Peter Dorman begins his post by making the laissez faire argument:
The standard theory, which can be found in any labor economics textbook and even some principles texts, is that dangerous work is essentially a non-problem. Workers will only take a dangerous job if they get extra pay to compensate. This gives employers an incentive to make jobs safer up to the point when the cost of more safety exceeds the added wages they would have to pay. The result: the level of safety is efficient (marginal benefit of safety improvements equals marginal cost), workers with a taste for risk are efficiently matched with employers whose technologies make risk harder to reduce, and workers in dangerous jobs are no worse off than those in safe ones, since they have bigger bank accounts to keep them happy. There are two big regulatory implications: interference, like enforcing safety rules, is against the interest of both workers and employers, and the wage-safety tradeoff can be used to calculate a “value of statistical life”, which comes in handy whenever cost-benefit analyses need to be performed. I kid you not. The “consensus” view in the profession is that a wealth of empirical data validates this theory. The Environmental Protection Agency holds earnest discussions on whether the value of a life should be raised or lowered a bit, or perhaps set at a higher level for some groups (like the rich) compared to others.
Peter sort of trashed this in his book from eleven years ago:
This book provides a critical survey of conventional economic approaches to occupational safety and the analysis of environmental risk in general. The author concludes that unsafe work is not voluntary, that markets do not compensate workers for risk, and that attempts to put a monetary value on life and health are futile. He attributes the shortcomings of economic orthodoxy to its underlying approach to human decision-making and social interaction, and demonstrates that useful alternative approaches are available. The analysis is used to identify policies that combine effective regulation with democratic values.
The rest of his blog post and his Markets and Mortality are now required reading.