Articles and Analysis


Using the Generic Ballot to Forecast the 2006 House and Senate Elections

Topics: 2006 , The 2006 Race

[Today's Guest Pollster's entry comes from Alan I. Abramowitz, the Alben W. Barkley Professor of Political Science at Emory University in Atlanta, Georgia. He is also a frequent contributer to the blog Donkey Rising.]

In order to predict the outcome of the 2006 House elections, I create a model incorporating both national political conditions and candidate behavior. Pre-election Gallup Poll data on the generic ballot and presidential approval are used to measure national political conditions and data on open seats and challenger quality are used to measure the behavior of congressional candidates. The model is tested with data on U.S. House elections between 1946 and 2004. A simpler model based only on national political conditions is tested with data on U.S. Senate elections from the same period. Based on the estimates for the models, I forecast the 2006 House and Senate election results.

The dependent variable in the House forecasting model is the change in the percentage of Republican seats in the House of Representatives. The model includes six independent variables. The percentage of Republican seats in the previous Congress is included to measure the level of exposure of Republicans compared with Democrats in each election-the larger the percentage of Republican seats in the previous Congress, the greater the potential for Republican losses. A variable for Republican vs. Democratic midterm elections is included to capture the effect of anti-presidential-party voting in midterm elections. Net presidential approval (approval - disapproval) in early September is included to measure public satisfaction with the performance of the incumbent president, and the difference between the Republican and Democratic percentage of the generic ballot in early September is included to measure the overall national political climate. The actions of congressional candidates are measured by two variables: the difference between the percentages of Republican and Democratic open seats and the difference between the percentages of Republican and Democratic quality challengers, defined in terms of elected office-holding experience.

The model does a very good job of explaining the outcomes of past House elections-all of the independent variables except the percentage of Republican seats in the previous Congress have statistically significant effects and the model explains 87% of the variation in House seat swings since World War II. Even after controlling for presidential approval and the actions of strategic politicians, the generic ballot variable has a substantial impact on the outcomes of House elections: a 10-point advantage in the generic ballot produces a swing of approximately nine seats in the House with all other independent variables held constant.


House Forecast
We can use the results in Table 1 to predict the outcome of the 2006 House elections. Based on a net approval rating for President Bush of -17, a Democratic advantage of 12 points in the generic ballot, and a Democratic advantage of 2% in open seats, and a Democratic advantage of 3% in challenger quality, the model predicts a Democratic gain of 29 seats in the House of Representatives.

Senate Seat Change Model
The dependent variable in the Senate model is the change in the number of Republican Senate seats. The independent variables are the number of Republican seats at stake in the election (a measure of exposure), a variable for Republican vs. Democratic midterm elections, net presidential approval in early September, and the difference between the Republican and Democratic percentage of the generic ballot in early September. Variables measuring candidate behavior are not included in the Senate model because data on challenger quality is not available for Senate elections and relative numbers of Republican and Democratic open seats had no impact on the outcomes of Senate elections when it was added to the model.


The results in Table 3 show that the Senate forecasting model is not as accurate as the House forecasting model, explaining only 65% of the variance in the outcomes of Senate elections since World War II. This is not surprising since the model does not include any variables measuring candidate behavior. Moreover, Senate seat swings are probably influenced more by chance because there are far fewer contests in each election and a larger percentage of these contests are competitive.

Despite the limitations of the Senate model, however, the results indicate that three of the five independent variables have significant effects. In the Senate model, in contrast to the House model, seat exposure is the single strongest predictor of outcomes. This is consistent with the results of previous models of Senate election outcomes such as Abramowitz and Segal (1986). According to the results in Table 2, for every additional seat that the Republican Party has to defend in a Senate election, it loses an additional 0.8 seats.

While the effects of the presidential approval variable are not quite significant at the .05 level, the generic ballot variable does have a statistically significant, and substantively important, impact on the outcomes of Senate elections despite the fact that the question asks about voting in House elections. The results in Table 2 indicate that an advantage of 10 points in the generic ballot produces a swing of about two seats in the Senate with all other independent variables held constant.

Senate Forecast
We can use the results in Table 2 to predict outcome of the 2006 Senate elections. Democrats need a gain of six seats to take control of the Senate. Based on a net approval rating for President Bush of -17 and a Democratic advantage of 12 points in the generic ballot, the model predicts a Democratic gain of 2.5 seats in the 2006 Senate elections. The main reason why the predicted Democratic gain is relatively small is that only 15 Republican seats are being contested this year.

Both national conditions and the behavior of candidates influence the outcomes of U.S. House elections. President Bush's low approval ratings and especially the large advantage that Democrats currently enjoy in the generic ballot suggest that Democrats are very likely to regain control of the House of Representatives in November. Democratic gains are also likely in the Senate but it will be difficult for Democrats to pick up the six seats that they need to take control of the upper chamber because only 15 of the 33 seats up for election in 2006 are currently held by Republicans.



Very interesting study. I think we all know that the Dems will gain more than 2.5 seats, but I think the model is really good. The reason for higher Dem gain in the Senate is based of not just the national environment but the weakness of GOP candidates in a number of races they should win (Montana)



The House forecast is pretty decent. Making a model using all of the available polling data on the house, it is possible to determine the mean number of expected seats that the Democrats should pick up either through Monte Carlo simulation or via a straight probabilistic analysis (sampling error should be normally distributed). While these two values should be the same, they end up being slightly different due to the fact that it is not possible to win a fraction of a seat in the house. The result of this analysis is that the Democrats should pick up 27 seats in the Monte Carlo simluation and 33 in the straight analysis. Thus, my analysis suggests that the dems will pick up 27 seats with a 95% confidence interval of 22-32 seats.

In the Senate, there is a 32% chance of Democratic control and 44% chance of a 50-50 tie (counting Sanders and Lieberman as Dems). The numbers in the Senate fluctuate a lot more due to the ever-changing polling data and the affect that a single race has on the overall composition of the Senate.


Craig VanGrasstek:

Have you tried tweaking the model to distinguish between first and second mid-terms? Making adjustments for those VPs who became president through death or resignation of their predecessor, I define the second mid-terms of the past few generations to be 1938 (FDR), 1950 (Truman), 1958 (Eisenhower), 1966 (LBJ), 1974 (Ford), 1986 (Reagan), and 1998 (Clinton). In these elections, the average seat losses for the president's party are 35 in the House and 6 in the Senate. Interestingly, those numbers are just at the high end of the forecasts that one commonly sees for 2006. I would be interested to see if such an adjustment in the dummies produced a different R-squared in your postdictions, and a different forecast for next week. It would also be interesting to know whether the other factors that you have already plugged in help to explain why this one factor does a much better job of predicting outcomes for the first five of these second mid-terms than it has for the two most recent ones.


Gary Kilbride:

I'd love to see Abramowitz' model backfitted to include Mark Mellman's mismatched seats factor. IMO that's the underplayed or ignored variable. In '96 I looked at '94 expecting a flurry of aberrational results. It was red districts voting red. This year I'm becoming increasingly optimistic of control, but red districts voting blue and knocking out entrenched incumbents en route doesn't make a heck of a lot of sense, not in the numbers being projected, by models or subjective forecasters. Not unless the close races break blue.

Let me retreat to 2004. Alan Abramowitz emphasized the advantage Bush had as an incumbent with his party in power only one term. I had posted that point repeatedly on progressive sites as a warning, but was rebuffed in favor of number of yards signs, and how the Redskins fared in their final game before the election. This year in a second term midterm the situational advantage is blatantly in favor of Democrats, but the edge was reversed in 2004. Looking to 2008, presidential elections with one party in power exactly two terms have produced the tightest popular vote finishes in modern American history -- '60, '68, '76 and '00. Only '88 varied, when voters rejected tank man and wanted a third Reagan term.

Here are the concluding paragraphs from an Abramowitz PDF, several months before the election in 2004. I remember reading this at the time and somberly realizing Abramowitz had probably isolated the correct weighting, that Bush's advantage as a first-term incumbent was being dragged down by the unusually polarized electorate, but probably not enough for a flip. I'd like to rerun '04 with a charismatic challenger.

www.forecastingprinciples.com/ Political/PDFs/Abramowitz.pdf

"George Bush�s net approval rating of -1 percent is well below the average of +11.7 percent for all incumbents running for reelection and far below the average of +19.4 percent for all first-term incumbents. In addition, the estimated 3.75 percent growth rate of the U.S. economy during the first half of 2004 is below the average of 4.5 percent for all presidential election years since World War II. The time-for-change model predicts a Bush win in 2004 primarily because he is a first-term incumbent.

According to the results presented above, a first term incumbent typically receives a boost of just over five percentage points compared with a candidate whose party has controlled the White House for two or more terms. Despite this advantage, however, George Bush has been running no better than even in the polls with his Democratic challenger, Massachusetts Senator John Kerry. In 25 national polls conducted during the month of July, Bush was favored by an average of 44.5 percent of likely voters compared with 46.3 percent for Kerry and 3.4 percent for independent candidate Ralph Nader.

It is entirely possible that George Bush will overcome his current deficit in the polls and go on to win the 2004 presidential election as the time-for-change model predicts. Given the stability of the race for the past five months, however, and the fact that Bush�s largest average lead in any month was just over one percentage point, it seems unlikely that he will receive anything close to 53.7 percent of the major party vote in November.

The poll results suggest that George Bush may not receive the full advantage that normally accrues to an incumbent president when his party has been in office for only one term. As many commentators have noted, the U.S. electorate is exceptionally polarized in 2004. George Bush enjoys overwhelming support from his co-partisans but he is not getting anywhere near the level of support that a first-term incumbent typically gets from independents and opposing partisans. In fact, the difference between President Bush�s approval rating among Republicans and Democrats is the largest ever recorded by the Gallup Poll."


Carl Leichter:

House gain 29 v Senate gain of 2.5

I suspect the reason why the House estimate seems reasonable, while the Senate estimate seems low might not have anything to do with the validity of either model. It might be due to the divergent nature of the data sets themselves.

There will be 435 House seats, but only 33 Senate seats up for contention in any given election. This means that there is about 13 times as much data available for developing a generalized House results model. I believe that is why the House results estimate seems to be more in line with our expectations. Since there is more data available for it, the House model is more accurate than the Senate model.


Alan Abramowitz:

Thanks for all of the comments. The Senate forecast is almost certainly too low. Senate elections are simply less predictable due to the small number of seats involved and the larger proportion of closely contested races. This year, more Republican seats than predicted are in jeopardy because of factors outside of the model such as the fact that a Republican incumbent in Pennsylvania is far more conservative than his state's electorate and a Republican incumbent in Montana is caught up in the Abramoff scandal. As a result, Republicans are likely to lost 4-6 seats in the Senate instead of 2-3.


Rick Matland:


One thing I'm curious about is the fact that your model (and some of the others) don't distinguish between 46-94 and later periods. It implicitly assumes the model is the same, but Charles' work, Gelman/King, and others have pretty clearly shown the vote/seat swing ratio is lower today than it was in the past (the cube rule is definitely dead). Largely because of redistricting acumen increasing we shouldn't expect the same swing, or should we? Shouldn't your model account for these changes?



John Caple:

I think what this model is saying is that everything being equal the Dems would pick up 29 seats.

However, we know that everything is not equal. Specifically, there are lower numbers of swing seats today than ever. Just based on that I think you say that 29 is probably an upper bound. The actual number is likely to be lower.

I think you also have to ask the question if GOP GOTV efforts and spending differences in local races can blunt the overall political climate.

If you believe that both the structural rigidity and GOP advantages are real issues you get an election that is very close at around the 15 seat change. If you believe they don't matter much at all you get a 30 seat change. If you are somewhere in the middle you get a change in the low 20s which is where most middle of the road predictions are.

The good news is that the model seems to fit the CW...


Post a comment

Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.