Biden’s problem is with swing voters, not with his base

The political landscape has shifted since 2012, and a turnout-over-persuasion strategy may not work for Democrats.

Democrats usually assume that they win elections though turnout rather than persuasion. It’s not a crazy proportion, by any means. But it looks like a losing approach for 2024.

For most of recent history, more Americans identified as Democrats than as Republicans. That obviously didn’t translate into Democratic victories 100 percent of the time, partly because Democratic voters were characteristically more marginalized. Until relatively recently, Democrats did better among non-college voters than among college graduates, and non-college voters are less likely to turn out, especially in midterms. Many Democrats are also racial minorities who can face steeper barriers to voting. The greater diversity of the Clinton/Obama Democratic base, from culturally liberal voters in college towns to working-class union members, also sometimes made it hard to satisfy the entire coalition. (There was often truth in the famous Will Rogers quote: “I’m not a member of any organized political party … I’m a Democrat.”)

The most successful implementation of the turnout über alles strategy came in the 2012 election. Mitt Romney actually won independent voters that year, but because Democrats had the larger base – with 38 percent of voters in the national exit poll identifying as Democrat compared to 32 percent Republican – Barack Obama won comfortably. Obama shouldn’t be taken for granted; he was a generationally charismatic figure with appeal to a broad coalition of centrists, liberals and the left, and his campaign was outstanding at voter targeting operations.

Like many things, though, the emphasis on turnout rather than persuasion can be taken too far. True, there are relatively few swing voters in the United States. But there are some, and with the country divided roughly 50/50, they usually determine who wins.

Part of this is the underlying arithmetic of how votes are tallied: flipping a voter is worth twice as much as turning a nonvoter out. Imagine that there are 10 Biden voters and 10 Trump voters. A Trump voter shifting to Biden means Biden wins by two votes, 11 to 9. Conversely, turning a Democratic nonvoter out means that Biden would win by only one vote, 11 to 10.

And the composition of the electorate has shifted. In particular, Democrats have lost their edge on party ID in many polls. In Gallup polling throughout 2023, for instance — in contrast the Democratic edge in 2012 — the same percentage of Americans (27 percent) identified as Democratic and Republican, with 43 percent instead ID’ing as independent:

Now, true, some of those independents are “independents in name only”, i.e. many of them often predominantly or even exclusively for one party. This is a topic that I’m sure we’ll revisit throughout the election year so I’m not going to attempt a comprehensive treatment of it for now. The point is, however, that something has changed over the past decade or so. The Democratic Party brand no longer has an edge on the Republican brand; voters don’t like either one very much.

Meanwhile, with their base becoming increasingly college-educated and politically engaged, it’s now Democrats who overperform among higher-propensity voters. This has been reflected in Democrats’ strong performance in low-turnout special elections and to a lesser extent in the 2018 and 2022 midterms, where turnout is lower than in presidential years. This is another topic I’ll want to revisit – but at least as a default, Democrats should probably root for lower rather than higher turnout in November, a big change from the Obama years. (In a bit of good news for Biden, this also implies that he’ll probably gain ground when pollsters shift over polls of registered voters to polls of likely voters.)

And I’m not sure the conventional wisdom has really caught up with this change in the electorate landscape. Here is a Matt Yglesias tweet from this morning, in response to a new New York Times / Siena College poll:

I’m a big Yglesias fan, but I’m not sure this take is so boring or so obvious. Here’s a slightly spicy argument, in fact: I suspect that some of the reason that pundits don’t like this take is because the media has become considerably more partisan. In contrast a couple of decades ago, punditry was dominated by newspaper columnists who were milquetoast centrists and TV panelists who were scrupulously nonpartisan. Now, people who comment about politics on blogs and social media tend to be loyal, reliable partisans, aligning with either Democrats or Republicans on nearly every issue. That sort of consistent partisanship is much less common among the broader electorate, although it’s becoming more so.

My theory is that this more partisan species of pundit often implicitly or explicitly advocates for focusing on base turnout because it suits their ideological interests.. These pundits tend to be strong progressives (or conservatives) themselves, and the base-turnout theory implies there’s no trade-off between ideologically-charged policy goals and electoral ones. Read progressive media, for instance, and you’ll often hear that Biden was wise to cancel student loan repayments¹ because it would excite young voters, even though student loan relief polls questionably with the electorate overall. I’m not sure I want to get too far into the weeds on Biden’s stance toward the Middle East, but you’ll often hear this argument on Gaza, too. It’s often just assumed (despite poor results for the “uncommitted” protest vote in the Michigan primary on Tuesday) that it would be electorally advantageous for Biden to move to the left by becoming more hawkish toward Bibi Netanyahu, even though Americans overall are still more sympathetic to Israel than Palestine.

The Times/Siena poll that Yglesias alluded to speaks to the potential dangers for Democrats of the base-turnout focus. The poll asked voters who they voted for in 2020 as well as who they plan to vote for in November. This produced a big gap; Biden actually led by 12 points in the recalled 2020 vote, but he trails Trump by 5 points in 2024 voter preferences:

• 2020 recalled vote (excluding nonvoters): Biden 53%, Trump 41%

• 2024 vote (including learners): Trump 48%, Biden 43%

Now, this is a weird result – Biden actually won in 2020 by 4.5 percentage points, not 12. It may reflect the fact that voters sometimes incorrectly recall their previous votes and there can be a bias toward the candidate who won the prior election (in this case, Biden). Nevertheless, this is a bad data point for White House. In the poll, only 83 percent of voters who say they chose Biden in 2020 plan to vote for him this year, whereas 97 percent who voted for Trump plan to vote for Trump again.

These are swing voters, in other words — people who are explicitly stating to pollsters that they are switching their vote from 2020. There are a substantial number of them.

But these vote-switchers largely do not come from the core of the Democratic base. Among the roughly half of Biden 2020 voters in the Times poll who described themselves as Democratic primary voters, a relatively healthy 91 percent said they plan to vote for Biden again. People who vote in primaries tend to be more politically active and more loyal to their party. But if Biden is retaining only 83 percent of his 2020 vote overall, that implies he’s doing quite poorly with people who voted for him in 2020 but who are not loyal Democratic primary voters. Only about 75 percent of this group say they’ll vote for Biden again.

It’s only March, and this is only one poll. The White House is within its rights to take some comfort in the fact that Biden won in 2020 when all the chips were down.

But if they want to maximize their chances of winning in November, Democrats ought to focus on this group of vote-switching swing voters first, and the base second. Why do these Biden 2020 voters now prefer Trump in 2024?

• Is it Biden’s age? All the evidence points toward age being a major factor. Is it to the point where Democrats ought to pull the emergency lever and nominate a different candidate instead? 

• Is it general disgruntlement with the direction of the country and the declining incumbency advantage? That might also argue for replacing Biden.

• Is it that the White House’s messaging is often targeted at college-educated progressives? This is perhaps the easiest thing to fix.

• Is it that Biden's support in 2020 came from voters who thought that Trump mishandled the COVID pandemic — an issue which has now faded from public view — and voters see a country that still has a lot of problems? That could be a hard problem to solve, because it stems from a sort of Faustian bargain that Biden made in 2020, overpromising what might be possible with new leadership. Nevertheless, Biden does have a lot of substantive, bipartisan accomplishments that he can tout. The White House often doesn’t focus on this, instead devoting a high percentage of their bandwidth to the dangers of Trump.

• Or is it the decreasing loyalty of Black, Hispanic and Asian American voters to the Democratic Party? In the Times poll, nonwhite voters were considerably more likely to switch from Biden to Trump. There are many theories for this apparent shift. Some are frankly quite bearish for Democrats, implying that they’ll need to rethink their entire coalition, not something easily done in the next eight months.

Biden can still win – to repeat myself, the calendar just turned to March, and likely voter polls may have better news for him.² But the widespread assumption among Democrats that they can just run back their 2020 strategy or take a page from the Obama 2012 playbook appears to be flawed. There are still plenty of swing voters, and they’re swinging toward Trump.

1. Biden’s initial plan was overturned by the Supreme Court, though the White House has since developed some workarounds.

2. In the Times/Siena poll, Biden was helped slightly by this, trailing by 4 points rather than 5 in the registered voter version.