Goodbye, Iowa. Hello, Michigan. Which Democratic primaries will matter most in 2028?
South Carolina will vote first, which could help moderates. But Michigan and Virginia may hold the real power.
The presidential nomination process is a strange “game.” Early states like New Hampshire and South Carolina don’t actually have very many delegates, and Democrats use a highly proportional system, so it’s very unlikely that they’ll prove to be the tipping-point states in determining who wins the majority at the 2028 convention. They matter mostly to the extent that the media and “party elites” (e.g., donors, elected officials, partisan influencers) believe they matter and treat them as reliable signals about the viability of the candidates.
And sometimes those expectations can be gamed. In 1992, Iowa Sen. Tom Harkin received 76 percent of the vote in the Iowa caucuses, with “uncommitted” in second. But this produced little in the way of momentum for Harkin because Bill Clinton and the other candidates had taken their ball and gone home. The media went along with this small conspiracy to deny Harkin his momentum1, camping out in New Hampshire instead. Clinton dubbed himself “The Comeback Kid” despite finishing second there to Paul Tsongas — and the media went along with that framing too.
So I don’t want to treat the order of states too mechanically. Although “momentum” can be measured statistically though indicators like national polls, candidates will seek to manipulate expectations and exploit various biases in media coverage. Moreover, the early states aren’t typically that predictive of the nominee. Many of them have a lifetime success rate below 50 percent in picking the eventual winner.
But having built a model of the primaries in 2020 — honestly, maybe the most complicated forecast we’ve ever designed — I have quite a few thoughts about the new calendar. Don’t worry: we’ll have a lot more coverage of the midterms soon. (Yes, we’re finally making excellent progress on the model.) But this is the first newsletter that I’m going to label as a Model Talk column related to the 2028 election.
That’s because last week, Democrats proposed their 2028 primary calendar:2
South Carolina: Saturday, January 22, 2028
Nevada: Tuesday, February 1
New Hampshire: Tuesday, February 8
New Mexico: Tuesday, February 15
Michigan: Tuesday, February 22
Virginia: Tuesday, February 29
Super Tuesday: March 7
Did Democrats forget about Iowa? Nope, after a colossal screw-up in reporting results in 2020, Democrats relegated it outside of the early-voting window. Meanwhile, New Hampshire has been demoted to holding the third primary.
New Hampshirites will complain that the state has a law requiring it to have the first primary.3 I don’t see why the Democratic Party or anyone in the other 49 states should really care about that4 — though along with Iowa, it might try to hold an earlier, unsanctioned contest. If these states seek to defy the calendar, the question is whether candidates and the media play along.
In 2008, Michigan and Florida jumped the queue in the primaries, and were punished with having their delegates stripped.5 Hillary Clinton and Barack Obama pledged not to campaign there — though Clinton sort of cheated on this promise in Florida, while Obama wasn’t even on the ballot in Michigan. Clinton’s wins in these states were treated, at least in my recollection, with an extreme asterisk by the media, providing only some fraction of the momentum that they usually would.
Putting minority-heavy states first could help moderate candidates
Nobody has a right to complain too much about the new calendar, though some progressives are kvetching about it. It’s reasonably well balanced between different parts of the country and the different racial groups that make up the Democratic Party. And it probably diminishes the importance of the first four states, regardless of their order, as compared to #5 and #6: Michigan and Virginia, which will offer opportunities to build up momentum before Super Tuesday, when far more delegates will be at stake.
Still, overall, I would say the new calendar puts a slight thumb on the scale in favor of more moderate candidates.
In most cases, you might associate concerns about racial equity with the more left-wing factions of the Democratic Party. There is, however, an ironic twist when it comes to the perpetual debates about the primary calendar: it’s usually the moderates who express more need for racial balance.
The reason is a phenomenon that might seem counterintuitive, but is well-known to any veteran of the process. White voters overall are more conservative than members of racial minority groups. But among Democrats, the ideological ordering is reversed. White Democrats are generally really liberal, while Hispanic and Black Democrats are more moderate. You’re seeing that play out in Michigan right now, where Haley Stevens is hoping for a strong performance among Black voters to come back against Abdul El-Sayed’s coalition of white liberals and Arab-American voters in the state’s primary next week.
In data I gathered from the Cooperative Election Study from 2016 through 2025, 65 percent of white Democrats identify as “liberal” or “very liberal”. But only 44 percent of Hispanic Democrats and 41 percent of Black Democrats do so.
All of these figures probably lowball the liberal share we’ll see in 2028, both because the Democratic Party is becoming more liberal and because more progressive voters are usually more enthusiastic than moderates and turn out at greater rates in the primaries. But I don’t see any particular reason to expect a reordering of the relative ratio of progressives to moderates among the different racial groups within the party.
Let’s take the rest of this in three stages. Since this is a Model Talk column, hopefully you’re not averse to a little bit of obsessive detail. Because in the primaries, small details tend to matter.
First, I’ll show you some data on the ideological makeup of the new First 6 states. (And the other 44 states also.)
Next, I’ll show you the historical results from the early-voting states, plus now demoted Iowa.
And then, I’ll describe some findings from our experience in actually modeling the primaries about how much the order of the states matters in the first place, and even speculate a bit about how this might affect some of the leading candidates.