The economics and politics of valuing life

February 17, 2011
Binya Appelbaum's NYT article on the various different values of a human life which are used by government agencies to justify regulations.

" data-share-img="" data-share="twitter,facebook,linkedin,reddit,google,mail" data-share-count="false">

I love Binya Appelbaum’s NYT article on the various different values of a human life which are used by government agencies to justify regulations.

The first thing to admire about the piece is that it doesn’t dwell on ethics or philosophy, as most such stories do — there are no rhetorical flights of fancy about the government trying to put a dollar value on love, or that kind of thing. Instead, Appelbaum goes on a tour of government agencies, looking at the numbers they’re using now, how those numbers differ from other agencies, and how they have changed over time:

The Food and Drug Administration declared that life was worth $7.9 million last year, up from $5 million in 2008, in proposing warning labels on cigarette packages featuring images of cancer victims…

The Bush administration rejected a plan in 2005 to make car companies double the roof strength of new vehicles, which it estimated might prevent 135 deaths in rollover accidents each year…

Last year, the Obama administration imposed the stricter and more expensive roof-strength standard, and it published a new set of calculations showing that the benefits outstripped the costs.

Most of the difference came from the increased value of human life. By raising that number to $6.1 million from a figure of $3.5 million in the original study, the Obama administration rendered those 135 lives — and hundreds of averted injuries — more valuable than the roofs…

Agencies are allowed to set their own numbers. The E.P.A. and the Transportation Department use numbers that are $3 million apart. The process generally involves experts, but the decisions ultimately are made by political appointees.

The Office of Management and Budget told agencies in 2004 that they should pick a number between $1 million and $10 million. That guidance remains in effect, although the office has more recently warned agencies that it would be difficult to justify the use of numbers under $5 million, two administration officials said.

This kind of behavior leaves the agencies open to charges of inconsistency and capriciousness: if at first you don’t succeed in making your cost-benefit calculation work, then just try again with an arbitrarily higher number for the benefits involved.

But I think that this is a case where the perfect is the enemy of the good. As Manchester University professor Robert Hahn notes in the article, “the reality is that politics frequently trumps economics”. That’s a fact of life. And in a world where political considerations are ultimately going to power many if not most decisions, using dollar values for lives saved is a good way of keeping such arguments grounded in reality.

Sure, businesses don’t like it when the FDA ups its value for a life saved by acetaminophen warning labels to $7 million from $5 million, and it’s entirely possible that the FDA changed the valuation only so that it could provide an official justification for a decision it had already made. The fact is, however, that these calculations are always messy at the best of times. It’s easy to point to the value-per-life part of the calculation, because that’s a hard number. But how on earth is the FDA meant to calculate the number of lives saved by adding a second warning label to acetaminophen bottles? The error bars there are going to be much bigger than the differences in value-per-life numbers.

In that context, a little bit of fuzziness in the $5 million to $10 million range seems entirely reasonable to me. It’s regulators’ job to make judgments, not to simply sit at a desk with a calculator and determine which of two numbers is larger. And at the same time it’s reasonable to ask regulators to justify their judgments using math. So sometimes they’ll use a slightly higher number, and sometimes it’ll be lower. Giving regulators a bit of wiggle room gives them the ability to do their jobs, while restricting that wiggle room allows a simple smell test to be applied.

None of this is exactly pretty, and it’s easy to see why Appelbaum couldn’t get straight answers out of the technocrats he talked to. But if anything the amount of wiggle room is smaller than I would think reasonable:

In December, the E.P.A. said it might set the value of preventing cancer deaths 50 percent higher than other deaths, because cancer kills slowly. A report last year financed by the Department of Homeland Security suggested that the value of preventing deaths from terrorism might be 100 percent higher than other deaths.

Both those numbers could and arguably should be significantly higher, I think. Dying of cancer is a particularly gruesome — and expensive — way to go. And the cost of the terrorist attacks of September 11 is well up in the trillions at this point — getting on for a billion dollars per initial life lost.

So color me impressed that the US government has found a way of getting things done and remaining empirical in an atmosphere which by its nature is always going to be highly political. It comes as no surprise that the Obama administration is using values higher than the Bush administration did — that’s part of what Obama meant when he promised to toughen up government regulation of corporations. I’m just happy that there’s a culture in Washington of basing these decisions on some kind of numerical argument.

(On which matter I have one quibble with Appelbaum’s piece. He says that if companies must pay lumberjacks an additional $1,000 a year to perform work that generally kills one in 1,000 workers, that would impute a $1 million value on a human life. I don’t think that’s true: you should take the present value of $1,000 per year before you multiply by 1,000. So the imputed value of human life here would be much higher than $1 million, depending on how long the average lumberjack works at his job.)


Comments are closed.