Was Democrats’ central strategy of running against ‘ultra-MAGA’ the key to electoral success? Yes and no — and the results carry warnings for both parties.
Was Democrats’ central strategy of running against 'ultra-MAGA' the key to electoral success? Yes and no — and the results carry warnings for both parties.
A t the end of the day, every election is about two things: turnout and persuasion. If you have a big-enough edge in getting people who already agree with you to vote, you win; if both sides turn out a competitive number of their voters, the winner is the candidate who persuades the persuadables. It is impossible in an election postmortem to entirely separate the two. Getting people to care enough to turn out is itself a form of persuasion, and so is getting the other side’s voters to stay home because they think it’s not worth showing up. Still, it is important in understanding the outcomes to make our best estimate of how much of the result was driven by turnout.
The 2022 results and the conclusions one might draw from exit polls suggest that Republicans won the turnout battle and lost the persuasion war. Thus, if you believe the exit polls, we had the most Republican electorate in the past two decades of midterm and general elections, but it was an election in which Donald Trump was massively unpopular and in which independents backed the Democrats in most of the key races — something that the president’s party hasn’t managed in decades. There is, however, more than one way to look at the data. There are well-known limitations to exit polls. Perhaps we can learn more by examining turnout more carefully.
How Many Voters?
Before the Trump era, the conventional wisdom was that Republicans would benefit from lower turnout, and Democrats from higher turnout. There were good reasons for this: Republicans tended to have the support of settled adults with families, who reliably turn out, while Democrats tended to depend on sizeable numbers of college kids, poor urbanites, and other less regular voters who require effort to turn out.
That assumption has come under scrutiny as educated suburban homeowners — the most reliable voters — have been trending Democrat, while Trump built a segment of his coalition from voters lower on the educational and economic ladder who are often less engaged in politics. Like Obama, Trump drew support from people who would come out to vote only for him and could not be persuaded to show up for other elections. After much focus, during the Obama era, on how the midterm electorate was significantly older and more conservative than the general electorate, turnout in 2018 was the highest for a midterm since the 1914 dawn of popular Senate elections and the 1919 passage of women’s suffrage. Half of all eligible voters showed up, compared with 36.7 percent in 2014. Before 2018, the highest midterm turnout since 1982 was 41.1 percent, in 1994.
That heavy turnout mostly favored Democrats, who swept the House and the midwestern governors’ races and ran lots of close races in states such as Texas, where big blue turnout drove a bunch of down-ballot Democratic wins. Yet Republicans held on in major statewide races in Texas, Florida, and Georgia and broke the usual pattern by gaining seats in the Senate and knocking off four incumbent Democratic senators (in North Dakota, Indiana, Missouri, and Florida). The 2020 general election featured enormous turnout as well, also the highest in a century for a presidential election — with bad consequences for Donald Trump but surprisingly good outcomes on Election Day for Republican House and Senate candidates.
What happened this time around? Let’s start with the raw numbers, drawn from the U.S. Elections Project, run by University of Florida political-science professor Michael McDonald. Just under 112 million Americans voted, down 5.58 percent from the 118.6 million who voted in 2018. But there were significant variations in turnout among states.
Turnout was up the most, not surprisingly, in New Hampshire, Arizona, and Pennsylvania — all states with both hotly contested statewide races and more races than each state featured four years ago. Turnout was down the most in states such as Tennessee, North Dakota, Indiana, West Virginia, and Missouri, all of which had major Senate races in 2018 but not in 2022.
Of course, in order to properly interpret shifts in voter turnout, it is necessary to consider the baseline. Not every state’s population of eligible voters stayed the same, or grew at the same rate, between 2018 and 2022. That is particularly true during the Covid era, which saw elevated mortality rates and a lot of relocation between states. On the other hand, shifts in birth rates, though they may affect overall population, will not show up in the population of eligible voters for another 16 to 18 years; shifts in immigration rates take three to five years to affect the pool of eligible voters in federal elections and most statewide elections.
Booming populations in Idaho and Utah are what you’d expect. Somewhat surprisingly, given relocations and retirements, Florida’s population of eligible voters expanded by only 3.03 percent between 2018 and 2022, less than the state’s rate of population growth. The District of Columbia stands out for the vivid collapse of its entirely urban electorate. D.C. lost over 10 percent of its eligible voters, with nearly all of that decline coming between 2020 and 2022.
When we adjust for the change in voter population size, we can directly compare the share of all eligible voters who turned out. Of course, some states (such as Minnesota) have a traditionally high baseline of turnout, while others (such as Texas) have a traditionally low baseline.
Looked at this way, turnout was down 6.4 percent nationwide. Pennsylvania, New Hampshire, and Arizona still stand atop the list, along with Michigan. They were joined by two states with an open if not competitive governor’s race (Hawaii and Arkansas) and Maine, where a former two-term governor was trying to unseat the incumbent.
The Anti-MAGA Voter Thesis
Is there a more identifiable pattern to explain why turnout was up or flat in some states while being dramatically down elsewhere? Outgoing AFL-CIO political director Michael Podhorzer thinks so. Podhorzer writes from deep inside the Democrats’ activist wing, and his analysis and rhetoric should be evaluated through that prism. His thesis, however, is presented with significant data and worth considering. Unlike some of the more pie-in-the-sky Democratic pundits, he doesn’t argue that Joe Biden or the progressive agenda were somehow secretly more popular than polls reflected, or that abortion and the Dobbs decision were magic bullets for Democratic turnout. He argues, instead, that the decisive factor was a “Blue Undertow” of better-than-expected Democratic turnout — but only in some states.
Podhorzer’s basic thesis runs as follows:
- The normal midterm electorate before the Trump era favored Republicans, but the dramatic increase in turnout in 2018 and 2020 was driven by “anti-MAGA” voters motivated not by any positive agenda but by alarm at Donald Trump and everything his personality and approach represent in American politics. As Podhorzer frames the mindset of these voters, “All that was needed to confound the usual midterm rout for the president’s party was making sure that 2020 voters understood that, just as they didn’t want Trump for President, they certainly didn’t want his criminal accomplices and MAGA fascists to take over Congress and their state capitals.”
- The traditional indicators all pointed to a red-wave election. There was no big red wave across the decisive states because it was precisely those states that experienced 2018-level turnout driven by these “anti-MAGA” voters.
- If the red wave had been stymied solely by bad candidates turning off Republican and independent voters who split their tickets, those voters would have voted Republican down the ballot, but in fact, in what Podhorzer calls the “MAGA Statewide Competitive states,” Democrats also did better down-ballot in state legislative races. This suggests that they actually turned out voters who preferred a generic Democrat to a generic Republican.
- Turnout was down nationally because there were only 15 “MAGA Statewide Competitive states,” while lower turnout in the other 35 states produced results that look more like a red-wave midterm.
If Podhorzer’s theory is correct, then a lot of what Democrats did in 2022 to boost the nomination of “stop the steal” Republicans, obsess over January 6, and have Joe Biden give speeches about “ultra MAGA” turned out to be savvier strategy than focusing on public policy, the Democratic agenda, the economy, or the kitchen-table concerns of ordinary Americans. His theory does not exonerate Republicans of having nominated dreadful candidates, although it suggests a different causal mechanism for how those candidates hurt the party: not because they did poorly with otherwise-Republican-supporting voters, but because they drove turnout that hurt the whole Republican ticket in their states.
PHOTOS: Midterm Elections Day 2022
Let’s examine the evidence marshaled by Podhorzer. His list of 15 “MAGA Statewide Competitive states” is composed of states classified by the Cook Political Report with Amy Walter, Sabato’s Crystal Ball (of the UVA Center for Politics), FiveThirtyEight, and others “as very competitive at the state level (Senate and/or governor), and in which at least one statewide candidate was identifiably MAGA (whether or not they were endorsed by Trump).” The states he lists are Alaska, Arizona, Colorado, Georgia, Kansas, Maine, Michigan, Nevada, North Carolina, New Hampshire, New Mexico, Ohio, Oregon, Pennsylvania, and Wisconsin. Here is the raw-vote turnout table again, with the data for those states shaded in blue, and a separate line for the total turnout across MAGA Competitive and Non-MAGA Competitive states, using Podhorzer’s definitions.
Collectively, turnout was up 1.37 percent across those 15 states, while it was down 9.31 percent across the other 35 states plus D.C. But those 15 states also saw a 1.96 percent increase in their voter-eligible population, compared with 1.25 percent for the rest. When you adjust for the population figures, the “MAGA Statewide Competitive” states saw turnout down by 0.58 percent, while the rest of the states saw turnout down by 10.76 percent.
Podhorzer’s theory has some explanatory power, but it also seems a bit overdetermined and not entirely consistent with the numbers. To start with, his definition of which states had a “MAGA Statewide Competitive” race sometimes seems arbitrary and potentially results-driven. For example, he classifies Oregon as a “MAGA Statewide Competitive” state, presumably because nut job Jo Rae Perkins was running for the Senate, but anyone who watched Oregon politics in 2022 knew that Perkins was less competitive statewide than were Darren Bailey in the Illinois governor’s race or Dan Cox in the Maryland governor’s race. The reason why Oregon was competitive was the governor’s race, in which the Democrat faced off against both a conventional Republican candidate and a third-party campaign from a dissident Democrat. One could ask similar questions about why he counted New Mexico or Colorado but not Minnesota. On the other hand, the states he identifies did not deliver consistently high turnout; turnout was down significantly from 2018 in Kansas, and down more than the national average in Alaska and Ohio. You would be hard-pressed to identify signs of strong Democratic turnout in the Ohio election results.
Podhorzer’s evidence from down-ballot races is also inconsistent. Start with House races:
In House races across the 15 states he identifies, Republicans in 2022 received 109.9 percent of their 2018 vote totals, compared with 105.6 percent elsewhere, while Democrats in 2022 received 94.1 percent of their 2018 vote totals, compared with 80.4 percent elsewhere. In other words, yes, those states saw generally less dramatic Democratic falloff from the 2018 blue wave. In seven states, all of them on Podhorzer’s list, House Democrats actually got more total votes than they did in 2018, although in five of the seven, Republicans increased their vote by a wider margin.
What about the two parties’ share of the vote?
This is where it is harder to see Podhorzer’s thesis as a consistent dynamic. Democrats improved their share of the popular vote in House races compared with 2018 in only four states: Maine, Colorado, New Hampshire, and Alaska. Alaska is an unusual case, because it has just a single district, it changed its voting system, and Republicans were running Sarah Palin instead of eternal incumbent Don Young. Across the “MAGA Statewide Competitive” states generally, however, there was still nearly a four-point swing in the House vote, with very dramatic shifts in the Republicans’ favor in Pennsylvania, Arizona, and Wisconsin and greater than a five-point improvement for Republicans in Nevada, New Mexico, and Oregon. That’s not really consistent with the idea that Democratic turnout remained competitive with 2018 in those states; it’s more consistent with the thesis that Republican failures at the top of the ticket were the failures of those candidates. By contrast, Republicans lost a number of key House races outside of those states, including closely watched bellwether contests in Virginia, Washington, and Rhode Island.
How about state legislative races? One of the states where Democrats flipped a legislative chamber — the state senate in Minnesota — was not on Podhorzer’s list. Turnout was down by 4.98 percent in Minnesota despite a competitive governor’s race featuring a somewhat fringy Republican. Podhorzer’s thesis seems supported in Pennsylvania and Michigan when you look at how Democrats took control of the chambers, but then we look at the cumulative popular vote for state legislative chambers across 13 key states (all on Podhorzer’s list) compared with 2018:
There was a distinct shift toward Republicans between 2018 and 2022 in state legislative races in every one of these states, swinging at least two points in every state but Michigan and Georgia. There were particularly powerful swings in Pennsylvania, Arizona, Wisconsin, North Carolina, Ohio, and Nevada. Even acknowledging that — as in House races — it is tricky to compare the statewide popular vote given redistricting and uncontested races, the popular vote for Republicans across these states looks a lot more like red-wave turnout than like a blue undertow.
Podhorzer also makes much of the idea that there were few ticket-splitters. But he still has to acknowledge, first of all, that a lot of people voted for stronger Republican candidates such as Brian Kemp while leaving the weaker candidate’s race (in Georgia, Herschel Walker) blank, which is a halfway form of ticket-splitting. Second, he needs to take account of some races in which there was obviously a lot of ticket-splitting. In New Hampshire, for example, Chris Sununu got 57 percent of the vote while Democrat Maggie Hassan got 53.5 percent.
What It All Means
Podhorzer’s model is not quite all it is cracked up to be, but even if we discount the strongest version of his argument, he makes a persuasive case that Democrats did better in turning out voters in a number of key states than the exit polls would suggest. We certainly should not treat the exit polls as gospel, given that they are adjusted to fit the final outcome, and at least the national outcomes did not end up the same in the final count as they looked when the exit polls were released before the West Coast was done voting. On the other hand, “the exits undercounted California” doesn’t tell us much about the statewide exit polls in places such as Georgia or Ohio.
The observation that Democrats did better than expected at turnout in some of the races they won brings us back to a central fact about politics: Being a successful politician is not just about turning out your voters and persuading swing voters to support you. It is also about avoiding giving people a reason to turn out against you. Compared with positive turnout and persuasion, negative turnout is the factor that is typically least in a candidate’s control, because your opponents can always find ways to fire up their own base. Still, polarizing and incendiary figures can and do drive turnout for their opponents.
One of the central political fallacies of the entire Trump and MAGA worldviews is that all votes on the other side are “the Left” — a monolithic entity driven by liberal media narratives — and that turnout on that side can be affected only by laws expanding or contracting the ease of voting (legally or otherwise). So much of the political malpractice of the past five years has been driven by this refusal to consider the effect of Republican candidates and officeholders on the strength of negative partisanship.
Few figures in American political history have done more to drive negative turnout than Donald Trump. When you take the turnout data and set it next to the list of candidates who evoked the specter of Trump’s worst characteristics as a public man, it is not hard to conclude that Trump’s supercharging effect on Democratic turnout in 2018 and 2020 had a lingering effect in 2022, especially in states such as Pennsylvania and Arizona where Trump’s stop-the-steal fingerprints were all over multiple high-profile statewide Republican candidates.
This is not, however, all good news for Democrats. To start with, it suggests that the voting majorities that gave Democrats the House in 2018 and the presidency in 2020 are entirely dependent on the choices of their adversaries rather than built on voter enthusiasm for the Democratic agenda or the people running the party. With Biden still unpopular and polls suggesting declining identification with the Democratic Party, Democrats will have to put all of their 2024 eggs in one basket: convincing voters that Republicans are running Trump again, even if they don’t.
Consider the flip side of Podhorzer’s turnout thesis: In the majority of states in the country, the anti-MAGA vote stayed home because there just wasn’t an election on the ballot that could energize them. Every single incumbent Republican governor and senator in the country was reelected, most of them handily. As much as Biden tried to turn traditional Republican stances on abortion and economic issues into “ultra-MAGA,” voters in over two-thirds of the country didn’t buy it.
It is not surprising that anti-MAGA voters didn’t turn out in the deepest-red or deepest-blue states without competitive races, or that they were not roused to alarm by the most moderate Republicans. But consider some of the candidates whose states fell outside of Podhorzer’s list. Blue-undertow turnout didn’t materialize in Utah, where the whole campaign against Mike Lee was framed as anti-MAGA, or in New York, where Kathy Hochul hammered Lee Zeldin for voting against certifying the 2020 election but won by only 5.6 points in a low-turnout race in a deep-blue state. It didn’t show up in Texas, a hotbed of heavy anti-MAGA turnout in 2018, in spite of all manner of national and statewide media efforts to turn Greg Abbott into a menacing hate figure.
And of course, it didn’t show up in Florida. In 2018, turnout in the Florida governor’s race jumped to 8.2 million voters from 5.9 million in 2014. The perennial purple state produced all manner of high hopes for Democrats, and with Ron DeSantis running comically pro-Trump ads in his primary race, it was easy to paint him as a generic MAGA candidate. He won by 0.39 points over Andrew Gillum despite an FBI probe hanging over Gillum’s head.
With a full-court press by the national media aimed at making DeSantis the new “worse than Trump” figure and DeSantis himself happily embracing much of the MAGA policy agenda and careful to avoid publicly offending Trump, one might have expected any anti-MAGA turnout strategy to get some traction against the Florida governor. His opponent, Charlie Crist, may have been an uninspiring and transparently insincere opportunist and two-time statewide loser, but as a sitting congressman and former governor and state attorney general, Crist was a plausible-enough potential governor to be the receptacle for votes driven by negative partisanship, just as creaky old Joe Biden was in 2020.
Instead, while DeSantis was adding 600,000 voters to his winning total, Democratic turnout collapsed. Crist drew nearly a million fewer voters than Gillum did. Val Demings drew 800,000 fewer voters than Bill Nelson did in 2018. Democratic House candidates in Florida drew 400,000 fewer votes than in 2018. At the state legislative level, the vote shifted 2.1 points toward Republicans in the Florida House and 6.2 points in the Florida Senate. Anti-MAGA voters in Florida simply did not behave as if they viewed DeSantis in 2022 in the same way that anti-MAGA voters viewed Trump or the Trumpier candidates in other states.
A national race would feature some different dynamics for DeSantis, or for that matter for Abbott or Brian Kemp. But it should worry Democrats that their central political strategy of running against Trump and “ultra-MAGA” — which worked against a lot of Republicans in major races in 2022 — simply didn’t work against Ron DeSantis.