What statistic is the best indicator for midterm election outcomes?

The common response might be: unemployment. After all, it's the number hat defines our weak economy. The average person on the street might not be able to tell you 2010's quarterly GDP growth, or the year-over-year change in federal revenue, or overall decline in home values. But most people know where unemployment is: It's near 10 percent, and it's not budging.

But according to political scientist Seth Masket, the relationship between the unemployment rate and midterm losses is weak. Below is a chart comparing unemployment to midterm election results from the last half century. Three lessons here. First, every president except for Bush II lost House seats in the first midterm. Second, there is a weak correlation between unemployment level and House losses. Third, the only midterm election in the last fifty years with unemployment anywhere near double digits was in 1982, when Reagan lost 26 seats, so beware pundit explanations like, "Unemployment this high usually results in big losses for incumbents." Better to say, "Historically, unemployment this high is nearly without precedent, so we don't know what it 'usually' does."


Forget about unemployment for a minute. There's another statistic that receives less media attention but offers a clearer indicator of midterm elections. It's real disposable income -- the money you have after taxes, adjusted for inflation. Here's the relationship between RDI and House losses (data from the Bureau of Economic Analysis.) It's not perfect correlation, but the trend is clear. When folks feel richer, they reward the party in power.


Masket is tracking income between the third quarter in the year before the election and the second quarter of the election year. If if divide Q3-2009 into Q2-2010 figures, we get 0.015 -- or 1.5 percent income growth. On the X-axis, that would put 2010 between Reagan 1982 and Clinton 1994.

You might be wondering why real disposable income figures aren't even worse, considering the severity of the recession and the stickiness of our high unemployment. First, the government has sent hundreds of billions of dollars in checks to Americans to keep them spending, through the stimulus and unemployment insurance and Census payrolls. Second, consider the "real" in real disposable income means adjusted for inflation. In the 1970s and early 1980s, high inflation gnawed at families' buying power. Today, inflation is so low we're worried it might turn negative. Low inflation is bad for the economy because it means families and companies are slow to spend money. But low inflation is OK if you've got money to spend because it keeps prices down.

As my colleague Dan Indiviglio recently reported, the latest income statistics add insult to injury, with government money winding down and private incomes slow to pick up the slack.

Of course, no two elections are alike, and one statistic -- even one key statistic -- in an election just an ingredient in a very large stew. But as you watch the returns tomorrow, remember this indicator as evidence that Democrats are running into a strong headwind. It is the number that best gets at the question, "Are you better off now than you were a year ago?"