What the Political Dorks Say

From Sabato's Crystal Ball

Alan I. Abramowitz

Guest Columnist

Could Democrats be heading toward an electoral disaster comparable to the 1994 midterm election in which they lost 54 House seats and 8 Senate seats, turning control of both chambers over to Republicans for the remainder of Bill Clinton's presidency? Nobody is predicting such a dramatic turnaround in party fortunes just yet. But while a Senate majority appears to be out of reach, some GOP strategists now see a chance for their party to regain control of the House of Representatives in next year's midterm election. And they're not alone. In a recent column, Charlie Cook of The National Journal, one of the nation's well-known and respected political analysts, warned that President Obama's sinking poll numbers along with growing resistance among voters to the President's policies and a painfully slow economic recovery could lead to massive Democratic losses next year. The fact that Democrats are struggling to pass a major overhaul of the nation's health care system just as they were in 1994 is also contributing to the feeling of "deja vu all over again" among many party leaders and supporters.

But while there are some important similarities between the current political situation and the circumstances that preceded the 1994 Republican victory, there are important differences between the makeup of the American electorate now and the makeup of the American electorate then, differences that make a repeat of the 1994 outcome highly unlikely.

The most important difference is that nonwhites make up about twice as large a share of the electorate now. Figure 1 displays the trend in the racial composition of the electorate in presidential and midterm elections between 1992 and 2008 based on data from national exit polls. Two patterns are evident in this graph. First, whites generally make up a larger share of the electorate in midterm elections than in presidential elections--the presence of a presidential race appears to be a more important motivation for voting among African-Americans, Hispanics, and other nonwhites than among whites. This means that the nonwhite share of the electorate in 2010 is likely to be lower than the all-time record of 26 percent that was set in 2008. Second, however, the data show a clear upward trend in the nonwhite share of the electorate in both types of elections. This means that the nonwhite share of the electorate is almost certain to be higher in 2010 than it was in 2006.

Needless to say, the racial transformation of the American electorate has important implications for the prospects of the two major parties. The weakness of the Republican Party among nonwhite voters is a much bigger problem for the GOP today than it was back in 1994. In that year, 86 percent of the voters were white while only 9 percent were African-American and only 5 percent were Hispanic or members of other racial minority groups. But in 2006, the most recent midterm election, only 79 percent of voters were white while 10 percent were African-American and 11 percent were Hispanic or members of other racial minority groups.

Based on the average rate of change in the racial composition of the electorate over the past two decades, by 2010 we can predict that no more than 76 percent of voters will be white while at least 11 percent will be African-American and at least 13 percent will be either Hispanic or members of other racial minority groups.

The Republican Party is even weaker among African-American, Hispanic, and other nonwhite voters today than it was in 1994. In the 2008 House elections, Democratic candidates won 94 percent of the vote among African-Americans, 70 percent of the vote among Hispanics, and 73 percent of the vote among other nonwhites according to national exit poll data. And recent polling data indicates that Republican support among all of these groups remains extremely low.

Based on the 2008 results and the projected racial make-up of the 2010 electorate, Republican candidates would have to win almost 60 percent of the white vote in order to win 50 percent of the overall national popular vote in 2010. That would be even more than the 58 percent of the white vote that Republican candidates received in 1994 and much more than the 54 percent of the white vote that Republican candidates received in the 2008 House elections.

Of course the results of House elections are not directly determined by the national popular vote--they are determined by the results of 435 individual contests of which only a small fraction are actually competitive. Nevertheless, there is a very strong relationship between the national popular vote and the outcomes of House elections. In 1994, when Republicans gained 54 seats to take control of the House, GOP candidates received about 53 percent of the national popular vote; in 2006, when Democrats gained 30 seats to regain control of the House, Democratic candidates also received about 53 percent of the national popular vote. For the 32 elections between 1944 and 2006, the correlation (Pearson's r) between the Republican share of the national popular vote for the House of Representatives and the Republican share of House seats is.93. This means that the Republican share of the national popular vote explains over 86 percent of the variation in the Republican share of House seats during this time period.

In all likelihood, in order to regain control of the House of Representatives in 2010, Republican candidates would need to win at least 50 percent of the national popular vote. Given the likely racial composition of the electorate in 2010, however, that is going to be extremely difficult.

Republicans are likely to make at least modest gains in the House of Representatives next year. That is the norm in midterm elections. Since World War II, the president's party has lost an average of 23 House seats and 2-3 Senate seats in midterm elections. The size of this loss varies considerably, however, based on such factors as the president's popularity and the numbers of House and Senate seats that each party holds going into the election--the more seats a party holds, the more it is likely to lose. Right now, those factors are pointing to a Democratic loss of about 20 seats in the House and 1 or 2 seats in the Senate. But the size of the Democrats' losses will ultimately depend on how voters evaluate the performance of the president and his party next fall.

What we can say with considerable confidence at this early date is that the steady growth of the nonwhite electorate and the continuing weakness of the Republican Party among nonwhite voters make a repeat of the Democrats' 1994 electoral debacle highly unlikely.