Skip to main content
Obama States and Unemployment: Confusing Cause and Effect

I like Ross Douthat, the New York Times’s newest regular columnist, but the case he presents today on economic conditions in blue states and red states is misleading. Typical of the piece are claims like these:

Meanwhile, California, long a paradise for regulators and public-sector unions, has become a fiscal disaster area. And it isn’t the only dark blue basket case. Eight states had unemployment over 11 percent in June; seven went for Barack Obama last November.

This is true — states which voted for Barack Obama have higher unemployment rates than those that don’t. In fact, the difference is statistically significant: Obama states have an average of 9.3 percent of their population unemployed at the moment, versus 7.9 percent for McCain states.

The problem is that Douthat is confusing cause and effect. Those states may very well have voted for Barack Obama because they had higher unemployment. Obama blew out McCain in high-unemployment states like Michigan, Oregon and Nevada — states which are normally much more competitive. He won, to much surprise, Indiana and North Carolina, two states with unemployment rates well above the national norms. Indeed, although the Obama states have higher unemployment now, they also did so by a statistically significant margin in November, when Obama was elected.

The other, probably more obvious flaw with Douthat’s logic is that Barack Obama has no more authority to govern Vermont than he does Wyoming. Instead, the voters in each state elect governors and legislatures to represent them. And sometimes, these decisions do not match their Presidential preferences: Vermont has a Republican governor, for instance, while Wyoming has a Democratic one.

If we look at who was in charge of each state as of Demcember, 2007, when the recession began, we see no difference in unemployment rates between states with Republican governors and states with Democratic ones.

We can also look at the composition of each state’s legislature. States that had Democrats in charge of both chambers of their legislature as of December 2007 have somewhat higher unemployment rates than those with Republicans in charge. However, the difference is not statistically significant. States with mixed legislatures — where the higher and lower chambers are split between the parties, or where one or both chambers is subject to a tie or coalation government — have higher unemployment rates than either Democratic or Republican states, perhaps suggesting that the real enemy in tackling the unemployment crisis is not partisanship but gridlock.

Nor do states with Democratic governors have larger budget gaps, although states with Democratic legislatures do:

We’re going to have something of a natural experiment taking place over next 12-18 months; the states have adopted a variety of strategies in response to the economic crisis, and we’ll get to see which ones recover more quickly than others. But so far the evidence, contra Douthat, is rather mixed.

Nate Silver is the founder and editor in chief of FiveThirtyEight.