Bryan Caplan  

Why So High? Economics and the Value of Life

Drowning Redheads is Wrong Eve... Imagine there's no economic in...
Economists are widely-seen as heartless.  Their use of the phrase "value of life" is often seen as damning confirmation of this heartlessness.  Nice people say, "You can't put a value on a human life" and change the subject!

What's striking, though, is that when you successfully cajole non-economists to put a dollar value on a human life, their numbers are vastly below the economic consensus.  Economists' standard estimate is around $7,000,000.  Non-economists' usually say under $1,000,000.  Their reasoning varies, but I've heard garbled versions all of the following:

1. People can't pay more than they have, and most people have well under a million dollars.

2. People earn around $40,000 per year after taxes, and most have under 25 years left to work.  So if you multiply annual earnings by remaining working life, most lives are worth under a million.

3. People need most of their income to live.  So if you measure the value of life by people's willingness to pay to stay alive, it's probably no more than $10,000 per year.

Why are economists' numbers so much higher?  Part of the reason is that economists detect blatant flaws in all three of the popular approaches:

Flaws with #1: When people bid for a house, their willingness to pay is emphatically not limited by their current possessions.  Credit markets allow them to borrow against future earnings.  When we estimate people's willingness to pay for their own lives, we should perform an analogous exercise.  Maximum willingness to pay is limited by the present discounted value of everything people will EVER own - their time emphatically included. 

Deeper flaw: Willingness to accept is just as valid a measure as willingness to pay.  For small purchases, these numbers are fairly similar.  But for large purchases, they differ dramatically.  (Of course if you take this caveat too seriously, economics comes close to endorsing the view that every life is infinitely valuable, since most people will not agree to die for any sum).

Flaws with #2: This approach ignores opportunity cost.  You shouldn't just include the value of the time people choose to sell.  You should also include the value of the time people choose not to sell.  In other words, if you're going to apply this method, you should measure potential earnings, not actual earnings.

Flaws with #3: It's simply not true that people in the First World "need most of their income to live."  People can physically survive while couch surfing, eating beans and rice, and working three jobs.  And if this austere lifestyle is their only way to stay alive, most submit to it.  This is a key lesson of every horrific war of the twentieth century.

If you make economists' recommended refinements, all three popular value of life approaches yield much higher answers.  But they're probably still short of $7,000,000.  How do economists get to their preferred answer?  By using a totally different metric: Willingness to pay (or accept!) for a small change in the probability of staying alive.  Something like:

4. How much would you pay to reduce your probability of death by 1 percentage-point?  Alternately: How much someone have to pay you to increase your probability of death by 1 percentage-point?

$70,000 is a very reasonable answer to questions like this.  Once you buy that idea, simply multiply by 100 to get a $7,000,000 value of life.

But why is approach #4 superior to refined versions of approaches #1, #2, and #3?  Truth be told, superiority is not clear-cut.  The strongest defense of approach #4, though, is that human beings rarely choose between certain death and a giant pile of money.  Instead, they face an endless series of choices between a small probability of death and a modest pile of money.  So approach #4 is the best way to quantify the value of life decisions we encounter in the real world.

This remains true even if we know with certainty that one person will die as a result of a choice.  Only individuals value stuff.  So if each person faces a 1% chance of death, summing the cost each person assigns to his own 1% risk makes sense.

Or does it?  Parting thought: Almost everyone - economists and non-economists alike - strangely neglects a big part of the value of life: The value people place on the lives of the people they care about.  Human beings worry about each other, and most who die are missed. 

Yes, Social Desirability Bias leads us to exaggerate how much we care about strangers and casual acquaintances.  But we plainly genuinely care about family and friends.  So economists shouldn't just ask how much you'd pay to reduce your risk of death by 1 percentage-point.  They should instead ask how much everyone including yourself would pay to reduce your risk of death by 1 percentage-point.  Unless you're a total jerk, $7,000,000 a life is probably a serious understatement.

COMMENTS (16 to date)
Philo writes:


We've never met, but I'd certainly chip in some positive amount to reduce your chance of death by 1% (over the next year; I'm afraid your chance of death eventually is 100%!). Most of us have family members and friends; you also have lots of fans. (Just think of what Justin Bieber's life is worth!)

Finch writes:

Morbidly, would you take a dollar for increasing the chance of death this year of someone you have never met on another continent by 1%, particularly if you could do it at a large scale? Is it the case that at some distance from oneself one's concern drops to zero?

If it doesn't drop to zero, that's going to make self-interested decisions increasingly hard to make as population rises, so I'm inclined to believe it really does drop to zero. At least outside of one's immediate focus.

Trevor Adcock writes:

But if people are willing to spend all of their potential lifetime earnings to keep themselves alive then there is nothing left over for them to pay to keep their friends and family alive.

There are only so many resources in the world and the value of all human lives cannot total more than the present value of all potential output in the world.

Alex Godofsky writes:


There are only so many resources in the world and the value of all human lives cannot total more than the present value of all potential output in the world.

Why not?

If a bunch of oil under Saudi Arabia disappeared, would everyone's lives become less valuable? That seems like a curious result.

Dave Tufte writes:

;) And absolutely never, ever, bring up the subject that at least part of the interest in "gay marriage" is opening up some of that value to people you like rather than people you're related to.

People will point out that you must be heartless to say that too ... and that you can't put a value on relationships ... and then change the subject.

FWIW: A friend who is an independent health insurance agent was very, very, busy after gay marriages were allowed in our state. Funny that ... it almost makes you think that it's all about subsidization of healthcare access through the tax code, and not love.

Dave Tufte writes:

In re-reading my above comment, I wonder if the satire comes through.

In my defense, as a libertarian, I have no problem at all with gay marriage in any form. But as an economist, I feel it's my duty to point out that there's an awful lot of fungible, liquid wealth available to someone who's self-serving and takes advantage of changes in marriage laws.

Wanjing Chen writes:

Everyone 's price of "value of life" is different. we have learned in Econ class that to charge the maximum he or she is willingness to pay is called perfect price discrimination, which in this case depends on present discounted value.

Jason Scheppers writes:

I propose that the value of life should be thought more like a supply and demand curve. A 1% decrease in probability death should not be assumed to be 10 times a 10% decrease probability of death. The value of life varies with time and with the purchasing power of the person in question. One should assign the zero value of life to all those that commit suicide, right?

I believe a lifetimes earnings has to come into play to some extent, but I agree you cannot limit the price to lifetime earnings, because of the reasons you cite. I am concerned that many of the method used miss the trade offs that have to be made. Not accounting for the trade offs allow the appraised value of life to be too high.

Finch writes:

In aggregate we cannot value people more than our (potential) ability to pay because "[t]he value people place on the lives of the people they care about" is necessarily value we don't place on our own lives. It's not some new source of ability to pay.

MingoV writes:

Using a single value for the worth of a life makes little sense to me. An 82-year-old with multiple chronic diseases, who is in chronic pain, may put a zero value on life.

The medical field does not use a single value for a life. It uses quality-adjusted life years. A value is assigned for a year of healthy life, and then the effects of age and illnesses are factored in. If the one year value is set at $100,000 (it can be set at any value), then a healthy 50-year-old with a life expectancy of 35 years has a (remaining) life worth of $3,500,000. A very unhealthy 50-year-old may have an adjustment factor of 0.5 and a life expectancy of 12 years. His life worth is $600,000. A healthy newborn with a 90-year life expectancy would have a life worth of $9,000,000.

terrymac writes:

If you care about someone, one method of showing that love is to improve their well-being. If gay marriage laws allow one to do so more easily - such as ending the tax bias, per the recent ruling against DOMA, that's no more objectionable than if one were to love and reward a person of the opposite sex in similar fashion.

Self-interest cannot be a dirty word for gays and lesbians, but a sacred matter for heterosexuals only.

Fazal Majid writes:

The survey method is still depending on stated vs. revealed preferences.

A better way would be to see how much people are willing to pay for safety features on cars like airbags, blind spot detection radars or enhanced child car seats. Better yet - people who choose scooters over cars in congested urban areas because they can get through traffic faster, at the risk of higher fatalities in case of an accident. Another good revealed preference is the wage premium for dangerous occupations.

Granted, humans suffer from cognitive biases that lead them to overestimate low probability events and underestimate more frequent ones.

James writes:

Maybe you will agree to get a million dollars exchange 1 year of your life. But how about get 70 million dollars and none of live. In my opinion, through the upper of marginal of the rate of the utility exchange to live, the utility must be higher than a million dollars.
To sum up, the utility of live can't be calculated by 1000000*70=70000000.

Bryan writes:

Value to whom?

My life is what makes values possible. Without it, there's no question of value to me.

AS writes:

You can't take the 1%=$70,000 ratio and extrapolate to 100%. The curve may be nonlinear or people make irrational decisions at small percentages. It's already known that people have a "kink" in their utility curves when asked to place value on very small gambles.

Better to use the positive externalities argument. Each human life benefits others around them through trade.

Mark V Anderson writes:

Bryan, Where do you get %70,000 to reduce probability of death? It sounds very high to me. People seem to routinely takes risks that increase their probability of death that much. I don't think the same people would risk $70,000 as easily.

Of course AS is right too, that there is no way the curve is linear.

Comments for this entry have been closed
Return to top