Poll Forecast: Slight GOP Edge in House for Midterms Could Be Overcome by People of Color and Progressive Women Candidates

Photo Source Ed Schipul | CC BY 2.0

Heading into Tuesday, the race to control the U.S. House of Representatives remains extremely tight according to CounterPunch’s election modeling based on a combination of polls and fundamentals. Our modeling accurately forecast in 2016 that Michigan and a handful of other toss-up states could put Trump at or beyond the 270 electoral votes necessary to become President. I have been #BlueWave skeptical since at least March when we published “Numbers Suggest Democrats Are Not Currently Set to Take Back the House of Representatives.” Republicans still hold a slim 220-215 lead for House control in our projection if no race is considered a toss-up. Including toss-ups in the model, likely more accurate, currently puts Democrats at 202 seats to 196 for Republicans with 37 toss-ups. The Generic Congressional Ballot average, usually a fairly accurate indicator of national popular vote totals, puts Democrats up by 7.3%. Our way of taking that metric, as always, involves averaging the latest poll from all firms in the field in the last ten days without weighting or adjustment. Other forecasters have long set 7% as the most likely vanishing point between the GOP holding the House and Democrats taking back control. A final seat count and Generic Ballot projection, taking into account small changes from the final day of polling, will be posted at the end of this Twitter thread around 10am on Tuesday.

District polling in the field in the last three weeks, and especially in the last few days, shows that four kinds of Democratic candidates, many of whom are seriously outperforming the model’s baseline expectations, could put Democrats over the top. Another kind of candidate, seriously pushed by and expected to do well by establishment Dems who control the Democratic Congressional Campaign Committee (DCCC), is not faring quite so well on average.

The four kinds of congressional candidates that should bolster Democratic hopes are people of color, candidates with excellent ground game, progressive women, and those who are running against a half dozen or so Republican candidates mired in one kind of scandal or another. To win Tuesday, Democrats in the closest 100 contests need to perform better than our model’s baseline expectations (based on fundamentals such as a district’s partisan lean, 2016 House vote, incumbency advantage, state and national generic ballot polling, and polling in similar districts) by an average of about 5.5%. Taken as a whole, Democrats running in these contests, if polling in the last three weeks in 71 of the 100 contests is reasonably accurate, appear to be doing 4.9% better than baseline. Seventeen Black, Latina or Latino, Arab, Asian, or Native American candidates in close races with recent polling are performing 6.7% better, on average, than the model’s expectations. Meanwhile, seventeen candidates with military, intelligence, or national security backgrounds are running just 3.47% ahead of expectations, well short of the five and half point average above baseline we expect is needed for Democrats to control the House.

Of course, there is overlap between some of these kinds of candidates—overlap that includes things like a progressive woman running a campaign with incredible ground game against very long odds. I’ve picked Jess King in PA-17 as an upset win over Republican Lloyd Smucker in a very Republican leaning district in part because of an army of volunteers that has knocked on around 200,000 doors while she’s run a pitch pipe perfect campaign deep in the heart of Mennonite country. Three candidates with national security backgrounds are people of color—Andy Kim NJ-3, Gina Ortiz Jones TX-32, and Gil Cisneros CA-39. Kim was a national security adviser in the Obama administration. He’s tried out a mix of moderate and somewhat progressive positions for his campaign while also running a ground game that has knocked on nearly a quarter million doors. Ortiz Jones, an Air Force captain in Iraq with subsequent intelligence experience, would be the first Filipina woman in Congress, but is running seriously behind in polling after being an early favorite of centrist Democrat partisans. Cisneros is running about even with our expectation and is in a very tight race against Republican Young Kim who would be the first Korean-American woman in Congress.

Randy Bryce, meanwhile, is a Latino candidate running on a wonderfully progressive domestic agenda in Wisconsin’s 1stdistrict where House Speaker Paul Ryan is retiring. Bryce would be expected to lose by about 7% according to baseline modeling that takes into account the district’s large R+5 partisan lean in Cook Political’s PVI ratings along with how big Ryan won in 2016. (An R+5 district represents an average of the district voting for President at a rate 5% greater than the U.S. as a whole over the last two cycles, meaning, generally, 5% less for Democrats and a 10% spread all other things being equal.) The most recent poll, from a partisan Democratic firm, has Bryce up by 1%. Wherever candidates reasonably fit into multiple categories, we have taken their numbers into account for each category.

Toss-ups are toss-ups for a reason. No forecaster in their right mind thinks they can accurately call all 435 individual congressional races with so many variables at such a level of precision. Starting with the third column in from the right “CALL” you can see our projection for each district beginning with the ones most favorable, according to the model, for Democrats (shaded in blue) and moving down through 100 races with toss-ups beginning at line 203 (shaded in gray) and super-toss-ups (shaded in dark gray and with less than a 1.5% lead for either side). Democrats need to reach the magic number of 218 to win back the House, though it’s unclear how much they would be able to accomplish or keep Trump from accomplishing with that slim of a majority.

A few races that should jump out, going down the list, for the narrative being told here are GA-6 where Black Lives Matter “movement mother” Lucy McBath has polled quite well in recent days, including a 2% lead in a New York Times/Siena College poll ending yesterday. Last year, Democrats hyped a bland, centrist candidate Jon Ossoff in a special election in GA-6 that they poured millions into only to have tight polling showing the race even turn into a four point loss on election day. As in 2016 for Hillary Clinton, Democrats struggled to get minority, especially young minority, voters to the polls for Ossoff. According to one accounting, the 18-29 year old early vote in Georgia has quadrupled with Stacey Abrams at the top of the ticket running for Governor. While McBath is running on a mostly moderate platform in this R+5 district, another candidate in a R+5 district is showing that one need not eschew left politics to remain competitive.

Leslie Cockburn, a long-time investigative journalist for such outlets as CBS and PBS and sister-in-law to CounterPunch co-founder Alexander Cockburn, is not only running on progressive domestic issues such as Medicare-for-All and opposition to new oil and gas pipelines in VA-5, she also quite clearly spell out a progressive vision for foreign policy including, among other items, equal treatment of Palestinians. She’s handled the inevitable charges of anti-semitism like the seasoned media professional she is. Cockburn showed a 1% lead in a recent NYT/Siena poll, and our model currently has her down just two and half points overall, well within striking distance. It doesn’t hurt that her opponent Dave Riggleman is now known nationwide for an embarrassment involving Big Foot erotica.

For my brief analyses of fifty of the closest races such as these, check out this Twitter thread here. (By the end of the day, I’ll try to have all of them up on my blog here for easier reading in one place.)

Most likely outcomes
A. Our primary forecast is for a tight election with one side or the other with 222 seats or less. The GOP’s negligible three seat lead in our model, without considering toss-ups, could well give way to late momentum for Democrats in individual seats such as Il-14 where Lauren Underwood, a young black woman who served in Obama’s Health and Human Services Department, came seemingly out of nowhere for a six point advantage in a NYT/Siena poll ending yesterday.

B. A moderate BlueWave giving Democrats 223-230 seats. To get to this range, Democrats will likely need outperform their current Generic Ballot expectations nationally and win in a bunch of places that have a heavy partisan lean toward Republicans.

C. Democrats 205-212 seats, with 223-230 for the GOP. Generic Ballot Polling over projected Democrats totals by three points or more in both 2006 and 2014 based on the RealClearPoliticsaverage of polling in those years. If this has happened again, expect a long, bad night for Democrats similar to November 8, 2016.

D. A Big Blue Wave of 231 or more seats for Democrats is possible if, as in 2010, Democrats perform better than their expectedRealClearPolitics Generic Ballot average by more than two percent to land them up near a double digit victory. Of course, 2010 was a wave year for Republicans as 2006 was for Democrats. It seems most likely that high enthusiasm for one party or another may infect polling numbers a bit to over-inflate how much they will win by rather than the likelihood of not thinking the leading party will do as well as it turns out to perform.

One particular factor that could heavily determine which route things will go is turnout for 18-34-year-olds. A pollster I spoke with who is working privately for Democrats in battleground House races this cycle suggests what FiveThirtyEight has pegged as the average/median figures (currently 37-38 seats) fits their numbers, but may be closer to the Democrat’s ceiling. “If people under 35 vote, it’s a very different picture” than if they do not, the pollster told me. If it goes as well as hoped, Democrats could go above 230 seats. However, “[i]t flips to 229-206 the other way [for Republicans] if under 35 come out in low numbers.” That would mean a Democratic pick-up of just 11 seats, very similar to the floor for Dems in our model.

Unfortunately, even with strong help from Big Foot in not just one (VA-5) but two (MN-8) critical contests, the big #BlueWave, said to be sweeping Democrats back into solid control of the House of Representatives Tuesday, may be more the stuff of forecaster fiction than foregone fact. Democrats in Washington seemed to learn nothing from Hillary Clinton’s loss and have, in key congressional districts, pushed as many CIA Democrats, prosecutors, centrist flunkies, anti-Labor lawyers, Bay-area venture capitalists, gun nuts, and Distinguished Professors of Civility as possible. Our projected outcome is almost a perfect 50-50 toss-up with thirty-seven races where one side or the other in our model leads by less than three and a half percent. If Democrats do manage to come out on top in the wee hours of Wednesday morning, they will likely be deeply indebted to candidates of color and the handful of progressive women who sneaked by watchful eyes in D.C. during primary season.