The Richmond Fed's Region Focus magazine has a great interview with W. Kip Viscusi. Viscusi's an expert on risk and its application in regulatory and policy analysis. Author of the well-regarded and widely used text Economics of Regulation and Antitrust . This is a one of a series of interviews in the Focus; the inteviewer Aaron Steelman always has excellent questions:
RF: How much discipline does the market impose on companies to act in a responsible way with respect to worker safety?
Viscusi: There are three major sources of financial incentives for job safety. By far the most important is the market. Workers on dangerous jobs generally perceive that they are dangerous. This drives up their wages and gives the company an incentive to make the workplace safer. If you look at it empirically, this dwarfs everything else that is going on. The number two player is workers’ compensation. The premiumsfor workers’ compensation are now in the $30 billion a yearrange. Particularly if you are a large enterprise, your workers’compensation bill goes up if you have a bad accident record.We found that in the absence of workers’ compensation, worker fatality rates would go up by one-third. So that’s a very large effect. Then, third, we get to the Occupational Safety & Health Administration (OSHA), which issues health and safety regulation for the Department of Labor. You are looking at zero effect in the early years of the agency, and maybe something like a 1 percent to 2 percent total effect on safety in recent years. It’s very small.
So, overall, the market exerts the most discipline on companies to protect workers. Every death on the job generates significant wage premiums in effect. But if a worker falls to his death because the scaffolding is poorly constructed, OSHA goes in there and imposes negligible fines compared to this. It is just not a big player. Workers’ compensation is a different case. In general, I think it is a very good program. The question is who pays for workers’ compensation? Even though companies pay the bill directly, that bill is passed down to workers who receive lower wages.
What we found is that workers are willing to accept a wage cut that exceeds the costs of the premiums because they value the insurance more than the actuarial costs of the insurance. It’s similar to people who pay more than the expected payout for auto insurance because they value having that protection. Also, workers’ compensation is a highly efficient insurance program. It has very low administrative costs, so it pays out something like 80 cents on the dollar, which is tremendous. In addition, companies get value from the program because it protects them from being sued by their employees in case of an accident. They avoid a lot of litigation as a result, and I think that is a significant benefit.
Here's a bit of the history of U.S. regulatory analysis:
RF: How does one properly derive an estimate of the value of a statistical life for use in cost-benefit analysis? How can those estimates be used to improve public policy?
Viscusi: The main technique used by economists is to look at the money-risk trade-offs reflected in the decisions that people actually make. One context is the labor market, where workers are paid more for dangerous jobs. Another context is the product market, where people pay less money for a relatively unsafe product or more money for a relatively safe product. I have looked at both contexts. But most of my work has focused on the labor market because we have a lot of data on workers’ wages, which we can match to the risks of those jobs.
Controlling for other aspects of the job, we find that workers are in fact paid more if they work in hazardous jobs. This is not a new theory. Adam Smith developed this in 1776. But it was only in the 1970s that economists started estimating the relationship. My current estimate puts the value at $7 million per statistical life. What that means is that if you face an annual risk of death on the job of one chance in 10,000, on average you get paid about $700 extra per year.
During the Carter administration, I worked in the Regulatory Analysis Review Group and was the Deputy Director of the Council on Wage and Price Stability, which was responsible for regulatory oversight at that time. I suggested to OSHA that they use statistical estimates such as these to value the benefits of OSHA regulations. They said, “No. It would be immoral to put a dollar value on human life. Absolutely not.” Then in 1982, OSHA proposed a hazard communication regulation that for the first time would have required the labeling of dangerous chemicals in the workplace and sent it over to the Office of Management and Budget (OMB) for review.
President Reagan had just set up this group within OMB that looked at new regulations and required that the benefits be greater than the costs. OMB looked at this and said this is all very interesting but the costs are greater than the benefits. Because OSHA had argued that putting a dollar value on life was immoral, they instead said that when calculating the benefits of improved safety due to the regulation, they were only going to estimate the cost of death. The cost of death was the present value of lost earnings plus your medical costs after you are killed on the job. Well, you can call it the cost of death or you can call it the value of life, but it’s still the same thing. OSHA appealed the decision to the vice president, who was in charge of all such appeals. He said it was a technical issue and needed to be settled by an expert.
I was asked to settle the dispute between the two agencies over the regulatory impact analysis. It was pretty easy. What I did was adopt every one of OMB’s assumptions with the analysis except for one thing: I used my value of life number instead of the cost of death number. Doing that increased the benefits by a factor of 10. Once you used the economic value of life numbers, the regulation had benefits greater than the costs, and the regulation was issued. So after that, regulatory agencies started using the numbers. Part of the reason was that it was good economics. But a big part was that it often made their benefits look large, and that’s what carried the day.
A related issue I have been working on recently is whether old people’s lives are worth less than young people’s lives. I was at a conference and suggested that the answer was yes, because of shorter life expectancy and lower quality of life. This generated a lot of discussion. Since that time, I have looked more closely at how the value of a statistical life varies with age. It turns out that it doesn’t really drop off the table as you get older. In fact, workers at age 60 have a higher value per statistical life than workers at age 20 because they are richer and can do more things that they enjoy. To take one example, I buy cars with all these additional safety features while my son drives around in a topless Jeep Wrangler. Why would this make sense if his value per statistical life was higher than mine?
Comments