Prediction markets allow people to bet on the outcome of elections and thus aggregate people’s expectations of what will happen on election day.
Prediction markets in the PollyVote
We are aware of only one prediction market forecasting the national popular vote in U.S. presidential elections, namely the Iowa Electronic Markets (IEM). Polly calculates a weekly average of daily forecast data from the IEM in order to reduce outlier effects of short-term spikes.
History of prediction markets
Prediction markets are often thought of as a relatively new forecasting method. Yet, such markets have been around for at least half a millennium, long before the emergence of scientific polling. The earliest account of political markets dates back to 1503, when betting on who will be the next pope was already considered old practice. In the US, political betting was common in the 19th, where public bets on a candidate were considered a sign of support. However, the heyday of election betting was between 1884 and 1940, when large-scale markets on presidential elections were operated. These markets not only provided accurate forecasts of the election outcome in an era before scientific polling, they were also widely popular among investors and journalists alike. At certain times, the trading volume in these markets exceeded that in the stock exchanges on Wall Street and major news outlets such as the New York Times, Sun, and World reported the betting odds as forecasts of the election outcome on a nearly daily basis.
Forecast accuracy of prediction markets
Prediction markets are an effective means to forecast elections. Evidence from different countries and elections shows that markets often provide more accurate forecasts than established benchmark methods such as polls, quantitative models, and expert judgment.
However, prediction markets do not meet the accuracy levels of the combined PollyVote forecast. The following chart shows the mean absolute error of the PollyVote compared to the IEM vote-share for the three elections from 2004 to 2012, calculated across the remaining days to Election Day. That is, at any given day, the chart shows the average error that one would have achieved by relying on either the PollyVote’s or the IEM’s forecasts until Election Day. For example, if one had relied on the IEM forecast starting 96 days before the election, an average error of 1.2 percentage points would have resulted. In comparison, the corresponding error of the PollyVote would have been 40% lower, at 0.7 percentage points. Note that the error of the PollyVote was consistently lower than the error of the IEM.