David I. Levine

Industrial Relations

BA 254C


Compensating Differences, and Alternatives

Define compensating differentials as the higher pay that a company must pay under perfect competition to compensate for bad working conditions.  Compensating differences are also called equalizing differences.

Economists often use the model of perfect competition as a benchmark. This model makes the following strong assumptions:

Assumptions about method:

    • Examine equilibrium, after adjustments to shocks have taken place.  (Ignores transitions.)
    • The equilibrium is unique and stable. (Ignores multiple equilibria, where history might matter.)

Assumptions about agents:

    • Agents are infinitely rational. (Ignores bounded rationality.)
    • Agents maximize fixed and exogenous preferences.  (Ignores social learning.)
    • Preferences are independent of others' actions and preferences.  (Ignores fairness or concerns for relative status.)

Assumptions about markets:

    • All wages and prices adjust instantaneously. (Ignores frictions that might lead to recessions.)
    • Agents are price takers. That is, they have no market power. Thus, anyone can move costlessly to an identical seller or buyer if they don't like this one.  (Ignores monopoly, as well as all costs of switching and searching.)
    • There are no externalities or public goods. That is, my production and consumption decisions do not affect others.
    • There are no transaction costs.
    • Information is perfect.  (Ignores problems of incentives/moral hazard, and problems of selection, screening, and adverse selection.)

Under perfect competition, the need to pay compensating differentials implies that employers have an incentive to clean up cheap-to-fix hazards. If a hazard is worth $100 to the marginal employee, but costs $90 to clean up, then profits rise if the employer cleans up the hazard. Otherwise, the firm would need to pay $100 to attract workers.

If a hazard is only worth $89 to the marginal employee, but costs $90 per worker to clean up, then the marginal worker prefers the higher pay. The perfectly competitive outcome would be $89 of above-market pay, and the hazard.

If a particular worker did not care about the hazard, he or she would still get the higher pay because the marginal worker cared, and the employer needed to pay the differential to the marginal worker. For example, consider a deaf worker in a noisy engine room that raises the risks others will lose their hearing. Such a worker does not care as much about the hazards yet receives wages higher than he or she would at an alternative job. The payments that non-marginal workers receive to compensate the marginal worker for hazards are called "rents." That is, the deaf worker receives $89 more at this job than at an identical quiet job place because the marginal worker at a noisy engine room must be compensated for the risk of loss of hearing.

The theory of perfect competition has three cheerful implications. If the assumptions hold:

1. The free market maximizes these rents to workers who don't care much about the hazard.

2. The market gives incentives to clean up hazards that are cheap to clean up, and distasteful to workers.

3. The market sorts workers into the best-fitting jobs. Employees who are tone deaf work as ushers at concerts of the aging Rolling Stones. In general, employees who find the conditions least disagreeable go there.

The implication is that under these assumptions safety regulation is a bad idea. If the government requires an employer to pay the $90 per worker to clean up the hazard, wages will fall by $89. The marginal worker is essentially indifferent (less noise is worth $89, and wages fall by $89.) The deaf worker and other inframarginal workers are worse off -- wages fall by $89, but they valued the hazard at < $89. Moreover, the higher labor costs of $90 per worker is larger than the decline in wages, so profits and employment at this firm will fall. (Employment may not decline nationally if the disemployed workers put downward pressure on wages; then the result will be normal levels of employment but slightly lower wages.)

Testing for Compensating Differences

For empirical tests, economists typically run:

log (wage) = A · individual characteristics such as age, education, and sex +

B · undesirable job characteristics such as danger, dirt, noise

For example, the coefficient on "Work in Alaska" is positive, suggesting a compensating difference for the cold weather. Many studies find compensating differences for dangerous work; that is, employees who work in dangerous jobs receive higher wages (e.g., see the review in Brown, 1980). At the same time, many studies find no such compensating difference.

Discussion questions

1. Why might compensating differences be present in only half the studies?

a. What are statistical problem such that the theory of compensating differences is true, but high wages still go along with good working conditions and high benefits on average?

b. What are exceptions to the assumptions of perfect competition that might lead to compensating differences not being paid?

2. For each exception to the assumptions of perfect competition noted in (1b), describe a carefully tailored policy to meet address the problem you describe.


Here are some partial answers:

Information limits: People are unaware of hazards.

Implication: We expect people to learn about most types of hazards as they spend time on the job.  Dangerous jobs should have higher turnover and/or wages that rise with tenure (Viscusi, 1991). 

Cognitive limits: People incorrectly process information about rare hazards or about hazards that apply to themselves (Tversky and Kahneman, 1988). Specifically, due to cognitive dissonance people in dangerous jobs may be happier denying the danger, even if it costs them income or raises their overall risks (Akerlof and Dickens, 1982).

Implication: If we ask people in dangerous jobs the dangers they perceive, their perceptions will be below objective evidence.

Heterogeneity among preferences: Some people are risk-lovers and like heights (for example). Deaf people may not mind working around dangerously loud noises. Compensating differences only need to be paid if there are not enough workers who do not mind the hazard.

Implication: We can ask people their preferences and find sorting in the workplace.

Heterogeneity in ability to pay.  Assume people who have higher incomes over their lifetime buy safer jobs, just as they buy safer cars and more smoke alarms.  If we do not measure earnings capacity correctly, we will see high incomes going along with high safety.

Testable implication: If we control for unobserved but constant ability by examining a person's job changes over a career, the compensating difference for danger should be more apparent (Brown, 1980).

Differences in bargaining power: At some jobs or workplaces employees have high bargaining power. This bargaining power might be due to high skills, unions, working with valuable and fragile technology or products, etc.  (Theories of bargaining go by names such as insider-outsider theories, conflict theories, ability-to-pay theories, and resource dependency theories.) Efficiency wage theories, where high wages buy higher effort or turnover or perceived fairness also imply that jobs where effort, turnover, or fairness is high will pay high wages and have more safety. Conversely, low bargaining power might be due to absence of these features of due to discrimination.

More generally, the labor market may be segmented with good jobs (high wages + safety) for some, and bad jobs (low wages + lack of safety) for others -- particularly women and minorities. (Theories of segmented labor markets are also called theories of dualism and/or dual or primary&secondary labor markets.)



Akerlof, George A.; Dickens, William T. The Economic Consequences of Cognitive Dissonance. American Economic Review v72, n3 (Jun 1982):307-319.

Abstract: A decision model motivated by the psychological theory of cognitive dissonance is presented as a modification of the usual economic model of rational decision making. Cognitive dissonance theory rests on 3 premises: 1. Persons have preferences, not only over states of the world, but also over their beliefs about the state of the world. 2. People can manipulate their beliefs by selecting sources of information likely to confirm desired beliefs. 3. Beliefs once chosen persist over time. Although the model closely follows standard economic analysis in that people are fully informed and make decisions to maximize their own welfare, it yields different results from the standard analysis and provides better explanations for some phenomena. It is argued that the cognitive dissonance model suggests why noninformational advertising is effective, why Social Security and safety legislation are popular, and why persons at risk of flood and earthquake damage fail to purchase appropriate insurance. These explanations rely on the assumption that people may choose to believe something other than the truth where the perceived benefit of such a belief exceeds its cost.

Brown, Charles, "Equalizing Differences in Labor Markets," Quarterly Journal of Economics, 85, 1980.

Tversky, Amos; Kahneman, Daniel. Rational Choice and the Framing of Decisions. Journal of Business v59, n4, part 2 (Oct 1986): S251-S278.

Abstract: The use of normative analysis to predict and explain actual behavior has been defended in several ways. Despite of these defenses, however, the logic of choice does not provide an adequate foundation for a descriptive theory of decision making. In fact, the deviations of actual behavior from the normative model are too widespread to be ignored, too systematic to be dismissed as random error, and too fundamental to be accommodated by relaxing the normative system. Violations of normative theory are traced to the rules that govern the framing of decision and to the psychophysical principles of evaluation contained in prospect theory. It is concluded that no theory of choice can be both normatively adequate and descriptively accurate. As an alternative, a descriptive model of

Viscusi, W. Kip; Moore, Michael J. Worker Learning and Compensating Differentials.  Industrial & Labor Relations Review, 45, 1 (Oct 1991):80-96.

In the standard compensating wage differential model, workers value their wage and workers compensation components based on full job risk information. Market forces generate positive wage differentials as ex ante compensation for exposure to relatively high risk. Similarly, market forces generate wage offsets for the increases in ex post risk compensation embodied in workers compensation benefits. It is hypothesized that, in industries with relatively high levels of job-related injury risk, workers with longer job tenure will more clearly appreciate the degree of job risk than will newly hired workers. These longer working employees will thus be more willing to accept lower wages in return for higher workers compensation benefits. This hypothesis is confirmed by an analysis of quit behavior using 1981-1983 data from the Michigan Panel Study of Income Dynamics and 1981-1985 data from the National Institute of Occupational Safety and Health.