Earlier this month, Republicans took the mayoralty in McAllen, Texas. This is an 85 percent Hispanic city which Hillary Clinton carried by 40 points in 2016; that margin was cut in half by Trump in 2020 and now we have an outright GOP win in the city.

What’s going on? Wasn’t demographic change supposed to help the Democrats? Was that idea wrong?

Not exactly. The important thing is to think about demographic change in all its aspects, with all its moving parts, instead of relying on a one-size-fits-all interpretation. Then it is much easier to see how demographic change can help, hurt or even have no effect on electoral outcomes. The analysis below, reprinted from the Persuasion site with permission, provides a logical way of thinking about these possibilities so that results like McAllen (or, more broadly, Hispanic voting in the 2020 election) are less difficult to understand and process.

Rising racial diversity is an ongoing trend in the United States, as the latest Census Bureau data confirms. In fact, the figures suggest that over the past decade, the white population will have declined for the first time in the nation’s history.

This demographic change is generally understood to be beneficial to the Democrats’ electoral fortunes, as John Judis and I argued in our 2002 book, “The Emerging Democratic Majority.” That’s a reasonable viewpoint based on a very simple idea: If voter groups favorable to the Democrats (racial minorities) are growing while unfavorable groups (whites) are declining, that should be good news for the Democrats. This is called a “mix effect”: a change in electoral margins attributable to the changing mix of voters.

These mix effects are what people typically have in mind when they think of the pro-Democratic effects of rising diversity. But mix effects, by definition, assume no shifts in voter preference: They are an all-else-equal concept, as we were careful to stress two decades ago. If voter preferences remain the same, then mix effects mean that the Democrats will come out ahead. That is a mathematical fact.

But voter preferences do not generally remain the same. Therein lies the reason why, in some cases, rising diversity has not produced the dividends for Democrats that many activists and advocates anticipated—and, in other cases, has little to do with gains that Democrats do make.

Consider these results from the 2016 election. Estimates from the nonpartisan States of Change project indicated that purely based on underlying demographic change, the Democrats’ Pennsylvania margin should have increased modestly between 2012 and 2016. Of course, that didn’t happen. Instead, Hillary Clinton actually lost the state by 0.7 points.

A simple way to unpack this result is to look at contributions to the Democratic margin, or CDM, which can be calculated by multiplying an election’s proportion of voters in a given demographic by their Democratic margin in that election. The result can be compared across elections to see how demographic groups change in their contribution to the overall Democratic margin. 

Doing this for Pennsylvania in 2016 shows that shifts among white voters, driven by white non-college voters, produced a large negative contribution to Democrats’ margin in the state. That shift simply overwhelmed the modest positive CDM from rising diversity. This pushed the Democratic margin in the state toward the unexpected result of a Trump victory. The same pattern applied to Michigan and Wisconsin in 2016.

But, of course, these states are slow-growing and hardly at the forefront of demographic change in the country. Perhaps the story would be different in states, like Texas and Arizona, that had fairly sharp moves toward the Democrats in 2016 relative to 2012. Estimates by States of Change indicated that rising diversity alone would move these states one to two percentage points in the Democrats’ direction in 2016. But the actual shift in the election was far greater. That’s because the actual shift was driven, above all, by an increase in white support—chiefly white college-educated support—for Democrats in both states. Indeed, the shift in white voter preference in Texas was enough, by itself, to account for the Democrats’ improved margin in the state.

The 2020 election illustrates these points further. Take Arizona again, and Georgia, the Democrats’ two big breakthrough states in the election. In Arizona, a burgeoning Hispanic electorate was set to boost Democratic fortunes by about the same amount as in 2016. However, Arizona did not escape the nationwide pattern of Hispanic voters moving toward Trump. As a result, despite the underlying trend toward more eligible Latino voters and excellent turnout performance, the overall CDM of Arizona Hispanic voters actually went down by a little less than a point in 2020. On the other hand, the CDM of white voters, driven especially by white college-educated voters, improved by around 5 points—more than enough to account for the shift in Arizona to a Democratic advantage in 2020. In other words, it was educated whites, not Hispanics, who won Arizona for Biden. 

In Georgia, rising diversity was also set to improve Democratic fortunes. However, the most important part of the nonwhite vote, black voters, had a somewhat lower CDM in 2020, as well as slightly reduced voter share due to declining relative turnout. As a result, the CDM of black voters in 2020 probably declined slightly in the state. As in Arizona, the Democratic shift from a deficit in 2016 to an advantage in 2020 can be accounted for almost entirely by a sharp shift toward the Democrats among white voters, especially white college-educated voters. This shift was the major factor behind Democrats carrying the state.

There are four lessons here. First, while the effects of rising diversity do indeed favor the Democrats, these effects are fairly modest in any given election and can easily be overwhelmed by shifts in voter preference against the Democrats among unfavorable demographic groups, such as white non-college voters.

Second, even among favorable demographic groups, the electoral benefit to the Democrats can be completely neutralized by shifts against the Democrats within a demographic group. This was the case with the Hispanic vote in Arizona and many other states in 2020.

Third, in states where demographic change is rapid, it is easy to mistake shifts toward the Democrats in a given election as indicators of these underlying demographic changes. But as we saw for Arizona and Texas in 2016 and Arizona and Georgia in 2020 (there are many other examples), their pro-Democrat shifts were, in fact, driven by white voters.

Finally, the long-range effects of rising diversity are also an all-else-equal proposition. While cycle-by-cycle voter preference shifts can be volatile and even out over time, sometimes they result in a long-term shift against a party like the Democrats—think of the move of white non-college voters toward the Republicans in the 2000s. This can cancel or even swamp the pro-Democratic effects of demographic change over a lengthy period. 

In short, demographics set the playing field, but they are not destiny unless all else remains equal. And all else almost never remains equal. Therein lies a challenge for the Democrats that the simple fact of rising racial diversity cannot solve.