Occupational Safety and Health: a Problem of Double or Single Moral Hazard
Following Cooper and Ross (1985), this article contributes to the literature by
developing a principal-agent approach to the effect of imperfect information on safety
in the workplace. Firms and workers can influence the probability of workplace
accidents by taking precautions. To the extent that a worker's precaution cannot be
observed by a firm, and vice versa, contract designing is a problem characterized by
moral hazard. Double and single moral hazard lead to nonoptimal levels of precaution
from firms and workers, but not necessarily to an underprovision of precaution.
Furthermore, it is shown that regulatory policies aimed at promoting safety in the
workplace (by inducing increases in the levels of precaution) may not be
Some authors, such as Thaler and Rosen (1976), allege that the presence of wage differentials for risky jobs implies that public regulation of occupational safety and health (OSH) is unnecessary. They argue that risk premiums in dangerous occupations are enough to compensate a worker in the case of an accident and to induce the socially optimal effort needed to reduce hazards.(1) This argument against regulation of safety in the workplace holds in a world of perfect markets and complete information. However, it breaks down with incomplete information or imperfect markets, and the literature on accident prevention(2) shows that nonoptimal situations can occur. For instance, there are externalities; Manga et al. (1981) argue that risk in the workplace leads to indirect costs that should be taken into account in the firm's decision concerning safety expenditures. An obvious example is the indirect health cost that a child incurs when a pregnant woman is at risk.
Problems of imperfect information in the "market of workplace accidents" have also drawn special attention. In particular, Oi (1974), Diamond (1977) and Rea (1981) postulate that workers are not well informed about safety levels prevailing in different firms. Since market outcomes are not efficient in this case, the effect of various policies such as mandatory insurance (workers' compensation) and safety regulation should be analyzed. Diamond and Oi argue that mandatory insurance is justified because it raises the expected utility of risk averse workers, while the enforcement of safety standards is required in order to improve the level of safety in the workplace, which is suboptimal when workers underestimate risk. In a different model, Rea (1981) shows that, when workers underestimate risk, mandatory insurance and safety regulation may have the perverse effect of reducing welfare and the level of safety. In particular, wages may be reduced with mandatory insurance and, as a result, the level of safety may fall because workers will attempt to substitute wages for safer jobs.(3) Carmichael (1986) disagrees with the "uncomfortably ad hoc" assumption that workers underestimate risk. He integrates imperfect information by exploring the role of a firm's reputation in repeated games. His model suggests that it takes time for workers to learn about changes in safety levels in a firm which leads, generally, to an underprovision of safety. In contrast with Rea, Carmichael is able to make unambiguous statements about the welfare-improving nature of government intervention related to OSH. Among other results, he shows that an increase in the level of compensation benefits received in accident cases leads to an unambiguous improvement in welfare.
This article contributes to the literature by working out a new approach to the effect that imperfect information has on safety in the workplace. In particular, the problems of double and single moral hazard are examined in the "market of workplace accidents". …