In the previous post, I examined some details of voter turnout nationally and in some closely-contested states, and concluded that, although Trump won the election due to the strength of support from white voters, he did not actually bring in new voters to any great extent, and he actually received fewer votes overall, nationally, than Mitt Romney did in 2012. The decrease in turnout for Clinton, especially in key states, played a far more significant role in this election.
To get that white support, however, Trump used rhetoric that alienated a lot of voters, and this may cost the GOP in future elections. The next election may not feature the same depressed turnout as this one did. As mentioned previously, swing state results were very close, and nationally, while final results are not completely tallied, Clinton is ahead in the popular vote count by nearly 1.5 million votes.
In the next presidential election, there will be fewer white voters and more minorities. Trump showed hostility toward minorities in this election, making the Republican Party unattractive to this growing share of the electorate. Today we’ll look at the votes of the fastest-growing segment of the electorate, Latinos.
Support for Trump among Latinos started low, and stayed low
Trump opened his campaign by insulting Mexican immigrants, and from the beginning, Latinos viewed his candidacy very unfavorably. In the first public opinion survey of Latino voters after Trump announced his candidacy, just 16 percent of Latino voters said they would vote for Trump if he were matched against Clinton. In the following year, Latinos continued to have a very unfavorable view of Trump. By Election Day, just 18 percent of Latinos voted for Trump, according to Latino Decisions, a firm specializing in polling of Latinos.
Latino Decision’s estimate of 18 percent support for Trump among Latinos nationwide comes from an election eve poll the group conducted, consisting of 5,600 interviews of Latinos with strong turnout histories who said they had either already voted in early voting or that they would definitely vote in the election.
The national exit polls fall short with subgroup results
Why rely on this number, instead of the 29 percent reported in the national exit polls? Gary Segura and Matt Barreto, of Latino Decisions offer a long critique of the national exit poll numbers for Latinos. For a number of reasons, the national exit polls are not very good at accurately estimating the voting results of subpopulations including Latinos. You can read their critique here.
They bolster their argument citing other major pre-election polls of Latinos, which obtained similar results. For example, a Washington Post-Univision poll conducted at the end of October found Trump had the support of 19 percent of Latinos. A NALEO-Telemundo tracking poll found 14 percent who said they would vote for Trump (with support running between 14 and 18 percent in the several weeks running up to the election). Other polls found similar results. It is implausible, Segura and Barreto argue, for all of these polls to be off by 10 percent or more. (Despite a media narrative to the contrary, pre-election polls this year were not off by much—for the popular vote, only about 2 percent, which is better than the 3.2 percent error in 2012.)
More evidence that the Edison Research exit polls were off for the Latino vote comes from actual results from precincts or counties with large Latino populations, results were similar to the election eve poll of Latino Decisions. For example, in exit polls in six counties in Texas with high percentages of Latinos, Latino support for Trump ranged from 19 to 24 percent, with turnout being 2 to 10 percent higher than 2012 turnout. In polls of Latino voters in six precincts in Miami Dade, Clinton received the support of 49 to 60 percent of Latinos. These are precincts where Cubans traditional gave their vote to Republicans. Clinton’s performance among Latinos in these precincts bested Obama’s by 13 to 25 points, and Latino turnout was 6 to 17 percent above 2012 turnout. In Wisconsin precincts with heavy concentrations of Latinos, Clinton won 83 to 90 percent of the Latino vote.
Trump may have brought out new voters—to vote against him
In my previous post, I said it appears that Trump did not bring out a lot of new voters, as he said he would. I should add a caveat: 20 percent of Latino voters were new voters (verses 10 percent among all voters), in large part motivated by opposition to Trump.
In sum, in appealing to white voters with divisive rhetoric, Trump succeeded in actually underperforming Mitt Romney with the fastest growing part of the electorate—Latinos. This may come back to haunt the GOP in future elections.