INTRODUCTION

For the entire 2022 cycle, mainstream pundits expected Republicans to comfortably flip the House of Representatives. Looking back, it’s hard to blame them for their optimism. The GOP had already enjoyed surprisingly-good House performance in 2020, netting 14 districts and reaching 213 seats – just 5 short of a majority. Conventional wisdom defining out-party momentum in midterms also added to the Republican case. Historically, the average net House gain for the party not occupying the White House is 28. That’s not all. The GOP had enjoyed a successful 2021 election cycle, flipping Virginia’s governorship, and amassed a modest generic ballot lead going into the early summer.

Then the conservative Supreme Court issued the Dobbs decision, removing constitutional protections for abortion rights in a country that is largely pro-choice. The repercussions of the holding helped Democrats take a lead in generic ballot polling in the early fall. Ensuing special election overperformances in seats right of the nation, including surprise victories by Pat Ryan in New York’s 19th and Mary Peltola in Alaska, diminished the “red wave” argument. Many analysts, including ourselves at Split Ticket, began to consider the possibility of a “neutral” year. Other fundamental indicators, like Washington’s historically-accurate primary results, also pointed to a weaker national environment for Republicans.

Going into October, the special election buzz started to wear off and Republicans eventually gained back a modest advantage in generic ballot polling per Split Ticket. Biden’s approval remained low until election day – it’s still just 41.5% – and most assumed that economic issues would ultimately trump social concerns like abortion with the electorate. In other words, historic precedent and partisan polling (on both sides) suggested a shift back toward the GOP in closing weeks. If most of the key districts broke their way, the assumption went, Republicans would have a good night.

We bought into these suppositions to a certain extent. Our priors pointed to an R+2 environment consistent with a good night for Republicans, but never a wave. In the House, the GOP was projected to amass a comfortable 234 to 201 majority. Election night returns vindicated post-Dobbs fundamentals, suggesting that overreliance on fickle information like party insiders and unclear internal polling had caused a sort of confirmation bias among forecasters. According to the latest New York Times House projections, Republicans have 211 seats to the Democrats’ 194. While the path is tenuous, there are enough uncalled districts for Democrats to hold the house, a feat that conventional wisdom precluded just days ago.

Because this article spends a decent amount of time looking at where our House ratings went wrong, it’s important to recognize that had we based our expectations on Split Ticket’s in-house generic ballot tracker, brilliantly designed by partner Lakshya Jain, our reading of the national picture would have been more accurate. Jain’s agglomeration accurately predicted a neutral environment by accounting for partisan polling that “flooded the zone” for Republicans. We won’t know the final House popular vote number until California finishes counting — the GOP is currently winning it even when accounting for uncontested seats — but odds are Split Ticket’s final D+0.2 average will end up being close to the mark.

WHAT DID WE MISS?

Because ≈ 30 races technically remain uncalled, there’s a decent possibility that our total number of misses will increase. That said, some contentious Pacific Coast districts like Washington-8, Oregon-5, California-47, and California-49 appear to be turning out like we expected, so the current figure of 18 will probably be close to the final error. One should also note that a call is only considered wrong if the outcome contradicts the prediction (i.e., OH-09). In other words, over or underestimating a winning candidate’s performance doesn’t count as a complete miss.

Looking at the above map of confirmed errors, one thing stands out above all: Democratic candidates were underestimated across the board, but particularly in races rated TOSSUP prior to our final update. This makes sense considering we overshot Republicans in the House, albeit by slightly less than some other forecasts. In New York, which joined Florida in having a good night for Republicans, our predictions held up well and actually missed three GOP pickups. Two came in seats rated LEANS D before the vote.

In New England, which hasn’t elected a Republican to the House since Bruce Poliquin in 2016, the GOP will most likely come away with no seats once again. Democrat Chris Pappas (NH-01) exceeded his 2020 reelection margin despite last-minute rumblings of a close race against Karoline Leavitt. His moderate colleague Jared Golden (ME-02) is poised for a comfortable victory in an upcoming ranked-choice instant runoff. 

The GOP performed better in Rhode Island-2 and Connecticut-5, both comfortable Biden-won seats prone to down ballot Republican ticket splitting, but ultimately fell short in both despite our LEANS R ratings. If there’s any consolation in those two misses, it’s that Republican George Logan came closer to beating Jahana Hayes in CT-05 than Allan Fung did to defeating Seth Magaziner in RI-02, a delta we had suggested.

In Ohio-9 and Indiana-1, we placed too much stock in polarization and the favorable trends of the Trump-era among white working class voters. These assumptions, in conjunction with expectations of a Republican-leaning national environment, led us to overestimate GOP candidates JR Majewski (OH-09) and Jennifer-Ruth Green (IN-01) while underestimating the sticking power of Democratic incumbents Marcy Kaptur and Frank Mrvan.

Texas-34 was called incorrectly because Republican numbers suggested Trump’s 2020 regional Hispanic inroads would grow rapidly in 2022. There was a rightward shift from the 34’s Biden +15.5 partisanship this year, but it wasn’t strong enough for Republican Rep. Mayra Flores to overcome her Democratic colleague Rep. Vicente Gonzalez. The results, along with Flores’s modest June special election victory in the old seat, indicate that the Rio Grande Valley’s down ballot Democratic strength won’t evaporate any time soon.

Two misses came in the suburban seats of Pennsylvania-7 and Michigan-7, where the salience of social issues, coupled with strong statewide Democrats, helped bring vulnerable incumbents Susan Wild and Elissa Slotkin over the finish line. This dynamic also helped Greg Landsman unseat long-time Republican Rep. Steve Chabot in Ohio-1 and allowed Democrats to win open Biden seats like Pennsylvania-17, Ohio-13, and North Carolina-13.

Three misses came out west, including two in districts that had been rated LEANS R. Nevada’s 3rd was the least surprising, with Democrat Susie Lee on track to join her colleagues Dina Titus (NV-01) and Steven Horsford (NV-04) in reelection. In New Mexico’s 2nd and Colorado’s 8th, also marginal Biden seats, Democrats Gabe Vasquez and Yadira Caraveo more or less met all of the necessary prerequisites for victory that we laid out in our district rundowns.

So far, the only Republican victories that we called incorrectly occurred in New York, where Congressman Lee Zeldin took an impressive 47% of the vote against Democratic Governor Kathy Hochul. His comfortable win in Nassau County helped Republicans George Santos (NY-03) and Anthony D’Esposito (NY-04) post historic wins in open seats. The GOP last won the 4th in 1994. Further upstate, Republican assemblyman Mike Lawler outran his seat’s partisan lean to defeat Democrat Sean Patrick Maloney (NY-17), becoming the first DCCC Chairman to lose reelection since James Corman in 1980.

OVERARCHING OBSERVATIONS

Before discussing the state of the House as a whole, we have two observations regarding this year’s elections. Both will be touched on in more detail in future House articles. The first deals with reversion, or rather the argument that long-time trends do not unfold linearly. The second makes clear that candidate quality does still matter and that there is a relationship between moderation and overperformance.  

Temporary reversion in specific congressional districts stems from the concept of down ballot lag, a frequent subject of discussion at Split Ticket. The term states that political realignments do advance steadily in the long-run but often take two steps back toward ancestral partisan leanings, which are usually stronger down ballot, before moving forward. Large gaps in candidate quality in specific districts and fluctuations in the national environment tend to be the primary causes of this sort of backtreading — manifested also as crossover voting — in federal elections.

This year’s races in Ohio-9 and Texas-28 fit into this category quite well. In the first seat, a Republican-trending Trump +3 district that was Obama +17 in 2012, veteran Democrat Marcy Kaptur blew out the controversial JR Majewski by 13 points, outrunning Biden by double-digits in each of the 9th’s county portions. Kaptur joins Matt Cartwright (PA-08) and probably Jared Golden (ME-02) as Trump seat Democrats, a group that could grow.

Texas-28, like a Hispanic-majority Ohio-9, was also a Republican trending seat, shifting from Clinton +31 to a mere Biden +7 between 2016 and 2020. Like Kaptur, conservative Democratic incumbent Henry Cuellar mobilized decades of experience to outclass his opponent, culminating in a double-digit victory. Cuellar’s institutional presence was valuable enough to cause an average leftward district swing of 6.3 points.

The Last Politiquero ties perfectly into the second point, which revolves mostly around candidate quality, or in many cases the lack thereof. Split Ticket research by Lakshya Jain notes that nominees can pay significant penalties for extremism. Republican Rep. Lauren Boebert, from Colorado’s 3rd, fits easily into this category. She is currently on track to narrowly defeat Democrat Adam Frisch in a seat that a generic Republican would have won easily. Her race was rated LIKELY R going into the election, but no one foresaw the drama that has since unfolded there.

Uncalled Races & Democratic House Chances – Possible, But Stretched Thin

As of this writing, both parties have viable paths to controlling the House majority, but we should preface this section by pointing out, regardless of who wins, how impressive it is that the Democrats are even in contention in the first place. After all, they were written off for most of the cycle by the majority of forecasters, including ourselves.

Let’s start with the improbable, yet possible outcome: Democrats maintain House control with a 218 to 217 majority. On the map below, each of the dozen or so truly-uncalled races has been allocated using our best guesswork. At the end of the day, there are enough outstanding ballots to change the outcomes in any of the mentioned seats. It’s crucial to remember, then, that these are possibilities, not projections.

With Colorado-3 likely in the Republican camp and districts like Oregon-5 and California-27 moving in a similar direction, the Democrats need to win the following seats: California-13, California-22, California-41, Washington-3, Arizona-6, and Arizona-1. Keep in mind that OR-05 and CA-27 are not done deals either and could be substituted in for other seats on this list, especially AZ-06, if late mail-ins offer any surprising changes. This math also assumes Democrats hold competitive districts they are defending, like California-47 and -49.

From an upset standpoint, Washington-3 currently seems like a better option for Democrats than Colorado-3. Controversial Republican Joe Kent finished ahead of incumbent Congresswoman Jaime Herrera Beutler in this year’s jungle primary. The moderate Republican had voted for Trump’s impeachment. Since then, returns have shown Democrat Marie Gluesenkamp-Perez leading Kent narrowly. While the outcome is still uncertain, the Democrat’s chance of victory looks sizable considering outstanding ballots in Clark County. Predictably, like Boebert and other extreme candidates, Kent is underperforming Trump’s 2020 baseline.

The more likely outcome at present, from a statistical standpoint, is a slim Republican majority — probably 220 seats to the Democrats’ 215. That would technically constitute a chamber pickup for the GOP, but it wouldn’t be viewed as an overwhelming feat by any means considering the high expectations that were set for Republicans this year. A GOP edge that slim could generate control issues for Kevin McCarthy, whose fate as leader remains uncertain as dissatisfaction stemming from the recent election results flows throughout a party searching for a scapegoat. 

Author’s Note: results are still being reported, so all information discussed in this article (including total seats called incorrectly) should be considered current at publication (1 PM EST, Nov. 1st 2022).