A simple, surprisingly good way to predict elections
PUNDITS, journalists, activists and others spend a lot of time during election season trying to figure out what factors will sway the vote one way or the other. Could he be like a candidate? Differences in policy? Her advertising strategy? His habit of insulting large sections of the electorate?
Campaign choices are important, but political science research indicates that the state of the electoral landscape is largely shaped by factors beyond the control of the candidates. People vote afterwards, based on their opinion of how things are going for them and those around them. But their views are shaped by all sorts of things. As Christopher Achen, of Princeton University, and Larry Bartels, of Vanderbilt University, explain in Democracy for Realistsbook published earlier this year, voters sometimes punish politicians for bad weather, random misfortune – shark attacks apparently cost Woodrow Wilson votes in the 1916 election – and performance of America’s local soccer team.
Perhaps more importantly, voters are very myopic. They are concerned about their economic welfare, and especially about whether their income is growing or not. In making this judgement, however, the months immediately before the election are crucial. Messrs. Achen and Bartels believe that few variables matter anywhere as much as growth in disposable personal income, adjusted for inflation, in the six months leading up to election day. In fact, they note, only two variables—short-term income growth and incumbent party tenure—represent the most reliable predictors of division the existing party vote (see chart). These measures predict a close race in 2016. In fact, a careful electoral model compiled by Yale University’s Ray Fair, based on economic fundamentals, predicts a Republican victory in the November. While real income growth is running at a respectable 1.4% annual rate, the Democrats have been in office for the past eight years and the president is unpopular himself on the ballot.
Polls, in fact, show more and more Democratic influence, and perhaps a strong one. This difference is due to the fact that short-term economic growth and tenure are not everything; other factors—such as wars, social unrest, or the presence of an incompetent and historically marginalized candidate on the ballot—also matter. But if Mrs. Clinton wins on November 8, this surprisingly effective model of election performance indicates that the road to re-election will be very difficult. It would take disposable income growth of more than 2% in the months leading up to the 2020 election to offset the effects of 12 years of Democratic incumbency. The road to a second term will be smoother if Republicans again nominate a poor candidate. If Mrs. Clinton hopes to beat the garden variety Republicans in 2020, she’d better think hard about how to get Americans’ real incomes growing soon before the end of her first term.