The House in perspective

While Paul Simon found 50 ways to leave your lover, I’m going to catalogue 15 ways to think about how many House seats Democrats will lose next week. History ordains Democratic losses — in all but four midterms since the Civil War, the party controlling the White House has lost House seats. The question, of course, is how big those losses will be.

(None of what follows constitutes my prediction — whenever I write something like this, some fool runs off saying Mellman predicts X or Y — so to be clear, I am not predicting anything.)

Simple averages provide some guidance, though they prove far from simple, as when you start, and what you count, make a difference. Over the last 17 midterms, the party controlling the White House lost an average of 28 seats, while over the last 16 it’s 24 seats, and during the last 15 midterms the average drops slightly, to 22. Start the clock with the midterm of 1870 and the average jumps to 35 seats; begin in 1934 and it’s 32 seats.

Gallup notes than since 1946 (i.e., the last 16 midterms), the average loss for presidents whose approval rating is above 50 percent is 14 seats. Alas, President Obama joins the group whose approval is under 50, where the average balloons to 36 seats.

Of course, these averages ignore the fact that the more seats a party won in the previous election, the more it has to lose. Converting seat losses to percentages of seats won in the prior election is a straightforward way to maneuver around this problem. The results — on average, presidential parties lose 15 percent of their seats — translate into about 39 seats this year.

Models are more complex, combining a variety of historical data to predict this cycle’s results by comparing current conditions to those in the past. While some of these models incorporate poll data like the generic vote, others focus exclusively on fundamentals like the economy, exposure (the number of seats the party in power is defending), presidential popularity, etc.

On average, models based on fundamentals forecast a loss of 35 seats. Those relying on polls are more pessimistic, expecting Democrats to relinquish about 51 seats — though, depending on the final Gallup generic results, some might predict losses of over 60 seats.

One final perspective, offered by distinguished political demographer Rhodes Cook, is not a forecast, but puts 2010 into the context of 1994 by examining the different terrain on which Democrats are battling this year. Cook divided House districts into those won by the Democratic presidential candidate in the two previous races (blue districts), those won by the GOP in the two previous presidentials (red districts) and those delivering inconsistent results (purple districts). In 1994, Democrats lost 6 percent of the blue, 27 percent of the purple and 54 percent of the red seats. If Democrats forfeit the same share of each category in ’10 as in ’94, the party would turn over about 46 seats.

Whew — a lot of data and too many computations. What’s it all mean?

First, significant losses were the baseline for this year. Forget the stimulus. Forget healthcare. Forget global warming and partisanship (bi- or otherwise). Normal midterm effects and Democratic exposure, let alone the economy, all but guaranteed substantial losses and a House in play. Anyone expecting small losses this year was out of touch with both historical precedent and economic reality.

Second, with so many indicators suggesting losses over 39 seats, preventing a Republican takeover of the House will be an extraordinary achievement.

Finally, the range of predictions is large enough that our uncertainty about the final outcome must be similarly expansive. We will actually need to let the voters decide.

Mellman is president of The Mellman Group and has worked for Democratic candidates and causes since 1982. Current clients include the majority leaders of both the House and Senate.

---