Is there an Olympic home advantage?

Olympic home advantage?

I really enjoyed following the Olympic games in Paris this summer and one thing I noticed is the amazing perfomance of French teams and athletes. Don’t get me wrong, all athletes were perfoming outstanding, and maybe I am influenced a bit by the defeats of the German Basketball teams (both M and F) against France, but I started to wonder if there is such a thing as an Olympic home advantage.

I have done some work on referee bias, therefore I know that there is a well documented bias of referees in football (soccer) to advantage the home team, mostly attributed to social pressure, and high profile teams, maybe because of career concerns (see my paper on this). But is there also an advantage more broadly for athletes at the Olympics, where in many disciplines referees do not have a big influence? A quick literature research revealed: Most likely. But since I found a great data set collected by Randi Griffin, and I enjoy a bit of data analysis, I decided to give the hypothesis a basic test.

Which Olympic games to consider?

The first Olympics held by the IOC were in 1896. Should we consider all the Olympics since then? Especially in the early 1900s, athletes may have faced travel restrictions and not been able to attend the Olympics for this and other reasons. Then the home country might have an advantage simply because of this. I decided to plot the share of athletes that competed in the Olympics over time. It seems that the share only remained constant after the second world war. I decided to then only use data from after 1946.

The figure shows the share of athletes from the host country competing in the Olympics. The shaded areas are 95% confidence intervals.

Is there a home advantage? Country-level evidence

How can we test if there is a home advantage for athletes at the Olympics. My first idea was to simply comnpare the share of medals per competition that athletes from hosting countries won and the share that those from non-hosting countries won. The first panel in the figure below shows that athletes from hosting countries are much more likely to win a medal in a given competition.

The figure shows the average share of medals per athlete and competition in hosting and non-hosting nations. The spikes indicate 95% confidence intervals.

But hosting the Olympics is quite an expensive endeavor, hosting countries must be commited to the sport, this could also mean they spend more on athletic development or are economically better off than non-hosting countries. How should we countrol for this? I decided to calculate the difference in the share of medals per competition a country won between a given year and the previous Olympics. This difference should be roughly 0 if there is no home advantage. The figure below shows that non-hosting countries actually earned a similar amount of medals per competition and athlete in any two consecutive Olympics, but hosting countries earned more (though this difference is insignificant).

Difference in share of medals per competition

The first panel instead shows the difference between this share between the present and the last games. The second panel restricts the sample to only nations that ever hosted the games. The spikes indicate 95% confidence intervals.

A second concern I had was that hosting-countries might just be substantially different from non-hosting countries. Therefore in the second panel I restricted the analysis only to countries that have ever hosted the Olympics, but again, there is a slight advantage for hosting countries.

Athlete-level evidence

Prior to hosting the Olympics, a country may step up it’s search for newcomers and young athletes to compete in the “Home Games”. So maybe, we should look at the perfomance of individual atheletes in consecutive Olympics, one in their home country and one in a different one. I do this in the next figure. The first panel shows the difference between the average medals per competition of an athlete in Olympic games hosted by his home country and one abroad (home medals minus abroad medals). The figure gives the clearest picture yet of a potential home advantage and does not disappear when focusing on the hosting nations only (second panel).

Difference expected medals hosting nations

The first panel shows the difference in medals per competition for athletes of hosting and non-hosting countries between the previous Olympics and the present ones. The second panel restricts this analysis only to athletes from countries that ever hosted the Olympics. The spikes indicate 95% confidence intervals.

Additionally, I decided to not only control for past but also future performance. Of course now, the sample drops to athletes who competed three times in the Olympics, which is definitely not the normal amount. Nevertheless, the picture in the first and second panel below remains roughly the same.

The first panel controls for the performance in the Olympics after the home games. The second panel again restricts to only athletes from nations that ever hosted the Olympics. The spikes indicate 95% confidence intervals.

Is there a home advantage and why?

So, is there a home advantage? The data seems to point in this direction, but there are many other reasons, why the figures could show these patterns, most importantly, the controls for expected performance are not perfect and the athlete pool constantly changes. But if we were to assume that there is a home advantage, why is it there?

There are many reasons that could explain a great performance of atheltes in hosting countries. The data I used does not allow a detailed analysis of them, but I like to think that the home crowds push athletes such as the French basketball players to new hights (pun intended).