Problems with Aggregating Data when Calculating the Value of Statistical Life
Government agencies use a calculation known as the “value of statistical life” (or VSL) in order to assess the costs and benefits associated with reducing the number of deaths in different social sectors. Traditionally, this calculation has relied upon assessing how much individuals are willing to pay to decrease their risk of death. By dividing this statistic by the actual decrease in the risk of death, government agencies are then able to generate a rough estimate of an individual’s VSL. However, the efficacy of this calculation is greatly diluted by the fact that these individual VSLs are averaged into one aggregate VSL, a practice that ultimately treats different risk factors equally. In essence, this aggregation process oversimplifies the heterogeneity of affected individual parties, thus making it difficult for government agencies to effectively allocate resources toward death prevention. In order to more accurately assess the costs and benefits associated with reducing the average number of deaths, government agencies should disaggregate risk factor data. This would have important practical implications for many government-funded programs, from cancer research to health care benefits for impoverished Americans.
Currently, each government agency has a VSL that it uses in all its cost-benefit analyses, essentially giving each type of risk equal weight. For example, the Food and Drug Administration will use a VSL of $7.9 million, whether the cost-benefit analysis is for cancer drugs or for tomato standards (to prevent salmonella infections). By using the same VSL, the FDA is implicitly assuming that people will spend an equal amount of money to prevent death by cancer as they will to prevent death by salmonella. But in a circumstance where death is slow and painful, like for cancer patients, the VSL ought to be higher because people would be willing to spend more to avoid slow, protracted suffering. Thus, the equal treatment of different risks is flawed. By disaggregating data based on risks, the FDA can more accurately assess the benefits and costs to proposed policies, and devote more resources to risks that call for a higher VSL.
Since a government agency uses a single VSL, the cost-benefit analyses do not account for differences between demographics. But the wealthy have a higher VSL than the poor do, since the wealthy would be willing to pay more to avoid death than the poor would be willing to pay. That is not to say that the government ought to spend more money on programs that prevent the wealthy from dying since they have a higher VSL. Rather, disaggregating VSL by demographics will enable the government agencies to make policies more equitable, as they account for distortions in individuals’ stated preferences. Take Medicaid, for example. Low-income Medicaid recipients pay for a fraction or none of the cost of Medicaid. Thus, the market data that is used to calculate VSL is flawed, since if not for the government subsidy, they would be willing (if they are able) to pay more for healthcare. Then VSL should be set higher than the one calculated from market data, indicating that the benefit of Medicaid to low-income people is actually greater than what benefit-cost analyses currently suggest. By disaggregating VSL by demographic data, flaws in data collection are revealed, and government agencies can better tailor their programs in order to address the needs of different demographics.