The Backdoor Discrimination of Cost-Benefit Analysis

In recent weeks, an unusual convergence of events has served to elevate somewhat the public profile of cost-benefit analysis (CBA).  Before then, CBA was an obscure and highly complex tool of policy analysis—the kind of thing that hardcore policy wonks would wonk about when the subjects of their usual policy wonkery weren’t wonkish enough.  Foreseeable future events suggest that the public profile of CBA will continue to rise.

The process began in early January when word emerged from the Obama transition team that the then-President-Elect planned to name Harvard law professor Cass Sunstein to head the Office of Information Regulatory Affairs (OIRA).  A little known but powerful bureau in the White House Office of Management and Budget (OMB), OIRA supervises the entire federal regulatory apparatus, imposing its will by individually reviewing all major federal regulations through the lens of CBA.  OIRA’s use of CBA, has been the subject of quite a bit of scholarly criticism, notably by the Members Scholars of CPR.  Since the otherwise progressive Professor Sunstein has been a prominent advocate of cost-benefit, the subject got a bit of an airing when news of his likely nomination leaked.

The public profile of CBA was given an additional boost after the announcement by President Obama that the Administration intends to revamp the whole enterprise of federal regulatory review.  In a Memorandum to Heads of Executive Departments and Agencies issued on January 30, the President launched a review of Executive Order 12866, the Order that establishes the institutional framework for centralized regulatory review by OIRA, with an eye toward revising it significantly or even replacing it altogether.  Among the issues to be reviewed is the future “role of cost-benefit analysis.”

One CBA issue that deserves scrutiny in that process is the use of the value of statistical life years (VSLYs) as a method of discounting the benefits of protective regulation.  The reason:  the VSLY method lends itself to discriminatory treatment of different population subgroups.

VSLYs first came to public attention during the Bush Administration, and the practice was dubbed the “senior death discount,” because it effectively waters down regulatory protections that primarily benefit older people.  The discount works by applying a systematic downward adjustment for measuring the benefits of regulations that protect the health and safety of seniors.

Presently, regulatory agencies use something called a “value of statistical life” (VSL) instead of the senior death discount when attempting to measure the value of a regulation that will result in saving lives.  Significantly, the VSL that agencies use is uniform across all people who might be saved by the regulation—regardless of their age.  So, for example, EPA uses a VSL of $6.1 million.  Thus, a regulation that saves one life, statistically speaking, produces a benefit of $6.1 million.  EPA’s VSL was calculated by using a willingness-to-pay study to measure people’s valuation of avoiding certain risks.  Thus, for example, EPA’s VSL might be based on a study that found that individuals on average were willing to pay $61 to avoid a 1/100,000 risk of dying.

Most supporters of CBA have rejected the use of VSLs, however, on grounds that it’s imprecise.  They argue that since we’re all going to die eventually, the purpose of health and safety regulation is to extend lives, not save them.  Therefore, the mark of a successful regulation should be not how many lives it saves, but rather how many additional life years it preserves.  That might seem like dancing on the head of a pin, but here’s the point:  By gauging the value of statistical life years (VSLYs), instead of the value of statistical lives (VSLs), the cost-benefit equation changes significantly.  VSLYs effectively discount the value of lives saved, making it harder to justify regulating by cost-benefit standards.

That right there is worrisome enough.  But another cause for concern is the insidious result that flows from it.  The approach favors regulations that protect young people over those that protect older people.  And proponents defend that outcome on the merits.  In their view, it is merely a reflection of the fact that the elderly presumptively have fewer life years left, and therefore receive less benefit from regulations, and in fact probably should receive less benefit..

The support by CBA proponents of the senior death discount has drawn a lot of criticism.  The fact is that it treats people differently in significant ways, because of an attribute—namely, age—they cannot control.  Moreover, the senior death discount, like the whole enterprise of CBA, denies the intrinsic value of individual lives.  Undoubtedly, most Americans regard the “value” of their lives as more than the monetized “price” of the sum of their expected remaining life years.

A second CBA issue that deserves a bit more scrutiny is the fact that many proponents of CBA have also endorsed the practice of adjusting the value of statistical lives (VSLs) in ways that would effectively discriminate against the poor and racial minorities.  This adjustment has been referred to by many CBA proponents as “disaggregation.”

Recall from above that agencies like EPA use a uniform VSL for everyone.  As with the senior death discount, disaggregation begins with the observation that a uniform VSL provides an inaccurate measure of the value of human life.  In particular, CBA supporters note that the willingness-to-pay studies on saving human lives reveal an interesting quirk:  different segments of the population appear to value their lives differently.  Therefore, logic dictates that VSLs should vary across these different segments of the population, so that the benefits of regulations that disproportionately affect one of these segments of the population can be measured more accurately.

At this point, many readers may be wondering why CBA proponents are attempting to refine the concept of VSLs through disaggregation, even after they proposed jettisoning the use of VSLs entirely by replacing it with the senior death discount.  It is not clear whether these two concepts are mutually incompatible.  It may be true that many proponents of CBA believe that disaggregation and the senior death discount should be applied in a sort of discriminatory two-step.  In the first step, the VSL would be properly adjusted for the subpopulation that would be disproportionately benefited by the proposed regulation.  In the second step, the disaggregated VSL would be converted into VSLYs to reflect the age group within the affected subpopulation that would be disproportionately benefited by the regulation.

In any event, those proponents of CBA that endorse “disaggregation” seem willing to follow its logic no matter where it takes them.   For example, Professor Sunstein notes in a law review article entitled Valuing Life: A Plea for Disaggregation that willingness-to-pay studies indicate that rich people are willing to pay more to avoid certain risks than poor people.  Therefore, when conducting cost-benefit analysis for regulations that achieve regulatory benefits disproportionately for rich people (e.g., a regulation that limits the presence of certain toxins in caviar), then regulatory agencies should use a higher VSL when measuring the benefits of that regulation.  This, of course, means that regulations for the benefit of rich people will necessarily be more stringent than regulations for the benefit of poor people.

Of course, CBA supporters like Professor Sunstein are not suggesting that poor people are “worth less” than rich people.  Rather, they contend that the variance in VSLs is based on the fact that certain people value themselves less, and that society should respect these self-estimations.  But isn’t it possible, even likely, that poor people are willing to pay less to avoid certain risks because they have less to pay?  If you’re rich, and don’t place much value on a few hundred thousand dollars, you’re more likely to part with it to protect yourself and your family from a known hazard.  But if that’s more money than you can imagine having, you might have to take your chances.

And what if the willingness-to-pay studies determine that racial minorities are less willing to pay to avoid certain risks than are whites?  Doesn’t that decrease the benefit—for cost-benefit purposes—of protecting those groups?

As these examples illustrate, recent efforts to refine CBA into a perfect regulatory review calculating machine can have disturbing consequences.  If adopted, the senior death discount and disaggregation would have the effect of reformulating CBA so that it would systematically and surreptitiously discriminate against some of the most vulnerable subpopulations in society, including the elderly, the poor, and racial minorities.  Unfortunately, these are the very subpopulations that are most dependent on the protections offered by environmental, health, and safety regulations.

With Executive Order 12866 under review, the future role of CBA in regulatory review is at a critical juncture.  With luck, this review will lead to a broader understanding of the many disturbing aspects of CBA, including its amenability to being used to discriminate against the most vulnerable segments of our population.  Such a broader understanding will be necessary if the revision or replacement of Executive Order 12866 will include a decision to abandon or at least significantly modify the role of cost-benefit analysis in the regulatory process.