Curtis Gans has a major article up at the Washington Monthly (for which, I did a little editing) that predicts that the Democrats could defy conventional wisdom and do much better than expected in the midterm elections. His argument is similar to one I have been mulling over for several months, and I have even touched on some of his themes in previous posts.
I think the most important thing for analysts to remember is that the sample size for second-term midterm elections in the postwar era is so small that we can’t safely generalize. It’s not just that the sample size is small, though, it’s that there have been highly specific influences on each of the examples. The Republicans under Dwight D. Eisenhower had to deal with a deep recession in 1958. The Vietnam War had a huge influence on the 1966 elections. WaterGate was the main theme of the 1974 elections. Iran-Contra dominated the 1986 midterms (my memory was faulty here because the story broke after the midterms). A great economy offset the impact of l’affaire Lewinsky in 1998. And 2006 was a referendum on the war in Iraq, as well as a reaction to the federal response to Hurricane Katrina, the Terri Schiavo controversy, the Abramoff scandal, and other pent up frustrations with the Bush administration.
When we look forward to November, we don’t see a recession or a country deeply divided over a war that is going badly or a major scandal. The economy could be a lot better, but the stock market is at a near high. The president’s poll numbers could be stronger, but he isn’t in anything like the situation faced by Nixon, Reagan or Clinton.
So, the beginning point for analysts should be to consider how prior midterms might have gone if the administrations had not been mired in controversy. Without Watergate, would the Democrats have done so well in 1974? Without Iran-Contra, would the Democrats have retaken the Senate in 1986? Without the Lewinsky scandal, would the Democrats have retaken the House in 1998?
Most analysts properly focus on the tendency of Democrats to show up in much higher numbers in general elections than in midterms, but 2006 proved that this doesn’t necessarily mean that the Democrats will lose as a result. In 1998, we learned that the public may not reward the Republicans if their opposition becomes pathological.
So, I don’t think we can do very well in predicting what will happen in November by looking at the limited sample of previous postwar midterm elections. We know that we have a challenge in getting out the vote. Beyond that, we don’t know much.
Curtis Gans is correct that we cause to be hopeful.
