The following article by Ruy Teixeira, senior fellow at the American Enterprise Institute, politics editor of The Liberal Patriot newsletter and co-author with John B. Judis of the new Book “Where Have All the Democrats Gone?,” is cross-posted from The Liberal Patriot:

Cast your mind back to those heady days of 2012 when Barack Obama vanquished Mitt Romney and won a second term in the White House. In that election, Obama carried nonwhite working-class (noncollege) voters by a massive 67 points, while losing white college graduates by 7 points. That means Obama did 74 points betteramong the nonwhite working class than among white college graduates.

In the next two presidential elections, that differential steadily narrowed as Democrats did worse among nonwhite working-class voters even as they improved among white college graduates. In 2020, Biden carried the nonwhite working class by 48 points (19 points less than Obama did in 2012) while carrying white college graduates by 9 points (16 points better than Obama). That cut the Democrats’ positive differential between these two groups almost in half, down to 39 points.

Now it’s Biden running for a second term and, astonishingly, that positive differential may have entirely disappeared. According to the just-released New York Times/Siena poll, Biden is actually doing worse among the nonwhite working class, carrying them by a mere six points, than among white college graduates, where he enjoys a 15 point advantage over Trump. Amazing. There is perhaps no better illustration of the Democrats’ transformation into a Brahmin Left party, beloved by the educated but increasingly viewed with suspicion by the working classes of all races.

The Times data allow us to dig into the attitudinal differences that currently exist between white college graduates and the nonwhite working class and help explain these trends.

1. Biden’s job approval. Biden’s job approval among the nonwhite working class is a dreadful 34 percent. Among white college graduates it’s a comparatively healthy 47 percent. The nonwhite working class is actually closer to the white working class’ assessment of Biden (28 percent approval) than to white college graduates’ view of Biden.

2. Trump favorability. Trump has a 44 percent favorability rating among nonwhite working-class voters. That’s lower than the 55 percent rating among the white working class but still closer to that rating than Trump’s 29 percent favorability among white college graduates.

3. Personality and temperament to be an effective president. White and nonwhite working-class voters are very close on the assessment of Biden (42 percent vs. 46 percent say Biden has what it takes to be an effective president), way lower than the 63 percent thumbs up from white college grads. And on Trump’s capabilities, the two working-class sectors are almost identical (48 percent vs. 47 percent think Trump can do the job), way higher than the 29 percent among white college voters who agree.

4. Assessments of the economy. About three-fifths (59 percent) of the nonwhite working class characterizes current economic conditions as “poor.” That’s identical with views among the white working class and way more negative than among white college voters, only 38 percent of whom believe the economy is that bad.

In terms of whether the economy is better or worse than a year ago, a mere 15 percent of nonwhite working-class voters believe the economy has gotten better while 51 percent believe it has gotten worse. That’s close to white working-class voters, where 21 percent say better and 42 percent say worse. But for white college graduates views are much sunnier: 32 percent better and 29 percent worse.

Looking back further, the nonwhite working class deems the economy worse rather than better than four years ago by 67 percent to 12 percent, quite similar to the white working class at 74 percent worse/16 percent better. That compares to white college grads at 50 percent worse/30 percent better.

5. Personal assessments. Just 12 percent of the the nonwhite working class are willing to say that Biden’s policies have helped them personally. That’s actually slightly less than the 14 percent of the white working class who are willing to say the same thing. Both sectors of the working class trail white college voters, who are much higher (if hardly enthusiastic) at 28 percent on this measure.

An even starker contrast is on assessments of personal financial situation. Over half of nonwhite working-class voters describe their financial situation as only fair (33 percent) or poor (23 percent), as do white working-class voters (33 percent only fair/20 percent poor). But only a third of white college grads feel that way (27 percent only fair/just 7 percent poor).

6. Cost of living. Specifically asked to rate “prices for food and consumer goods” as an aspect of the current economy, 70 percent of the nonwhite working class assesses the situation as poor, as does 71 percent of the white working class, while white college graduates at 54 percent are significantly less negative. The same pattern is evident on gas prices. Fifty-five percent of both nonwhite and white working-class voters characterize the situation as poor, compared to only 35 percent of white college grads.

This convergence of views between the nonwhite and white working class, and their divergence from those of white college graduates, helps explain the trends we have been seeing. Clearly, the Biden years have been experienced by nonwhite working-class voters in a different and less pleasant way than they have been by white college graduates. As the data reviewed here suggest, a lot of this is about the economy. But there are other factors.

As I noted last week, as the Democrats have moved to the left on sociocultural issues, they are increasingly diverging from the comfort zone of the moderate-to-conservative supermajority of nonwhite working-class voters. Democratic positions in these areas are, however, congenial to white college graduates who have rewarded Democrats with increasing support. The accelerating “Brahminization” of the party is the result.

Can Democrats escape the negative effects of this Brahminization? It’s certainly possible though oddly their chances will be best in a relatively low turnout election, where their educated, engaged voters are more important and working-class voters less. And it is mathematically possible for Democrats to drive up their support among white college grads sufficiently to counterbalance whatever losses they might experience among working-class voters. Mathematically possible but not easy.

To get a sense of how heavy the lift could be here, consider a scenario where both the white and nonwhite working class move away from the Democrats by 10 margin points—slightly more than indicated by the Times data among white working-class voters but far less than indicated among nonwhite working-class voters. This can be simulated using States of Change data that allow both nonwhite and white working-class preferences by detailed subgroup (race, gender, age) nationally and within states to be estimated for 2020 and then moved to the right by the specified amount. Those adjusted estimates (with all other preferences held constant at their 2020 levels) are then applied to the projected structure of the eligible electorate in 2024 and subsequent elections.

In 2024, this shift toward Republicans among both nonwhite and white working-class voters produces a solid 312-226 GOP electoral vote majority. The states that move into the GOP column are Michigan, Pennsylvania, and Nevada by 3 points, Arizona and Georgia by 4 points and Wisconsin by 5 points—all six of the key swing states for the coming election.

Republicans also carry the popular vote, albeit by just a point. Thereafter, the GOP starts to lose the popular vote but continues to win the electoral vote through 2040. Again, the Democrats can conceivably counter these—or any other losses—by a sufficiently improved performance among other parts of the electorate. But it does give you a sense of the working class-sized hole the Democrats may be digging for themselves.

It might be worthwhile for Democrats to start seriously thinking about how they can de-Brahminize their party. The rewards could be great—and the penalties for not doing so even greater.