimate the total costs of occupational
injuries and illnesses to the United States in 1992. This study appears
to be the first to use national data to estimate these costs. We find
that the costs of occupational injuries and illnesses are considerable,
surpassing those of AIDS and nearly as great as those of cancer and
heart disease. Potential victims include any one of the roughly 120
million Americans who work for a living. Since the injuries and
illnesses occur at places of business, some of their costs are spread to
consumers in the form of higher prices throughout the economy, all
workers in the form of lower wages, and taxpayers. But despite the size
of these costs and the fact that so many people pay them, occupational
injuries and illnesses do not receive the attention they deserve
(Rosenstock 1981). By almost any measure, AIDS, arthiritis, Alzheimer's
disease, cancer, and heart disease receive far more attention than
occupational injuries and illnesses. In the course of four years of
medical training, the typical U.S. doctor receives six hours of
instruction in occupational safety and health. The national debate on
medical care rarely addresses occupational safety and health issues.
This is unfortunate. The potential for cost savings from prevention of
occupational injuries and illnesses appears to be significant. ...
1. An early summary of some of our findings was published in the medical literature (Leigh et al. 1997). We received numerous ideas for improvements. As a result, the numbers in the book do not precisely coincide with those in the 1997 study. We prefer our estimates here. These cost estimates are within 10 percent of those from the 1997 summary study. The counts of illnesses and injuries are within 1 percent of those from the 1997 summary paper. The greatest differences between the summary study and this one include these: The summary study included property damage ($9 billion), police and fire protection ($1 billion), and costs to innocent bystanders ($3 billion). None of these are included here.
2. The National Institute for Occupational Safety and Health (NIOSH) receives one of the lowest levels of funding for the nearly 20 National Institutes of Health and related agencies in the Centers for Disease Control. NIOSH research awards sum to roughly one-half of 1 percent of the National Cancer Institute (NCI), less than 1 percent of the National Institute on Aging (NIA), and roughly 7 percent of the National Institute on Dental Research (U.S. Department of Health and Human Services 1992). (There is some overlap between NCI and NIOSH spending. For example, some portion of any NCI spending on bladder cancer would likely have some benefit to a person who developed bladder cancer as a result of job-related exposures. But, in general, the overlap for NCI or NIA or any other institutions is not likely to be large. Among specialists within these fields, few focus on occupational factors. Moreover, if occupation is the focus of a grant proposal to the NIH, reviewers will generally send that grant to NIOSH, regardless of the specific disease being investigated. Finally, 85 percent of our costs arise from injuries, not illnesses.) Moreover, no private charities are available to fund research on occupational injuries and illnesses. By contrast, heart disease has the American Heart Association, cancer has the American Cancer Society, AIDS has the Ryan White Institute, and arthritis has the Arthritis Institute.
None of the federal government's flagship health statistics publications Advanced Data series on either injury-related data visits (Schapport 1994), or on hospitalizations (Hall and Owings 1994), or on emergency room visits (Burt 1995) include any categories for occupational injuries.
As another example of the lack of resources for occupations injuries and illnesses, it is notable that there are more fish and game inspectors in the United States than OSHA inspectors (McGarity and Shapiro 1993, 213).