A longtime political strategist weighs in on the X factor that could help Democrats in 2026 — and hurt them in 2028.
America is in the midst of a vast political realignment — not entirely caused by Donald Trump, but certainly accelerated by him. And the biggest indicator of how someone will vote is now based on education. Went to a four-year college? You’re likely backing the Democrat. Skipped it? There’s the Republican lever.
These are some of the most important observations in a new report on the state of American politics ahead of the 2026 midterms from Doug Sosnik, a former senior adviser to President Bill Clinton and now at the Brunswick Group, a global advisory firm. Sosnik has become well known in Washington for his slide decks of political analysis, and this memo and slide deck lay bare the degree to which education polarization has reshaped American politics.
In 2024, Democrats carried 14 of the 15 most college-educated states, while Republicans captured 14 of the 15 least college-educated states (Utah and New Mexico, respectively, bucked the trend).
A year out from the midterms and with Trump’s approval rating on the decline, Sosnik told POLITICO Magazine that he expects Democrats to do well. The House is within reach for the party, though the Senate will obviously be tougher to take.
But Democrats would be foolish to think they’ve solved their problems if they do succeed in 2026. College-educated Americans are higher propensity voters — they’re more likely to turn out in a midterm election. When the next presidential election arrives, the makeup of the electorate will once again shift.
“2028 and 2026 are completely different,” he said.
This interview has been edited for length and clarity.
As we look through this report, what results jumped out to you as something different from the political consensus heading into 2026?
There’s been a rule of thumb, actually going back to the 1950s but certainly in the period of time that I’ve been in politics now for 40 years, that in midterm elections, the president’s job approval is the single best predictor of the outcome. And what also has happened in the last 20-25 years is that the president’s job approval around Labor Day of their first year in office tracks pretty closely to their job approval for their entire term of office. And Trump, more than any other president, has a much narrower band between his low marks and his high marks of job approval.
If you go back to 1994, when Clinton had the same job approval as Trump does now, it was the biggest midterm election swing against the incumbent party since at least the early 1970s. And then in 2010, Obama had the same job approval as Trump, and that turned out to be a bigger swing against the incumbent party, I think the biggest since World War II.
But due to structural changes in our politics, which are largely due to a realignment in our politics based on education levels, even if the Democrats were to have a really great election cycle in the midterms, there’s going to be a limit to how many seats they can win back due to these structural changes.
If you look at Trump’s job approval on issues, he’s underwater on everything, particularly way, way lower now on the economic ratings, on inflation, and even immigration now is underwater. So you would think that his total job approval, currently around 44 percent, would be lower.
The bottom line is based on historical standards, Trump and the Republicans should be headed to a really bad midterm election. But because of these changes in our politics, due to realignment based on education, they’ll be more insulated than they would have been in the past from a tsunami-type of midterm.
Trump’s approval rating suggests Republicans are in for a rough 2026, at least in the House. But Biden and Democrats managed to largely overcome very similar numbers in 2022. What’s the likeliest outcome next year, and what would Republicans need to keep losses to a minimum?
We’ve only had seven swing states in the last three presidential elections. The other 43 are not. And the reason the other 43 are not is they skew towards either a high level of college education or low level of college education. And that’s why 86 percent of the states in this country are not competitive. What distinguishes almost all of those seven states is that they all tend to be in the middle on education.
So in 2022, there were probably three factors that helped Democrats. The first was, I do think that the Roe decision was a factor, especially for the higher percentage of the electorate that is college-educated that votes in the midterms. The second thing is that the basket of states in the Senate that were on the ballot were disproportionately in those battleground states where it wasn’t a layup for either party. The third reason is that the states overall were more favorable to Democrats. For instance, they picked up two governors’ races because you had Republicans in Massachusetts and Maryland, two of the most highly educated states in the country. So, I think it was the basket of states that enabled the Democrats to defy history. Those same conditions won’t be available to Republicans in 2026.
The Senate is obviously tougher for Democrats. What are their chances at winning back a majority?
Right now, there are only six Senate races that are considered highly competitive or somewhat competitive, and three of them are held by Democrats. So that’s a huge advantage for the Republicans in terms of limiting their losses.
Now for the Democrats to take the Senate back, they have to enlarge the map in Texas and Ohio. I think there are some reasons so far, at least, that there should be some concern by Republicans. They’ve done a poor job of recruitment. New Hampshire, which was considered a toss-up by the Cook Political Report, now leans Democratic because of poor candidate recruitment by the Republicans. They’ve done a poor job of recruitment in Georgia as well. And they’ve got a ticking time bomb in Texas, where if Attorney General Ken Paxton becomes the nominee, that state all of a sudden becomes winnable. The Democrats have also done a very good job in recruitment getting, without question, the best possible candidate in North Carolina and Ohio.
How does the fact that Trump is not a regular politician but a movement leader affect how the midterms might shake out?
There is some historical analog which should give some heartburn to Republicans. In almost 50 years, I think there have been probably three people that have been what I call movement candidates — Reagan, which was a political movement, that was probably the last time our politics had been realigned that way. And then Obama and Trump, and what happened for both Obama and so far with Trump is that movements are, by definition, built around the popularity of an individual, and that is largely proven to be not transferable in the case of Obama Democrats and in the case of Trump Republicans.
You look at Obama’s victory in 2008, I think it was the biggest political victory by a Democrat for president since Lyndon Johnson. If you look two years later, that was the biggest loss by a party since World War II. Because the people that voted for Obama were attracted to Obama. They weren’t necessarily attracted to the Democratic Party. Republicans got battered in 2018, lost over 40 seats in the House, and they underperformed considerably in 2022. So there’s no evidence up until now that this movement to Trump has transferred to other Republicans when he’s not on the ballot.
One of the starkest parts of your deck was the degree to which Democrats are now relying on college-educated voters, who also tend to participate more in midterm elections. Does this combination mean they have a better chance next year in a lower turnout scenario?