Our House forecasting model continues to show Republicans as having about a two-in-three chance — 67 percent — of taking over the House on Nov. 2, similar to the forecast in recent weeks.
The most likely number of Republican pickups is in the range of about 45 seats — although significantly larger or smaller gains remain possible. The model does not expect a clean sweep: Democrats are favorites in 4 seats currently held by Republicans. But Republicans are favorites in exactly 50 Democratic-held seats, according to the model, which would be enough to give them control of the House.
As in past weeks, the various indicators that the model uses to formulate its forecasts show varied patterns.
- The Democrats’ position in the generic ballot has improved slightly. But they remain about 6 points behind the Republicans among likely voters.
- Expert forecasters, like Cook Political and the Rothenberg Political Report, continue to move their characterizations of races slowly but steadily in the direction of Republicans.
- Local polling tends to be ambiguous. There were a handful of poor polls for Democrats this week — one, for instance, showed the outspoken incumbent Alan Grayson trailing in his race in Florida’s 8th congressional district. But in other districts, Democrats have released campaign polls showing themselves to be in a reasonably strong position, and Republican campaigns have not always countered them.
As we remarked last week, Democrats seem to be in a more vulnerable position in the Midwest than elsewhere in the country. Of the 50 Democratic-held seats that the model regards Republicans having at least even odds of gaining, 6 are in Pennsylvania, and 3 are in Ohio, and several seats in Michigan, Illinois and Iowa are also vulnerable.
While Republicans are more likely than not to take over the House, there are comparatively few individual seats in which they are overwhelming favorites: there are “only” 20 Democratic-held seats that the model assigns them at least a 80 percent chance of picking up, for example, and only 7 where they have at least a 90 percent chance.
But a much wider list of seats — possibly as many as 90 or even 100 — could conceivably fall into Republican hands. Unless there is either a shift in Democrats’ direction between now and Nov. 2, or the polling proves to underestimate them, the party will probably either lose the House or come perilously close to doing so. They have only about a 10 percent chance of holding onto 230 or more seats in the new Congress, according to the model, which might allow them to retain a reasonably functional working majority.
* * *
You may notice that we’re in the process of rolling out additional detail on the House forecasts. Each contest now has its own race page, for instance, which we have released in beta form — see for example the pages for Illinois’ 10th congressional district or Florida’s 24th.
In addition to charting the evolution of the forecast in each district over time, these pages also contain a list of polls, including those released by campaigns. Bear in mind as you browse these pages that local polling is but one of several inputs in our House model, and not always the most important one. In addition, the polls — particularly those released by campaigns — are subject to heavy adjustments of various kinds. We are working on various types of graphics that will convey a more comprehensive view of how our forecasts are calculated.