KEY POINTS FROM THIS ARTICLE
— Perceptions of the presidential race could have some impact down the ballot.
— Ticket-splitting is on the decline, but plenty of voters will vote for different parties for president and House, perhaps to the benefit of candidates from both parties.
— We are making 14 House rating changes, 10 in favor of Democrats and four in favor of Republicans. The changes don’t really impact our overall House assessment, which is that we are not expecting much net change in the makeup of the House.
Table 1: Crystal Ball House rating changes
Expectations and the House
With the conventions now in the rearview mirror, and incidents of violence in Kenosha, WI and Portland, OR dominating the news, political watchers are debating whether the presidential race is getting tighter. Compared to Joe Biden’s high points in late June and early July, the race does seem to be closer; compared to before the conventions to now, the story is still unclear, although a flood of new polling Wednesday was generally favorable for Biden.
Whatever prognosticators say — our own Electoral College ratings show Biden leading but just shy of 270 electoral votes, which is a little less bearish on Donald Trump than other forecasters — the public itself does not seem to see a clear favorite in the race.
This may have down-ballot consequences.
Two mid-August polls showed Biden leading Trump nationally by a comfortable margin, but respondents in those same polls were torn on who they actually thought would win. The Pew Research Center had Biden leading 53%-45%, but respondents, by a slim 50%-48% margin, thought Trump would win. Likewise, a CBS News/YouGov poll had Biden up 52%-42%, but when asked who they thought would win, respondents were torn, 41%-40% nominally in favor of Trump.
The USC Dornsife tracking poll, which attracted attention in 2016 for being more favorable to Trump than many other national polls over the course of the campaign, has started off this cycle by being a little more favorable to Biden than the national averages. Its most recent reading showed Biden up by 10 points. However, when respondents are asked about who they think others in their state will support, Biden and Trump are tied.
Betting markets also indicate more uncertainty than national polls. RealClearPolitics aggregates betting markets, and finds that oddsmakers have the race roughly even right now — this after Biden became a favorite in June and July while Trump was narrowly favored in April and May. This stands in contrast to 2016; Ohio State University’s Thomas Wood recently noted that Hillary Clinton was a much bigger favorite in betting markets in late August 2016 than Biden was in late August of this campaign.
The Pew and YouGov polls showed that slightly more Republicans are confident in Trump winning than Democrats are confident in Biden winning. Surely, some of this is a 2016 hangover after the Democrats’ Election Night nightmare. But some of it, also, must reflect legitimate uncertainty about the outcome overall and at least some lack of trust in Biden’s polling lead, as well as the reliability of the polls that undergird Biden’s lead (we feel it too, at least to some degree).
Time will tell if the public and the bettors do better or worse than the polls. But the reason we bring all this up in the context of a report on the U.S. House is because these expectations may have down-ballot repercussions.
We’ve previously noted in this space a study by respected political scientist Robert Erikson, who suggested that some high-information voters may be likelier to split tickets against the party of a presidential candidate they believe is strongly favored to win the election as a way to put a check on the person they believe is the likely winner. This sort of dynamic may have helped down-ballot Republicans in 2016, who may have benefited from (erroneous) projections of a Clinton victory, contributing to the ticket-splitting we saw in highly-educated, suburban districts where Republican House incumbents performed well even as Trump was significantly underperforming usual Republican presidential performance.
This time, Democrats may be better off if Biden is simultaneously leading, but the public and betting markets don’t see him as being favored. Republicans are defending some highly-educated suburban districts this time where Trump seems likely to do even worse than he did four years ago. How confident voters are in the presidential outcome could contribute to the House outcomes in these districts, particularly in these kinds of districts with high-information voters.
Back before the pandemic, and when it appeared that Bernie Sanders might be the Democratic nominee, I wrote about this dynamic from the perspective of concerned Democratic House incumbents. You could now apply to same argument to Trump and Republican House incumbents — and even to Republican Senate incumbents like Cory Gardner in Colorado or Susan Collins in Maine, who will need crossover votes to win. Arguing that they could be a check on the Democrats’ worst impulses could end up being a closing pitch for both embattled senators as well as some swing district Republicans, but the argument may have less juice if voters aren’t confident about the presidential outcome.
Meanwhile, Democratic incumbents in Trump-won districts may have an easier time separating themselves from the national party and generating crossover support if voters in their districts are not strongly assuming Biden will win the presidency. That leads us to this week’s rating changes.
House rating changes
The power of incumbency is sometimes dismissed in this nationalized era, and it is true that both parties are less likely to win districts that vote strongly for the other party’s presidential candidate than they used to. But there is still ticket-splitting, and sometimes a dramatic amount of it.
You don’t have to go back very far to find it.
In the last presidential election, we saw some major shifts at the congressional district level: 200 of the 435 districts saw at least a five-point change in Democratic and/or Republican presidential performance. There were 37 districts won by House Republicans in 2016 where Trump ran at least five points behind Mitt Romney’s 2012 showing and/or Hillary Clinton ran at least five points ahead of Barack Obama’s 2012 showing. On the flip side, there were 32 districts won by House Democrats where Clinton ran at least five points behind Obama and/or Trump ran at least five points ahead of Romney.
The House Republicans who won these 37 districts ran, on average, 15 points ahead of Trump. Meanwhile, the House Democrats who won their 32 districts ran 22 points on average ahead of Clinton. So change at the presidential level is survivable, particularly by incumbents. And there still will be districts that vote for one party for president and another for House.
A couple of Trump-district Democrats whose 2018 victories were at least somewhat surprising are Reps. Jared Golden (D, ME-2) and Joe Cunningham (D, SC-1). Golden’s district flipped from voting for Obama to backing Trump by about 10 points, while Cunningham’s district voted for Trump by 14 points, but that was down a bit from Romney’s 2012 showing.
We are moving both from Toss-up to Leans Democratic.
Cunningham faces a well-regarded challenger, state Rep. Nancy Mace (R), but Cunningham is also (at least to our eyes) cutting through the political clutter by running lighthearted, positive ads. More to the point, and based on the numbers we’ve seen, Trump seems likely to run significantly behind his 2016 showing in SC-1, and Cunningham probably will run ahead of Biden in the district as a well-funded incumbent. That’s enough to give him a slight edge, at least for now. It’s sometimes tricky extrapolating from special legislative elections, but results from a district in the area last month suggest that some Democratic trends may be taking hold in the Charleston area.
Golden, meanwhile, faces an underfunded though credible challenger, former state Rep. Dale Crafts (R). While ME-2’s single electoral vote seems likelier than not to remain in Trump’s possession — Maine and Nebraska award electoral votes by congressional district — the president may be hard-pressed to replicate his 10-point districtwide victory. Any little bit closer ME-2 gets will benefit Golden. As a side note, Maine continues to use ranked-choice voting for congressional races, and Golden benefited from the system in 2018 — he finished narrowly behind then-Rep. Bruce Poliquin (R), but took the seat based on second-place votes. However, only Golden and Crafts are on the ballot, so RCV shouldn’t play a role in the outcome.
While we’re moving a couple of first-term Democratic incumbents from Toss-up to Leans Democratic, we’re moving one the other way: Rep. Debbie Mucarsel-Powell (D, FL-26). At first blush, this move may seem like a head-scratcher: Hillary Clinton won this South Florida district by double digits, and Joe Biden probably will too. But that glosses over the unique local politics in the area, where Republicans are much stronger down the ballot, and Mucarsel-Powell faces a strong Republican challenger, Miami-Dade County Mayor Carlos Gimenez (R). Republicans released an internal poll showing Gimenez up five points at the end of July, and we have heard some concerns from Democrats about the race. Gimenez may be getting a boost not just from down-ballot Republican voters, but also from playing an executive role during the coronavirus crisis. Democrats will work to tie Gimenez to Trump, although there are some signs that Trump may be doing marginally better with Hispanics than four years ago, and voters of Cuban ancestry in South Florida have traditionally been very open to voting Republican.
We also are moving Rep. Mike Garcia (R, CA-25), the special election winner from earlier this year, from Toss-up to Leans Republican. He definitely benefited from a very GOP-leaning electorate in the May special election, but he also may generate some crossover support as Joe Biden remains likely to carry the district, which voted for Hillary Clinton by seven points in 2016. We also thought state Assemblywoman Christy Smith (D) ran a poor race in the special election, and we’ll have to see if she can improve. Republicans had Garcia up 48-41 in internal polling as of late July; Politico reported that Smith’s campaign indicated its internal polling also had Garcia ahead, but by less. That Democratic polling also indicated Biden was up 48%-43% in the district, similar to Clinton’s 2016 margin. This is another district with a significant Hispanic population, and it is more working-class than some other districts Democrats flipped in 2018, two factors that suggest it might not be markedly more Democratic in 2020 than it was in 2016.
If Garcia ends up winning twice this year in what is a hard district for Republicans, it would remind us of a pair of races from 2010, when Mark Critz (D) somewhat surprisingly held a Republican-trending Western Pennsylvania district against the same Republican opponent in both a special election and a general election despite 2010 being a bad Democratic year (Critz lost two years later after his district was dismantled in redistricting). Voters may also figure that because they already voted for Garcia earlier this year, they might as well just stick with him. Critz did do better in his 2010 special election than the general, winning by eight and then two points, respectively. Garcia seems very unlikely to repeat his 10-point victory with a presidential electorate, but we’re giving him a small edge as he tries to hang on and win a full term.
Moving CA-25 to Leans Republican also helps to balance out our Leans Democratic ratings in two other California House seats, those held by Reps. T.J. Cox (D, CA-21) and Harley Rouda (D, CA-48), that arguably could be Toss-ups themselves. If Republicans make up any ground in California this November as compared to two years ago, it’ll be one of these three districts in all likelihood where that happens.
While Garcia moves out of the Toss-up column, Reps. Steve Chabot (R, OH-1) and Scott Perry (R, PA-10) move in. Both face strong Democratic challengers in districts that may trend Democratic at the presidential level this year. If Biden has any chance of winning Ohio, he has to make inroads in this Cincinnati-based district, which voted for Trump by about seven points in 2016. Health care executive Kate Schroder (D) is running a spirited campaign against Chabot, who benefited from some missteps by his challenger in a competitive race last cycle. Biden also will hope to make gains in the Harrisburg/York-centered PA-10, where Rep. Scott Perry (R) faces state Auditor General Eugene DePasquale (D). Trump won PA-10 by about nine points, but Perry only narrowly hung on in the redrawn district in 2018 against a less heralded challenger than DePasquale.
Speaking of presidential changes, we may see another shift toward the Democrats in two open seats in once-strongly Republican suburban turf. Two strong candidates to flip from Trump 2016 to Biden 2020 are two highly-educated and diversifying open seats, GA-7 in the Atlanta suburbs and TX-24 in the Dallas-Ft. Worth Metroplex. We are moving both districts from Toss-up to Leans Democratic at the House level, based primarily on the belief that the strong Democratic trend we are expecting in both districts will be enough to carry Democratic nominees Carolyn Bourdeaux (GA-7) and Candace Valenzuela (TX-24) over the finish line. If Republicans had strong incumbents in these seats, we would feel differently, but both are open.
Additionally, we are moving three Safe Republican districts to Likely Republican. Two are in California: Reps. Tom McClintock (R, CA-4) and the vacant seat formerly held by Duncan Hunter (R, CA-50) that former Rep. Darrell Issa (R, CA-49) is trying to hold for the Republicans; the other is the vacant seat formerly held by White House Chief of Staff Mark Meadows (R, NC-11). While Republicans should be OK in all three of these districts, polling has indicated competitive races and Democrats have credible challengers in all three. The one that perhaps stands out is NC-11 in western North Carolina: Madison Cawthorn won a surprising primary rout over a candidate who is close to Meadows, and the 25-year-old has the potential to be a future star (he was disabled below the waist after an auto accident and was featured at the Republican National Convention). However, he’s also gotten into hot water through the revelation of a social media post in which he may have been a little too enthusiastic about visiting Adolf Hitler’s vacation home in Germany, as well as some accusations that he behaved aggressively toward women during his teen years. All three of these districts lean considerably right of center, which ultimately should protect the GOP candidates in these districts.
We pay attention to the top-two primary results in Washington state as something of a preview for the fall, and following the finalization of the vote tallies after the Aug. 4 primary, we wanted to take a closer look. Just like in California, all candidates from all parties run together on the Washington primary ballot, with the top-two vote getters advancing to the November general election.
Two of the state’s 10 congressional districts appear to be at least somewhat competitive, and the results in each were encouraging for Republicans.
In Southwest Washington’s WA-3, Rep. Jamie Herrera Beutler (R) got about 57% of the total two-party primary vote, with Democrats combining for 43%. Throughout the past decade (2012-2018), the two-party voteshare from the primary in this district either stayed the same or got a little more Republican in November, so Beutler seems to have a fair amount of breathing room in her rematch with Carolyn Long (D). We are moving WA-3 from Leans Republican to Likely Republican.
In neighboring WA-8, Republicans actually outvoted Democrats by a narrow 51%-49% in the two-party vote. One of the positive signals for now-Rep. Kim Schrier (D, WA-8) in the 2018 primary was that she emerged from a crowded primary in which the two-party vote was 51.7%-48.3% Democratic. At the time, this made us think Schrier would win, and she did, 52.4%-47.6% against Dino Rossi (R), a frequent hard-luck losing candidate in Washington politics. In WA-8, the two-party vote has been a decent proxy for the fall, although the district was not really competitive from 2012-2016 as popular Rep. Dave Reichert (R) did not face major challengers. The Republican nominee this time, veteran Jesse Jensen (R), does not seem like a top-tier challenger on paper, and Republican outside groups are not prioritizing the district. We wonder if there may be some hidden Republican opportunity here, though.
One potentially positive sign for Schrier: Turnout in the district was extremely high in the primary, about 245,000 votes. However, there were roughly 310,000 votes cast in the 2018 general election, and turnout in a presidential year should be even higher than that. The district consists of all or part of five counties: the majority of the votes come from the King County portion (this is the county that includes Seattle), and this was the only part of the district Schrier won in 2018. The remaining votes come from the other four counties. Combined, primary turnout in the other four counties was at 85% of their 2018 general election total; King County turnout in WA-8 was only 75% of its 2018 total. So perhaps turnout has more room to grow in King County, which could help Schrier. Still, we’re moving the district from Likely Democratic to Leans Democratic as a precaution. WA-8 is a presidential swing seat that voted for Obama by two in 2012 and Clinton by three in 2016 — unlike some of the other districts mentioned above, it does not profile as a district that is moving sharply one direction or the other.
Finally, and moving up the Washington coast to the Last Frontier, Alaska, Rep. Don Young (R, AK-AL) appears to be in an increasingly competitive rematch with Alyse Galvin, the nominal independent/Democrat who held him to a seven-point win in 2018. We’re moving that race from Likely Republican to Leans Republican. Young is the longest-serving member of the House and turned 87 earlier this year.
The big picture
Overall, we now have 232 districts at least leaning to Democrats, 192 districts at least leaning to Republicans, and 11 Toss-ups. If we split the Toss-ups roughly down the middle (6-5 Republican), we’d be looking at a 237-198 Democratic-controlled House, or a two-seat gain from the 235-200 Democratic House elected in 2018.
Our general feeling the whole cycle has been to not expect much net change in the House overall, although the Democrats appear a bit better positioned to net seats than Republicans do at this point.