This article appears in the June 23, 2014 issue of National Review.
As the World Cup approaches, soccer fans from around the world are preparing for a bacchanalian soccer binge that will inevitably lead to a deep emotional crisis for all but one nation. Even with the many disappointments, the rapture of fans in the winning nation is so great that the World Cup undoubtedly contributes positively to worldwide happiness. World Cup revenues, after all, are projected to be northward of a billion dollars, with tourism spending piling billions on top of that.
So who will the lucky winner be? Economics has a surprisingly large amount to say on the subject. Forget Thomas Piketty: By far the most important academic study out this year is Goldman Sachs’s massive “The World Cup and Economics 2014.”
In order to calculate the chances of success for countries in each round of the tournament, the Goldman Sachs economists who authored the study drew on data going back to 1960. Discounting friendly games and focusing instead solely on mandatory international matches, they tested the ability of several different variables to predict the winners of about 14,000 such contests. These variables included whether a team was a host to a match, whether a team was playing on its home continent, the number of goals scored by a team in its previous ten matches and the number scored on the opposing team in its previous ten matches, whether the match was a World Cup match, and a composite measure of a team’s success called the Elo ranking.
The authors developed the best possible econometric model drawing on those data, and used it to generate a prediction that Brazil has a 48.5 percent chance of being the World Cup champion. The accompanying table suggests that fans from other countries should be worried. For reference, the table below indicates in the far-right column the Goldman Sachs prediction for this year. To the left of that, for historical perspective, is the output of the model for the previous tournament.
Each column shows the predicted probability that each team would make it to a certain round in the tournament. So for Brazil in 2010, for example, the probability was 26.6 percent that it would win, 39.2 percent that it would make the final, and so on. The cells are highlighted to show how teams actually performed—for example, Germany made it to the second round, then the quarterfinals, and the semifinals, while Portugal made it only to the second round and Italy and France, despite good odds, failed to advance at all.
The model performed remarkably well in 2010, with the final including two of the top three teams, and the semifinals three of the top four. So it is quite likely that the semifinals will, after all of the drama, include Brazil, Spain, Argentina, and Germany.
And from that scenario, one can be sure of two things. First, if Brazil wins, the celebration in the home country will be so wild that the rest of us will simultaneously be sorry we are missing it and glad we are not there. Second, if Brazil loses, it will be one of the bigger upsets in soccer history.