By Daniel Gross
But there's a robust, and surprisingly accurate prediction business among economists for the summer Olympics. (I covered these sets of predictions in 2008 and 2004.) For the Winter Olympics, which feature fewer competitors and many fewer events, the number-crunchers typically put down their keyboards and spreadsheets. All except economist Daniel Johnson, who seems uniquely qualified to prognosticate on such matters. He was born in Canada, teaches at Colorado College, and (as his C.V. notes) is way into curling.
As Johnson explains, the model he uses is pretty simple. The economist looks at five variables: population, income per capita, climate, political structure, and host-nation advantage. The results are quite intuitive. The bigger a country's population, and the more resources per capita the country has to devote to athletics, the more world-class athletes it is likely to produce. Climate, of course, is a huge factor in the Winter Games. "If you grow up skiing, the chances are you'll be a better skier as an adult," Johnson says. "Even when you control for income, cold countries do better than warmer countries." The political climate factor is somewhat counterintuitive—the shrinking base of nations with one-party rule tend to do quite well in Olympic events. (Think Cuba and China.) Tyranny may not be a competitive advantage in global economics or cultural competitions, but it apparently remains one in global athletic tournaments. Finally, Johnson notes that in the Winter Games, the host nation generally wins about three medals more than might be expected, one of them gold.
Johnson adds that there is one "slush factor" that is more difficult to measure: a nation's sporting culture. In the Summer Olympics, Australia generally punches above its economic weight because the Aussies are a tough, buff, sporting nation. "The same holds true for Norway in the Winter Games," he says.
Johnson's predictions for 2010, which include only those nations expected to win 10 medals or more, can be seen here (PDF). Canada, which typically does well in the Winter Olympics, is expected to ride the host-nation status to a world-leading, 27-medal haul, though only five of them are expected to be gold. The United States is expected to tie Norway for second, with 26 medals, to be followed by the traditional European powers: Austria (25), Sweden (24), Russia (23), and Germany (20). The Unites States, Canada, Austria, and Russia are expected to win roughly the same number of medals they did in 2006.
Drill down, though, and you'll see that the model predicts some significant changes from 2006. The Nordic countries are expected to make big gains: Finland is expected to win 14 (four gold), up from eight (zero gold) in 2006; Sweden's medal haul is expected to nearly double, from 13 in 2006 to 24 this year; and Norway is expected to take home 26 medals, up from 19 in 2006. Another big mystery: Italy, which has fared poorly in population growth, income, and political climate over the past four years, is expected to win 19 medals, up from 11 in 2006. Is Tomba la Bomba coming back?
I'm not precisely sure what accounts for these shifts. One factor boosting Scandinavia's stock is the stability of the nations' economies in recent years. Some of these changes can also be explained by a recalibration of Johnson's model. Essentially, he filtered in the fact that Nordic and Italian athletes have done well in recent Olympics (in part because of the home-field advantage afforded by the recent games in Turin and Lillehammer).
The countries that Johnson expects to slide include Germany—the model projects its medal total will schuss from 29 in 2006 to 20 in 2010—and China. Despite China's impressive economic growth and impressive achievements at the Beijing Olympics in 2008, Johnson's model projects the country won't meaningfully boost its medal count in Vancouver. "They've been growing really rapidly, but China's per capita income isn't that great compared to many of the other countries," says Johnson. In addition, China hasn't devoted the type of resources to these Winter Games that it did toward the summer events on its home turf.
How much stock should we put into these projected medal counts? Over the past few Olympics, the model has a 94 percent correlation with the actual medal count per country, and about an 88 percent correlation with the actual gold medal haul per country. And unlike the ground-up projections that ESPN and Sports Illustrated make, it doesn't rest on gauging which athletes will win which events. It's just a matter of crunching numbers. "On average it's been large, rich, cold nations that win medals," said Johnson. "And they do so in a fixed proportion."
Of course, economic models are fallible. (Just ask AIG.) And there's reason to think that a model for the Winter Olympics—and with this Winter Olympics in particular—might have problems. First, the sample is much smaller than the that of the Summer Games. There are 86 events and 258 medals to be awarded in Vancouver; there were 302 events and more than 900 medals handed out in Beijing. A few slip-ups, or one set of heroic performances, can alter a country's performance beyond the model's historical margin of error.
Second, the Winter Olympics are overwhelmingly composed of individual competitions rather than team ones. Yes, there's hockey, curling, and the four-man bobsled. But pretty much everything else is an individual event. The Summer Olympics, by contrast, have a large number of purely team events (baseball, soccer, handball, basketball) and events contested by teams as well as individuals (gymnastics, track, swimming). In other words, a country's performance in the Winter Olympics is much more dependent on the performance of a few key individuals. The absence of a single athlete—say Lindsey Vonn—could mean the United States wins 10 percent fewer medals. And it may be a cliché, but the margin of error in winter sports is tiny. One slip on the ice, one missed gate in the slalom, one poorly executed double axel, and the favorite can go from sure thing to last place. All of which make predictions of any sort extremely difficult.
Third, Canada will undoubtedly get a boost as the host nation. But I'm wondering if the model shouldn't give the United States something of a host-nation boost. Vancouver is only about 30 miles from the U.S. border. Large numbers of American fans are likely to make the trek and will be chanting "USA!" and ringing cowbells. Athletes who live and train in the Western United States won't have to adjust to a new time zone or to a vastly different climate.
So let the games begin! In a couple of weeks, we'll check back with professor Johnson and see how his model performed this year.
0 comments:
Post a Comment