How do Olympic gold medals influence longevity?

I was intrigued by the new paper by Adam Leive, called “Dying to Win? Olympic Gold Medals and Longevity.”  The main results are these:

This paper investigates how status affects health by comparing mortality between Gold medalists in Olympic Track and Field and other finalists. Due to the nature of Olympic competition, analyzing performance on a single day provides a way to cut through potential endogeneity between status and health. I first document that an athlete’s longevity is affected by whether he wins or loses and then detail mechanisms driving the results. Winning on a team confers a survival advantage, with evidence that higher mortality among losers may be due to poor performance relative to one’s teammates. However, winning an individual event is associated with an earlier death. By analyzing the best performances of each athlete before the Olympics, I demonstrate that an athlete’s performance relative to his expectations partly explains the earlier death of winners in individual events: on average, Olympic Gold medalists expected to win, but losers exceeded their expectations. Conversely, athletes considered “favorites” but who fail to win die earlier than other athletes who also lost. My results are robust to estimating a range of parametric and semi-parametric survival models that make different assumptions about unobserved heterogeneity. My central estimates imply lifespan differentials of a year or more between winners and losers. The findings point to the importance of expectations, relative performance, surprise, and disappointment in affecting health, which are not highlighted by standard models of health capital, but are consistent with reference-dependent utility. I also discuss potential implications for employment contracts in terms of a trade-off between ex post health and ex ante incentives for productivity.

The paper is here, and for the pointer I thank the excellent Kevin Lewis.


Anybody know of an ungated copy?

An obvious mechanism is that winners dope more than losers, and that there's something systemic about individual sports (say that a higher fraction of them are endurance sports) that inclines them to use more dangerous doping agents like EPO, rather than less dangerous doping agents like steroids.

But there are things I don't understand without access to the paper. These are all track and field athletes, right? So they are likely heavy dopers by Olympic standards. What are the team track and field sports? I assume they deal with my suggested mechanism, but how? There are also likely systemic differences between sports in all sorts of respects.

Relay races.

Is that it?

I suspect a sizable factor in killing off champs is car crashes -- gold medalists get cars, losers take the bus.

David Wallechinsky's encyclopedia of the Olympics is full of car crashes killing off Olympians a few years later.

Suicides come up less often in the book. The Japanese bronze medalist in the 1960 marathon committed suicide in 1963 (I believe) because he couldn't cope with the national expectation that he would win the marathon in Tokyo in 1964. But that story stands out in my memory. In contrast, ex-Olympians dying in car crashes are a dime a dozen in Wallechinsky's accounts.

This is a completely reasonable thing to look for in their robustness checks. I'd at least have had an RA catalog all the causes of death for a particular year's medalists, and all the causes of death for a particular year's deaths.

I'm guessing it's easier to find the car crashes than the suicides, though.

Right, car crashes leave a public record.

Other early deaths are murky about cause and effect. Football player Lyle Alzado said he was dying of cancer due to steroids. (Others whispered he had AIDS.) Walter Payton had all sorts of health problems before cancer killed him. (On the other hand, I almost died of cancer at 38, but nobody ever accused me of looking like I was on steroids.)

PEDs are a not unreasonable suggestion.

I wouldn't be surprised if German academics have done a study by now of lifespans of East German Olympians. Of course, they tended to get their PEDs under more careful control than Western athletes who freelanced with Dr. Feelgoods.

Okay, now I've read Levine's study.


It's a careful one, but so careful it excludes most of the interesting factors, such as PEDs, car crashes, and war.

It's for male track and field finalists in 1898-1948, so no modern PEDs (which may have first played a role in throwing events in 1956). 53 war deaths, but they are excluded. Only 2 auto accident deaths (my recollection is that most of the Olympic hero car crash deaths are in the 1950-1975 era, typically in Second or Third World), but they are excluded. 2 murder victims are excluded.

Individual gold medalists average about one year less lifespan than non-gold medal finalists, but all the difference in mortality is from about age 75 onward. Relay race gold medalists live longer, however.

This study is interesting because it mostly contradicts earlier theorizing that high status makes you live longer. But, since all the individual event differences in mortality are out in old age, it's not all that interesting by itself.

Ha! Thank you, this is helpful.

Yeah, I should have read it before theorizing off of the abstract.

Anyway, it's a good study for what it's designed for. It's designed to follow up studies of whether baseball Hall of Famers and Oscar winners live longer than their colleagues who fell short of grabbing the brass ring. (There is some oft-cited British study about how bosses live longer than their underlings because, supposedly, low status just makes you want to roll over and die, or something.)

It would be interesting to study life expectancies in a field of striving that's more extremely Death or Glory. Life expectancies of ambitious mountaineers in the Himalayas, for example.

Athletes competing for Olympic medals can be very, very close to each other in skill and outcome: tenths of a second or tenths of a meter can separate gold medal from fourth place. Situations like Usain Bolt blowing everyone else out are rare. Even Michael Phelps had very close finishes against almost-as-good opponents during his dominating run.

So no, if Olympic athletes all dope, I wouldn't expect winners to be meaningfully more doped than losers. These things are won on the margins.

I'd think the opposite. Winning is dose dependent, and the winners are generally made up of those more willing to experiment and try novel doping strategies or endure worse side effects, both of which may increase long term risk. It's hard to look at the last fifty years of sport and not think that.

In any event, it's an obvious mechanism to consider. It might be wrong, but the paper must have done heavy lifting to rule it out to be taken seriously. Perhaps it did - I can't read it.

You're making a ton of assumptions about the way doping works that I don't think survive scrutiny. For example, you're assuming that it's just a "yes"/"no" or "more is better" kind of thing. Doping can be done incorrectly, which might be correlated with both slightly worse performances (losing) and dying earlier. Doping "correctly" would lead both to exceeding expectations and not suffering as many adverse effects later in life.

I find the argument in the OP much more convincing a priori. The fact that data backs it up is much more convincing to me than hypothetical correlations with doping.

I am assuming something close to "more is better," but I think that's a correct assumption. Certainly the introductions of new kinds of doping have led to new performance plateaus in various sports (football, hockey, and baseball have all had drugs define eras). And within some broad classes of drugs, like say steroids, more is better up to very high levels. Levels that would never pass Olympic testing. sometimes discusses this kind of thing, often in the context of cycling and running. Particularly when the drug-soaked Tour de France comes around each year.

I think the questionable part of my analysis is whether high levels of PEDs really would lead to lifespan problems, controlling for the other aspects of athletic life. There's reason to suspect that, but I don't think it's well understood. PEDs, other than testing for them, are sort of a taboo topic.

"Winning is dose dependent"

That was clearly true for Ben Johnson in the 1988 100m, who had improved so much since 1984 by juicing so massively that he got caught even by the tests of 1988.

The other star of 1988 sprinting, Flo-Jo, appears to have been clean until the offseason between 1987-1988. I saw her win a silver medal at the LA Coliseum in the 1984 Olympics and she was willowy compared to the massive gold medalist Valerie Brisco-Hooks, the wife of NFL star Alvin Hooks. At the 1987 world championships, Flo-Jo came in second to an East German lady who later revealed to be juicing. Without winning a gold medal, it was hard to make a living in sprinting back then, so after getting cheated by juicers in her two big chances, Flo-Jo called up Ben Johnson for some training advice in late 1987. She showed up in 1988 looking like Wonder Woman and set records that, hopefully, will never be broken. She retired when better drug tests were threatened in 1989 and died about ten years later.

I suspect the drug tests even out the playing field -- it's hard to get away with taking gigantic doses, so the effect is that everybody takes some steroids, but not as much as Ben Johnson or Flo Jo did in 1988.

Tests increase the reward for trying novel compounds and techniques.

> the effect is that everybody takes some steroids, but not as much as Ben Johnson or Flo Jo did in 1988.

This is probably correct, but they are probably taking more of other things.

In sports like cycling, track & field and even wrestling, the use of PEDs has been very nearly universal and common knowledge among the participants. The real miscreants in these drug-fueled contests are those that didn't use drugs. Competing when they knew they could not win against dopers, they merely added legitimacy to the proceedings. Now many of them rail about how they were cheated of their opportunity for glory when, in fact, they had no such chance in any case and were merely tourists in a country where they were reluctant to embrace native ways.

A British Olympian has just died from unexpected heart problems. Well, unexplained heart problems anyway. She seems to have known she had a problem.

I have to say that it is very charitable not to scream steroids looking at the pictures.

Ben Johnson is supposed to have got caught because of blood doping. He was taking his own blood, storing it, and re-injecting it before the race. Alas, while his body removed the steroids, the blood in the fridge remained pristine. Flo-Jo was never caught. Even though, obviously, she was taking something.

In fact testing has not had a good record lately. How many baseball players were caught by tests? The Feds seized the company records. It looks more and more like testing exists to divert the rubes, not to catch the guilty.

The current testing keeps the cheating within certain bounds. Ben Johnson was so out of control his eyes had turned yellow -- even 1988 level testing caught him.

It didn't catch Flo-Jo, but nobody has come close to her records in 26 years.

"I demonstrate that an athlete’s performance relative to his expectations partly explains the earlier death of winners in individual events - " What are the expectations ? That they would win Platinum?
It may be interesting to study Silver medalists. There are enough pictures circulating showing how disappointed they are . The Bronze medalists , by contrast , seem thrilled . Do they live longer?

"partly explains the earlier death of winners in individual events": I have always thought that the statistical meaning of "explains" is unfortunate. If used in writing for laymen, it can be tantamount to a lie. "Significant" has some of the same problems.

With how close most Olympic track and field events are, everyone who medals expects to win. If you don't go in with that expectation you're gonna lose. OTOH, silver medalists are the saddest Pandas. Gold medalists have gold, bronze winners made the podium and silver medalist came "this close" to winning it all.

Seinfeld has a routine on that.

Has anyone seen Prior Approval since yesterday?

You mean, Prior to Today?

Why do you ask, because somebody peppersprayed TC?

I heard he got his a** kicked by some students. I just figured it was 2nd graders, though.

Where can I buy Olympic Gold Medals from? I want to extend my lifespan.

How could gold metal winners be a large enough data set to tell much?

They didn't just look at gold medalists (of which there must've been hundreds by this point in track and field).

And frankly, I wouldn't be surprised if you were able to get statistically significant results based on 50 data points.

"However, winning an individual event is associated with an earlier death."

Car crashes have killed a lot of Olympic gold medalists. I noticed that years ago while reading David Wallechinsky's "Complete Books of the Olympics." A lot of the heroes celebrated in the book were dead with a decade or two from car crashes.

My impression is that in the mid-20th Century winning a gold medal facilitated obtaining a car, including in countries where cars were rare and accident rates were high. Also, athletes tend to have a Need for Speed and might be highly competitive about things like passing on narrow highways.

In the U.S., where you don't need to win a gold medal to be able to afford a car, the history of track and field near its peak of popularity in the 1970s is full of car crashes:

"On May 30, 1975, returning from a party, [Steve] Prefontaine was driving on Skyline Boulevard, east of the University of Oregon campus near Hendricks Park when his orange 1973 MGB convertible swerved into a rock wall and flipped, trapping Prefontaine underneath it.[13]"

"Twice in 1977 Mary Decker escaped auto accidents unhurt; in one, her car rolled over four times."

She had trouble on the track, too.

Comment of the Year!

That was meant to appear under Yancey Ward's comment. Either damn software or damn me!

Hey, I was writing that in all seriousness:-) I thought about Mulp, too, but he left a rambling, incoherent comment in the next thread, so he isn't in jail, unless it has WiFi.

I don't know about Mulp, but, in seriousness, judging by what he writes, pa must be a lot older than the young man they showed in the video.

Comments for this post are closed