Economic Countdown To The Olympics 2: Predicting Olympic Medals
In the second part of our five-part series on The Olympics (Part 1 here) we ponder the impossible to predict - the medal count. As Goldman notes: Economists like to think that the toolkit of their profession helps them explain many things or, as some would claim, everything that is interesting about human behavior. In the context of the forthcoming Olympic Games in London, therefore, the key question is whether economic variables can help explain and predict success at the Olympics itself. At one level, this seems like a daft question even to consider. It is hard to imagine that economic variables could even begin to capture the kind of individual skill, mental determination and hunger that drive athletes to perform feats of unimaginable virtuosity that is the stuff of Olympic legends. But at the level of a country, it may be possible to identify the ingredients that unlock success at the Games. As British Paralympian Tim Hollingsworth explains: "...when you create a world class environment you are far more likely to create world class athletes." What is a 'world class environment' and how do we measure it across countries? Luckily, we have an answer in the GS Growth Environment Scores (GES), a broad measure of growth conditions across countries - and, indeed, this is what we find: gold does go where the growth environment is superior. The forecast leaves USA, China, and Great Britain battling it out for 'Most Golds' and USA leading China overall - but remember "the most important thing in the Olympic Games is not winning but taking part."
Jose Ursua and Kamakshya Trivedi, Goldman Sachs:
Gold Goes Where Growth Environment Is Best—Using Our GES to Predict Olympic Medals
Let’s Predict Olympic Success in London 2012
One of the ‘fun’ parts of building econometric models is that if they are reasonably good at explaining the past, they can then be used to forecast the future. In this case, we want to predict the eventual country medals tallies at London 2012.
Not all sports are created equal, and the relationship between Olympic success and the economic variables we have discussed may differ across sports for a number of reasons. It is more expensive to participate in some sports, such as Sailing or Equestrianism, and this may limit participation from low-income countries, but success may also vary on account of different traditions, culture and the demographic characteristics of the population or the geographic features of each country.
The baseline model that we use to do this is a panel regression featuring as explanatory variables our headline GES, a host dummy, population and lagged medal attainment. As a starting point, note that the model does fairly well at explaining medal outcomes from previous Olympics, or that it works well in what economists call ‘in sample’.
Chart 2 above shows the ‘in sample’ results, where we compare what the model would have predicted in the past four Olympic Games (which is as far back as our GES extend) versus what actually happened in terms of medal attainment. The 45° degree line represents a perfect forecast, and the fact that our scatter shows a high concentration of points around that line makes us quite confident that our model can satisfactorily predict Olympic success.
We are now finally ready to unveil our predictions for London 2012. Table 3 (below) shows our forecasts for medals in London 2012—gold on the left-hand side, and total number of medals on the right-hand side. In each case, we also show each country’s medal attainment in Beijing 2008, and the difference between the 2012 forecasts and the latter (a positive number means we expect the country to win more medals in 2012 than were attained in 2008).
Our forecasts reflect two very clear patterns revealed in our analysis. First, countries with superior growth environments and higher incomes are expected to win more medals, and, second, there is also a marked host effect that will likely bump up the number of medals attained by Great Britain (11 more gold medals, and 18 more overall). The US is expected to lead the group, with something close to 36 gold medals and 110 overall. Our estimates also predict that the top 10 ranks will include five G7 countries (the US, Great Britain, France, Germany and Italy), two BRICs (China and Russia), one N11 country (South Korea), and one additional developed and emerging market (Australia and Ukraine, respectively). Based on our analysis, these 10 countries will likely capture more than half of all medals attained during the Games. The rest of the countries in the list are a diverse mix from various parts of the world, broadly led by Europeans and including Asian, Latin American and African nations.
So, ultimately, it appears that gold does go where growth and the overall growth environment are best. For the next few weeks, our forecasts will be competing against other predictions for validation in the real world. Like many people around the world—economists and noneconomists alike—we will be eagerly following the medal outcomes of the long-awaited London 2012 Olympics. As Coubertin himself said, “the most important thing in the Olympic Games is not winning but taking part.” We believe the same applies to our forecasts, and although we would be delighted to see them hit their targets perfectly, we will have succeeded as long as they have provided some food for thought. The die is cast—enjoy the Games!