When people talk about the midterm elections, they often say unemployment will be the most important factor. An unemployment rate at 10 percent on Election Day 2010, the argument goes, will ensure heavy Democratic losses in the House and Senate. For sure, jobs and the economy are likely to dominate the coming year. But is the unemployment rate the number to watch as the election unfolds?
Maybe not. Glen Bolger and Jim Hobart of Public Opinion Strategies have a fascinating post in which they correlate presidential job approval with losses in midterm elections. Here's what they found:
POS, with the help of data from the Gallup Organization and information from our friends at National Journal and The Cook Political Report, took a look at how presidential approval effected mid-term House losses for the President’s party in every mid-term election since 1962. The results were staggering. If the President’s approval rating was 60% or higher, the President’s party picked up an average of 1 seat. If the approval rating was between 50 and 59%, the average loss was 12 seats. Finally, if the President’s approval rating was below 50%, the average loss was 41 seats (one seat more than the 40 seats GOPers need to win back control of the House).
I went back and looked at data for midterm election years to see how the unemployment rate corresponded with incumbent losses. When Democrats lost 47 seats in 1966, unemployment was at 3.6 percent, but LBJ's approval rating was below 50.
Or consider the 1994 and 2002 elections. In 1994, when Democrats lost 52 seats, unemployment was at 5.6 percent. That's less than the rate in 2002 (5.9 percent), when Republicans picked up eight seats. In retrospect, the more reliable indicator in those years was job approval: Clinton's stood at 46 percent on Election Day 1994, whereas Bush's stood at 63 percent in 2002.
In 2006, unemployment was actually lower still, at 4.5 percent. But by that time Bush was incredibly unpopular--his job approval had fallen to a dismal 39 percent. And the Democrats took control of Congress.
A lot of commentators focus on the 1982 midterms, where Reagan's approval was low and unemployment more than 10 percent. When viewed in the broader historical context, however, it seems that Reagan's 43 percent job approval was a better predictor of GOP losses.
The employment picture will likely improve this year. But Obama's job approval may not. And the latter number will be far more important to the future of the president's agenda.