China's Winning Ways

Before the opening ceremony of the Beijing Olympics, I wrote about economic models that aimed to pre-empt thousands of hours of television coverage, sappy features, and tape-delayed tension. John Hawksworth of PriceWaterhouseCoopers and Andrew Bernard of Dartmouth's Tuck School of Business both aimed to predict national medal hauls based on factors like home-field advantage, the size and growth of national economies, and past political affiliations. (You can check out the PriceWaterhouseCoopers projection here and the 2008 Bernard projection here.)

So how did the economists do in the individual projection event? Both performed something like the American track-and-field team—several high-profile triumphs and a couple of splashy pratfalls.

As was the case in 2004—here's the 2004 Olympics postmortem—both analysts underestimated the strength of the U.S. delegation. The final medal count from Beijing has the United States first in total medals with 110 (36 golds), followed by China (100 medals, 51 golds), Russia (72 medals), Britain (47), and Australia (46). Bernard gets a 10.0 for nailing the steady performance of the Americans, having projected that the United States would win 105 medals, down only slightly from the 2004 total (108). Bernard also nailed the number of golds the Americans would bring home, guessing 36 on the button. London-based Hawksworth, however, gets closer to a 2.5, predicting the U.S. medal haul would fall 20 percent from 2004, from 108 to 87.

In a year when all the major predictive factors favored China, both economists nonetheless seriously underestimated the host nation's performance. Hawksworth projected China would increase its medal haul from 2004 by 40 percent, from 63 to 88. Bernard saw China making more modest gains, up to 81 medals (with 37 golds), and coming in third behind Russia. Instead, China came in second in the medal total with 100, a 59 percent increase from 2004, and lapped the field in gold medals.

On Russia, the two analysts also diverged. Bernard predicted the country's medal take would hold steady from 2004, at 92 medals (including 25 golds), while Hawksworth guessed it would decline to 79—much closer to the country's total of 72. Bernard nailed South Korea and Cuba but misfired on Japan (40 projected, 25 won), Romania (17 projected, eight won) and Jamaica (fewer than six projected, 11 won). For his part, Hawksworth had France pegged (40), but he overestimated the performance of perennially emerging Mexico (three medals vs. a projection of eight) and underestimated the strength of Kenya's long-distance runners. (Kenya came in 18th in the nations' medal race but didn't register in Hawksworth's top 30.)

What accounts for the divergence between model and results? Both models include a home-country effect, which projects that host-nation athletes, fueled by crowds and increased investment, punch above their weight. Clearly, the combination of intense nationalism and rapid economic growth had a greater impact on China's results than might have been projected. Looking over the results from 2004 and 2008, there seem to be two other factors at work here: a post-hosting letdown or a prehosting boost. I've noted in the past that one of the pitfalls of forecasting (political and economic) is to extrapolate from recent results. Greece turned in a great performance when Athens hosted the Summer Games in 2004 with 16 medals. Both Bernard and Hawksworth predicted Greece would win 15 in Beijing. But the original Olympians had a very lame Olympics, winning only four medals and no golds. There may also be a sort of pre-Olympics effect at work, by which the next country in line begins to kick it up a notch in anticipation of its hosting duties. Both Bernard (31 medals) and Hawksworth (28) projected a modest Olympics for Great Britain, but the Brits—Olympic hosts in 2012—ended up fourth with 47 medals.

Another interesting trend: Despite the spread of wealth around the world and the increasing decentralization of the global economy, the rich generally stayed rich. The top 10 countries won 532 medals in 2004, compared with 519 in 2008. The top 30 won 796 medals, up from 776 in 2004.

Predicting Olympic outcomes from a top-down perspective (as the models do) or from a bottom-up perspective (consulting experts in every sport and picking potential winners) is an extraordinarily difficult endeavor. As I noted before the Games began, there's a big element of chance and randomness—think of the one one-hundredth of a second that separated Michael Phelps from Milorad Cavic in the 100-meter butterfly or the unexpected flubs of the American sprint-relay teams. And I still maintain that these models have to take better account of cultural phenomena, such as a nation's openness to immigration. Among the winners for the United States were gymnast Nastia Liukin (born in Moscow) and, my favorite of these Games, Henry Cejudo, the 20-year-old son of illegal immigrants from Mexico who became the youngest American to win gold in wrestling. Nothing is certain in this world. But it's highly likely that had Cejudo's family stayed in Mexico—or if they had been deported soon after their arrival—the United States would have won only 35 gold medals in Beijing.