The 2010 midterm elections don’t look good for Democrats. On November 2, the Democrats will have 57 seats in the Senate, the Republicans 41 and independents two. I project that when the dust settles, the Democrats will hold 53 seats, the Republicans 44 and independents three.

Thirty-seven seats are contested this year; nineteen are held by Democrats and eighteen by Republicans. Based on mid-August polls and on longer-term voting trends, I expect Republicans to win three Democratic-held seats: Delaware, Indiana, and North Dakota. Six other Democratic-held seats are in play: Arkansas, California, Illinois, Nevada, Pennsylvania, and Washington. Arkansas, where Senator Blanche Lincoln trails badly in the polls, most likely will turn Republican. The same is true in Pennsylvania, where Representative Joe Sestak, who upset Arlen Specter in the Democratic primary, faces an uphill battle against Republican Pat Toomey. But this race has narrowed considerably, and much could change before the ballots are cast.

Republicans also have vulnerable seats: Florida, Kentucky, Missouri, New Hampshire, and Ohio. Kentucky and New Hampshire lean toward the GOP, but Democrats have the edge in Missouri and Ohio. Perhaps the oddest race of the year is in Florida, where Governor Charlie Crist lost the Republican primary and is running as an independent. I favor Crist to win the three-way.

Democrats surely will take a hit in the House as well. My prediction reflects the challenge of incumbency in a period of economic recession: using current forecasts for GDP growth in 2010, I predict that Democrats will lose 25-30 seats, most likely leaving them 224 seats to the Republicans’ 211. The forecasting model is somewhat naive, though the Democratic and Republican parties seem to agree with its results.

These predictions reflect the political tenor of their moment. But election season is a frenetic time. As November nears, check this site for more analysis and forecasts from the experts.

• • •

While the elections got in full swing this summer, the Senate was busy debating Elena Kagan’s Supreme Court confirmation. President Obama’s second nominee to the high court faced greater public opposition than his first, Sonia Sotomayor. In July 2009 we polled respondents on whether they agreed with a potential confirmation of Sotomayor. We ran an equivalent poll about Kagan in July 2010. Overall, support for Kagan was considerably lower, perhaps mirroring Obama’s falling approval ratings. But, curiously, Hispanics bucked the trend and threw their support to Kagan more strongly than they had Sotomayor. Support for confirmation of court nominees:

Election Update

(Sept. 27, 2010)

In late July, I put forth my forecasts for this piece. Forecasting is a risky business, and certainly foolhardy that far in advance of the election. There are many twists and turns on the election trail that affect individual races, even the national picture. In July, Mike Castle, a popular House member from Delaware and moderate Republican, looked like a sure thing to win Joe Biden’s Senate seat in Delaware. Charlie Crist, a popular governor and (another) moderate Republican turned independent, seemed to have the inside position in the race for Florida’s open Senate seat. My forecast was for a close election in which Democrats would emerge with a slight lead in the Senate and House. It was based mainly on party identification and economic performance. The Democratic Party maintained an edge in party identifiers. And the early summer economic news, although not rosy, seemed brighter than the reports that greeted us at the beginning of the fall.

This election, we are told, is all about jobs. Few of my fellow political scientists, however, use unemployment or GDP-growth to predict midterm congressional elections. The staples of most forecasting models for congressional elections are party identification and popularity of the President. These pull the electorate in opposite directions. Many of my colleagues who contributed to this contest made their forecasts at the beginning of September. Perhaps the key input in those forecasts is a survey question called the Generic Congressional Ballot: “If the elections for U.S. Congress were being held today, would you vote for the Republican Party’s candidate or the Democratic Party’s candidate in your district?” In late August, Republicans led by as many as seven percentage points. That gap narrowed considerably over the month of September.