I’m not sure what represents a greater threat to Sean Trende’s “missing white voter” hypothesis for the GOP’s electoral problems: the misappropriation of his work by conservative stand-patters and double-downers who conveniently ignore Trende’s warnings about what it might take to appeal to these “missing” voters, or the chip-chip-chipping away at his data and conclusions by progressive analysts.

I noted earlier this week that TNR’s Nate Cohn had challenged Trende’s conclusions based on both regional and generational disparities in the GOP’s share of the white vote (i.e., getting a higher percentage of old white southerners does not cut much electoral ice over time). Now come the formidable Alan Abramowitz and Ruy Teixeira with an analysis at Larry Sabato’s Crystal Ball that questions Trende’s calculations:

Trende’s claim that Republicans have increased their performance among white voters is based on his calculation of a statistic known as the PVI, or Partisan Voting Index, for white voters. Essentially, this statistic is used to compare the political preferences of a given group to the electorate as a whole. The PVI for white voters compares the Democratic share of the white vote with the Democratic share of the vote in the overall electorate…..

Over time…the PVI for white voters has become increasingly negative, with an especially dramatic decline since 1992. There is no question that in comparison with the overall electorate, white voters have become more Republican over time. But the interpretation of this result is not as straightforward as Trende suggests. That is because the PVI for white voters reflects both the Democratic margin among white voters and the size of the nonwhite electorate.

In fact, the main reason that the gap between the Democratic margin in the overall electorate and the Democratic margin among white voters has increased over time is not because whites have become more Republican but because nonwhites, who are overwhelmingly Democratic, now make up a larger share of the overall electorate. As just one example, the PVI of the white vote in 2012 (-24) was far more negative than it was in 1988 (-13). Yet Democratic margins among both whites and nonwhites were essentially the same in each election. The real change: Nonwhites were just 15% of voters in 1988 compared to 28% in 2012. In other words, the rapid growth of the very Democratic nonwhite share of the electorate makes it seem like white voters are becoming more Republican than they actually are.

So the growing size of a heavily Democratic nonwhite vote has increased the racial polarization of voting above and beyond any actual Republican trend among white voters.

More generally, it’s another way of saying that a GOP strategy based on winning ever-more-historically-anomalous levels of support in a shrinking portion of the electorate doesn’t look much like a winner.

Ed Kilgore

Ed Kilgore is a political columnist for New York and managing editor at the Democratic Strategist website. He was a contributing writer at the Washington Monthly from January 2012 until November 2015, and was the principal contributor to the Political Animal blog.