February 26, 2012 by Brian
This was printed in the Boulder Daily Camera on February 25, 2012.
Which economic indicators best correlate with presidential election results? Last year New York Times statistician Nate Silver presented an elegant answer to this question. For the sixteen presidential elections since World War II, he computed the correlation between the incumbent party’s margin of victory and the value of 43 indicators in the first nine months of the election year. The results? Change in employment rates matter. Market indexes and oil prices don’t.
The Institute of Supply Management’s manufacturing index best correlates with incumbent party victories, with a 46 percent correlation. Close behind are changes in non-farm payrolls and changes in the unemployment rate – both above 40 percent correlation. Since World War II, incumbent presidents ran for reelection seven times. Only Jimmy Carter and George H.W. Bush lost – both when the unemployment rate increased.
Note that the change in unemployment rates matter, not the rate itself, which had zero correlation. Meanwhile, gain of the Dow Jones index had only a six percent correlation. Silver also found a 15 percent correlation between lower gas prices and an incumbent victory.
While the unemployment rate has been decreasing for about a year, it’s not necessarily a good sign for Obama. The rate has decreased partly because many have stopped looking for work. Classifying these people as unemployed would increase the unemployment rate by 1.25 percentage points, reports the Congressional Budget Office. Worse for Obama, economist James Sherk shows that despite job growth, this is the “weakest recovery in more than half a century.”