Government is raising the value of a life

The Environmental Protection Agency set the value of a life at $9.1 million last year in proposing tighter restrictions on air pollution. The agency used numbers as low as $6.8 million during the George W. Bush administration.

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 article is here.  If the goal is to give current people what they want, arguably this makes sense and perhaps it does not go far enough.  Death is…BAD.  If the goal is to maximize real gdp per capita, or most other macroeconomic indicators, it makes sense to value human life at replacement cost (and here) and this policy change does not make sense.  I'm not arguing for either standard and indeed I think they both lead to absurdities.  Instead the point is this: theoretical ordinal welfare economics and applied welfare economics, as represented by wealth measures, do not coincide as much as many economists like to think.  This gap becomes increasingly important as health care and safety provision increase, relative to the size of the economy as a whole.

What the Chinese have done is to neglect health care investments (until very recently) and basically maximize gdp growth.  They wanted to have fewer people anyway, so why spend money to keep ailing people around?  We find this horrible when presented in such explicit terms, and yet we admire their achievement of the end of growth maximization.


How much does the DoD value the lives of swarthy civilians in other countries at? $0.09 million or so?

There's a very simple way to calculate what dollar value we put on our own lives. It will just require a few volunteers, some wagers and some Russian Roulette. Assume a strong EMH.

..."yet we admire their achievement of the end of growth maximization"
Not the progressives, who in general frown on development at the cost of human and environmental degradation. Right wing, hates them for narrow patriotic reasons.
Other than businesses and people who buy into (neo)liberal project, I don't see anyone admiring them.

What size is the average Life Insurance policy in the US? Does that shed much light on the question?

If the FDA really believed their own $7.9 million figure, they would be a LOT more aggressive at approving new treatments.

I suspect it's just a publicity number and they don't really do any such calculation to support their own decision-making.

So which number will the Death Panels use to compare to the present value of future medical treatment? And how will the foregone SS benefits be accounted for in the equation?

"yet we admire their achievement of the end of growth maximization."

Yeah, the same way we admire Barry Bonds and Mark McGwire.

Um. does this mean that people who can't afford life insurance value their own life at zero?

And since they would (can) pay nothing at all to stop any number of other people's deaths, they value everybody else's life at zero?

Who knew the poor were so evil!

EPA $9.1m/life
FDA $7.9m/life
I sense an arbitrage opportunity!

Shouldn't the value depend on age? I would expect it to be non-monotone, with an early peak at say five years? (It seems wrong that newborns should have have the highest value, considering replacement and emotional attachment)

If you consider the behavioural economics of this, that people are more affected by risks that they can't control, it's optimal too have a value that exceeds revealed preferences.

A high value on human life is essentially a regressive policy that redistributes wealth to the rich. To the extent that it results in greater expenditure on "luxury" goods like safer roads (and spending millions on futile medical treatments), it means greater expenditure on the things that the well-off should relatively value more. I think this is part of the reason why third world countries are not livable places for rich first world people - a lot of decisions about spending on things like safety are made on the premise that lives are easily replaceable.

I'm going to have to correct my proposal as I realize that using volunteers would induce selection bias. Here's the way we need to do this if we want to be scientific (It's going to cost some money, but that's OK):

Start by assuming a liberal par value of 10MM per life. Give a professional negotiator 16.7MM. Select 10 people at random. The negotiator calls these randomly selected Americans and negotiates a game of Russian Roulette with a six shooter at whatever price he can (odds of dying 1 in 6). The key is that he gets to keep any money he saves if he is able to talk all 10 people into the game, thus he is motivated to get the lowest price possible. Then take the average price he is able to get people to play Russian Roulette for and use that as the starting par value for another round of 10 people with a different negotiator. Iterate until an equilibrium price is discovered.

Perhaps this sounds a bit cruel and unusual but if you don't like it don't ever say you're as curious to know the value of a human life as much as I am. I'm available to work on death panels, BTW.


What would the shape of the value curve look with respect to age? Just curious and guessing....


The average over all the groups sampled so far? That should converge, right?

Is this change real or nominal? Either way, Scott Sumner should be pleased.

@Rahul: then it just isn't clear why this is an iterative scheme, per se. Normal iterative techniques rely on the actual output of each step to converge to the answer; that's why the used it as their guess in the next step. I'm wary of the effects of changing the experimental parameters for each sample and then averaging them for a single data point.


If money available to conduct the experiments wasn't the limiting factor you could adopt a non-iterative solution. Send many negotiators out each with excess of money. Problem is you don't know the exact value of a life a priori. To maximize negotiator diligence the amount he is given should be only slightly more than the life-value.

These are not values of life. These are values of a statistical life (VSL), which is very different.

VSL is the value of a small reduction in risk, linearly extrapolated to a 100 percent risk.

Basically, the EPA has found, through many studies of individuals' actual behavior, that people will pay around $91 to reduce a risk of death of 1 in 100,000, for example.

Take $91 and divided by 1/100,000 and you get $9.1 million.

No one -- not the EPA, not the FDA, no one -- says that a human life is worth $9.1 million. Rather, they are saying that a reduction in risk of 1 in 100,000 is worth $91. The conversion to $9.1 million is a convention, and a highly confusing and unfortunate one at that.

It would be far better if the EPA and others would simply report that they are updating the value of the reduction of a risk of 1 in 100,000 to $91. That's how the EPA uses it, anyway. For example, the EPA proposes a rule to tighten a pollution standard. The effect of this tightening is to reduce everyone's risk of death by a small amount. How much is this worth? Use the $91 per person per 1 in 100,000 and calculate it.


As for the curve with respect to age, I was thinking how that might take an unintuitive shape. For instance, one might normally assume that a younger life would be worth more, yet such an experiment may show that youth doesn't value their lives any more than older people. Particularly since younger people tend to have less money they may be willing to settle for less than someone at middle age. (Though we could correct for wealth and income.) Someone who is very old, on the other hand, who is more likely to suffer from illnesses and whose quality of life may not be so great to start with, may demand significantly more money, not because they are more afraid to die but because only significantly more money would seem to make much difference to their quality of life. So the experiment may reveal that people tend to value their lives MORE as they age. (Which may explain our seeming upside-down value system regarding health-care spending.)

Dirk, Rahul: It turns out that the shape of the VSL curve with respect to age is an upside-down U, but a relatively flat one. VSL peaks at middle age and is lower for the young and old, but not by much: The difference between the peak at middle age and the values for the young and old is about 10 percent.

What method did they use here, the old price of a bicycle helmet technique?

Anyway, can anyone think of a greater misnomer in the Economic arsenal of terms than VSL? Seems to me that a VSL study or article will come out every 6-8 months or so, only to be roundly mocked in popular media. Can’t Economists and insurance folks come together with a term like marginal risk aversion premium, or something esoteric sounding enough that the plebeians won’t erupt into laughter?

Comments for this post are closed