It's 2004 all over again

And that might not be such a bad thing for Democrats, especially with their Electoral College disadvantage erased.

I wish I could give you a crystal-clear sense of what Democrats felt like in 2004 after George W. Bush was elected to a second term. But the truth is, my memories of that time are hazy. That’s partly because I was a twentysomething doing dumb twentysomething things, but mostly because it was before I really got into following politics — the catalyzing event, instead, came two years later, when Congress passed a law to essentially ban payment processing to online poker sites, then my primary source of income. So everything before 2006 is pre-history to me.

From what I recall, though, it was a time when Democrats felt like they were really on their heels and liberalism was in retreat.

In 2000, Bush had lost the popular vote and won by only 537 votes in a disputed outcome in Florida. This time, there was no question that he’d been elected legitimately, with Republicans’ first majority since 1988. Or almost no question: there was some denialism (from Robert F. Kennedy, Jr.!) and some poll unskewing (the Princeton Election Consortium site insisting with high confidence on a John Kerry victory). Still, I remember a lot of rhetoric about America being a “center-right” nation. I remember the Jesusland map. I remember our friends the latte-sipping liberals.

Of course, it didn’t work out so badly for Democrats. Bush’s second term was a disaster, marked by the failure of Social Security reform, the ongoing quagmires in Iraq and Afghanistan, and Hurricane Katrina. Democrats had a strong 2006 midterm, gaining 33 seats — including from some of the bastards who had taken away my poker. And then Barack Obama romped to the largest Electoral College and popular vote win of the 21st Century so far. Progressivism was in the ascent — until the Tea Party came along in 2010 and whipped Democrats back to reality again.

It’s hard not to see the parallels between Bush’s win in 2004 and Donald Trump’s last week. Like Bush, Trump won thanks partly to a surge of votes from Latino and Asian American voters. Like Bush, he’ll win the popular vote — probably by a margin of around 1.4 percentage points once all votes are counted. He’ll probably come just short of an outright majority, although it will be close, and Trump’s Electoral College margin was more impressive than Bush’s, who was only one state (Ohio) away from losing to Kerry. 

Certainly, the mood feels very different than after Trump’s first win in 2016. Democrats have approached the outcome in a more cerebral and analytical way than I was expecting, with manifestos about a new way forward for the party and unapologetic shifts away from “wokeness”. The party is ready to move past the Clintons, the Bidens and the Obamas — well, unless Michelle Obama decides to run, I guess — and it would be stunning if there’s any appetite to nominate Harris or Tim Walz again.

With a clear, undisputed outcome on relatively high turnout — likely in the range of 156 million votes, just a hair down from 2020 — there’s less talk about the term-limited Trump being an existential threat to democracy, rhetoric that may have been persuasive to people like me but never resonated with swing voters. And although you can find people who blame the New York Times for the outcome, for the most part, the vibe shift is in the opposite direction, away from academic polemicism and toward constructive strategic conversations about how Democrats can regain their footing for 2026 and beyond.

Trump is likely to be unpopular

And like Bush, Trump will probably have a challenging second term:

• Most incumbents are unpopular these days, especially in their second terms. The incumbent party has now lost three presidential elections in a row.

• According to the national exit poll, Trump was elected with a tepid 48 percent favorability rating. However, he actually won 9 percent of voters who had an unfavorable view of him. These people may have seen Trump as a superior alternative to Harris but will be less tolerant now that he has no more elections to run in. In the same poll, 44 percent of voters have a very unfavorable view of Trump, so the ceiling on his popularity is likely to be fairly capped.

• Trump’s most likely successor, vice president-elect JD Vance, is also unpopular.

• Trump’s plan to enact tariffs may significantly increase inflation, which we know voters are highly sensitive to.

• Democrats have the more passionate voter base that more reliably turns out in midterms and special elections.

• When he’s inaugurated on January 20, Trump will be just as old — actually a few months older — than Joe Biden was at the start of his term.

• And he’s inheriting all of the problems that Biden faced, like the unstable situation in the Middle East.

• Trump will likely emerge with only a very narrow majority in the House, one where his agenda will often be held hostage to the Freedom Caucus. Depending on how the last few races are called, special elections or absences could potentially even put control of the chamber on the line before 2026.

• Trump might not be interested in gutting the welfare state, but some Republicans in Congress will try to — and Democrats can target them in 2026.

• And thermostatic shifts in public opinion will begin to favor Democrats again. I don’t expect wokeness to make a comeback, but things like anti-immigrant sentiment may decrease, especially if Trump tries to carry out mass deportations.

Mind you, I wouldn’t count on this being a happy time for Democrats. Trump has a lot of unchecked power and a 6-3 Supreme Court majority. But electorally speaking, these are the seeds out of which comebacks are made, and there’s a good chance that Trump is another octogenarian who overreads his mandate and overreaches in all sorts of ways. It was an impressive win, but the country isn’t that different. Trump will get 3 to 4 million more votes than he did in 2020, while Harris will get about 6 million fewer votes than Biden.

And unlike in some elections, Democrats are chastened by their loss. At betting markets, a group of relatively moderate candidates — Gretchen Whitmer, Josh Shapiro, Mark Cuban, Pete Buttigieg, and Andy Beshear — have a combined 55 percent probability of being the 2028 nominee versus 17 percent for Gavin Newsom, who would be the closest thing to running Harris’s ideas back in a white male package. AOC is the only real left-wing candidate on the board, with a 5 percent chance. The party has some bad habits to overcome, especially the tendency to ostracize people who only halfway agree with them. But it’s a relatively rational party — one that was smart enough to recognize that Harris at least gave them a shot when Biden didn’t, while also lamenting that an earlier Biden exit might have given them more options. 

The 2006 and 2008 elections were the high-water mark for Democrats’ “Fifty-state strategy,” the idea that the party ought to compete pretty much everywhere, including by nominating moderates who were good fits for red states and districts. Such a strategy is less likely to succeed now because polarization is stronger and “candidate quality” effects are weaker. But that period was also marked by a spirit of inclusiveness, where heterodox liberal centrists like me felt welcome in the broad anti-Bush coalition. Losing the popular vote casts a pall: Democrats can’t even tell themselves the myth anymore that they’d be in the majority if everybody voted, not in an electorate where Trump turns out the more marginal voters. The party basically has no choice but to hang a HELP WANTED sign outside of DNC headquarters.

A level electoral playing field

There was also some hidden good news for Democrats in last week’s results. Their structural disadvantages in the House, Senate and Electoral College are now much less. Let’s take these one at a time.

The Electoral College

As predicted by the model, Pennsylvania was the tipping point state¹. Its vote count isn’t quite finalized yet, but Trump leads there by 2.0 percentage points. If my projection of a Trump +1.4 popular vote is correct, that means the tipping point is only 0.6 percentage points to the right of the popular vote. Michigan and Wisconsin are now left of the tipping point, meanwhile, 

And Georgia — even though it flipped from a Biden win to a Trump win — actually moved considerably toward the tipping point: the Atlanta suburbs and exurbs are one of the few places where Harris gained ground relative to Biden. And North Carolina is now only 2 points away from the tipping point. Even if Arizona becomes a challenging state for Democrats, that’s 308 electoral votes where Democrats are seriously set to compete in 2028.

Some observers have hypothesized the existence of a battleground state effect by noting that Harris’s margins in the seven swing states moved less than the national numbers. Overall, there was a 5.9-point swing toward Trump, assuming again that my popular vote guesstimate is correct. But Wisconsin swung by just 1.5 points toward Trump, North Carolina by 2.0, Georgia by 2.4, Pennsylvania by 3.2 and Michigan by 4.2. (Arizona and Nevada were close to the national swing, however.) Perhaps Harris’s ground game or TV ads were more effective than they were credited for?

Maybe. But I’m skeptical because the shrinking Electoral College/popular vote gap can instead be almost entirely explained by what happened in six noncompetitive states. Democrats’ erosion in California, where they won by 29.1 points in 2020 but are headed to “only” a 20.8-point win based on votes counted so far, cost them 0.9 points off their national popular vote margin — even though it didn’t hurt their Electoral College chances at all. New York cost them 0.6 points, and New Jersey and Illinois 0.3 points each. So they lost the most votes in the places where those votes were most wasted.

Meanwhile, Republicans impressively ran up the score in Texas and Florida — but now they have an excess of voters in those populous states. Their gains in Florida alone were responsible for 0.7 points of national vote swing, and Texas another 0.6. These six states then — the four blue states plus the two red ones — combined to reduce the Electoral College penalty for Democrats by 3.3 points, wiping it out almost entirely. Considering that we saw similar swings in the 2022 midterms — Democrats holding up relatively well in the Midwest, but having big shifts against them in New York and Florida — this is probably the new normal, a map remade by COVID-era migration patterns and racial depolarization.

The Senate

Considering the Senate’s importance to the Supreme Court and other things, I initially thought the losses Democrats took there were in some ways a bigger blow to the party than even another Trump term. A week later, though, the situation looks considerably better for Democrats than it did on Election Night, with the party securing narrow holds in battlegrounds Arizona, Nevada, Michigan and Wisconsin — though losing Pennsylvania. That’s still a 53-47 majority for Trump — importantly, one that’s big enough to withstand defections from Susan Collins and Lisa Murkowski. Still, there’s a path back for Democrats. Let’s break the incoming Senate into three broad categories.

• Harris states (38 seats). Democrats control 37 of 38 seats, with Maine (Susan Collins) as the lone exception.

• 2024 battleground states (14 seats). Democrats control 10 of the 14 seats, the exceptions being one of the Wisconsin seats (Ron Johnson), one of the Pennsylvania seats (where David McCormick is Senator-elect) and both North Carolina seats. Of these, 5 will not come up for re-election again until 2030, meaning they’re locked in place for whoever is chosen president in 2028. Democrats control 4 of those 5 seats — the races they just won in Wisconsin, Michigan, Nevada and Arizona. The GOP controls only McCormick’s seat in Pennsylvania.

• Solid Trump states (48 seats). Republicans now control all 48 seats, following Democratic losses in Ohio, Montana and West Virginia.

So, Democrats have a low ceiling in the Senate, while those 48 senators in deep red states give Trump a nice high floor. Democrats will have to try to compete in some of them. The best target is probably Ohio, which will hold a special election in 2026 as Vance advances to the vice presidency. Sherrod Brown lost his race by only 3.8 points in an R +1.4 national environment this year. In a typical midterm environment that favors the “out” party — say, a D +5 or D +6 map — he could probably be competitive if he wants to run again. By the same logic, Jon Tester (who lost by 7.4) could give things another try in Montana. In Alaska, Mary Peltola looks likely to lose the state’s lone U.S. House seat, though 20 percent of the vote is outstanding and Alaska has ranked-choice voting, which could reverse a small deficit. However, she won statewide in the last midterm, 2022. It’s one more option for Democrats against the relatively unpopular Dan Sullivan, although Peltola could be hurt if Alaskans repeal ranked-choice voting. And while I can understand why Democrats might want to give up on Florida at this point, there will be another special election there in 2026 if Marco Rubio becomes Secretary of State.

But setting those long-shot options aside — though Ohio isn’t really such a long shot if Brown runs — the math is fairly simple. Democrats have to win 3 of these 4 races to have a Democratic majority for a potential president Whitmer or Shapiro or Cuban in 2028:

• Susan Collins in Maine (2026)

• Thom Tillis in North Carolina (2026)

• Ted Budd in North Carolina (2028)

• Ron Johnson in Wisconsin (2028)

And yes, Democrats also have to hold on to all of their own seats, including what are likely to be challenging races in Arizona, Nevada, and Georgia. However, a party generally doesn’t lose its own swing state seats in a typically friendly midterm environment. Thus, Georgia’s Jon Ossoff and Michigan’s Gary Peters are probably in relatively good shape for 2026. There’s more downside risk to the party in 2028, and conditional on a Republican winning the presidency, they’ll almost certainly maintain a GOP Senate. But if Democrats win the presidency, you’d bet on just enough tailwinds to get them the Senate, too — not a cinch, obviously, but at say 60 or 65 percent odds.

Meanwhile, the complete sweep of Tester and other Democrats out of the solid Trump states does have one hidden benefit to the party: Democrats no longer have a soft underbelly of seats that were likely to fall at the next presidential year. (Tester and others benefited from competing in a string of strong D cycles — 2006, 2012 and 2018 — until their luck ran out this year.) And progressives no longer have Kirsten Sinema or Joe Manchin to complain about.

It’s true that the Senate tipping point is still a bit Republican-leaning: either North Carolina (2.0 points more red than the rest of the country) or Arizona (4.2 points), depending on whether you need 50 seats or 51. This margin, however, is considerably less than in heading into the 2020 cycle, when the same two states were at the Senate tipping point but leaned GOP by 5.8 and 7.4 points, respectively. Sonia Sotomayor should probably have listened to us and retired — but she may get bailed out if Democrats can finally knock off Collins or turn narrow losses in North Carolina into narrow wins. 

The House

That the House is still uncalled in an election where Trump won the popular vote is a favorable sign for Democrats. Now, the writing is on the wall: Polymarket shows nearly a 100 percent chance that the GOP will eventually win the chamber, as Democrats currently won or have leads in 213 seats, five short of the number they need. But the GOP bias built into the map in place from 2012-2020, buoyed by the GOP dominating the redistricting process after their strong 2010, is no longer.

This chart shows every race where the margin is somewhere between D +5 and R +8. (Uncalled races are highlighted.) As I said, Democrats are currently ahead in 213 races. But what if the national environment were 1.4 points better for them, meaning a tied national popular vote rather than Trump +1.4 — and this shift also resulted in a 1.4-point gain for Democrats in every House race? Then they’d win 218 seats, splitting the difference between the two that would project to be exact ties (CA-45 and PA-10).

In other words, there is no longer any bias in the median House seat. If Democrats win the popular vote by any margin in 2028, they’ll be favorites to win the House, and the same is true for Republicans.

And on a D +6 map — not so far-fetched for the midterms — Democrats would win 234 seats, meaning a 21-seat gain from this year. Now, their upside is somewhat capped: both sides used the 2020 restricting process to protect their incumbents, reducing the number of competitive seats. Still, especially in a parliamentary era with strong party discipline — and The Squad is both smaller and more cooperative with the rest of the party than the Freedom Caucus — a Democrat in the White House in 2028 would probably have a functional working majority.

So Democrats will get a relatively fair shake in 2026 and 2028, and what is probably going to be a decent electoral environment. Beyond spending about a billion dollars in trying to defeat Susan Collins, does that mean they can get away with only making incremental changes?

The truth is they might get away with it, because this loss was incremental. Add 2.5 points to Democratic margins across the board and Harris wins the Electoral College 286-252 and Democrats win the House 222-213 — although they’d still lose the Senate, regaining only Bob Casey’s seat in Pennsylvania. But I think that would be a mistake. One of the reasons that there tends to be a lot of mean-reversion in elections — majorities aren’t held for very long these days — is because healthy political parties react to their losses. Democrats got a clear message from Hispanic and Asian Americans that their votes can’t be taken for granted and from blue states that voters have limited tolerance for a left-wing agenda on crime, schools, taxes, and identity politics. That’s not to suggest that a dose of populism couldn’t help, and Democrats will need better messaging on Elon Musk’s role in the government, Trump’s cronyism, and how Trump’s policies benefit the rich.

But although Democrats who relied on unskewed polls may have been shocked by the result, and although there’s no telling what Trump’s second term could bring, this was a fairly ordinary election in many respects. And what normally happens is the losing party rebounds — so there’s always the next cycle for Democrats to look forward to.

1. Shapiro, she should have picked.