Donald Trump talked about his daughter in sexually explicit terms on a nationally syndicated radio program. In a leaked recording, taken shortly into his third marriage, Trump bragged that he could accost women without consequence as a result of his social status—and in fact, several women came forward over the course of his campaign detailing how he allegedly violated them. If an African American aspired towards the presidency with this kind of sexual baggage, his candidacy would have been dead on arrival.

Similarly, a black candidate whose business dealings were defined by accusations of nepotism, shipping jobs overseas, exploiting undocumented workers, stiffing U.S. contractors and exploiting bankruptcy and tax laws to evade financial and civic obligations would not be viewed as the kind of leader America needs. Indeed, an African American who seemed to lack basic knowledge about the major issues facing the country and possessed no experience in government would not even be on the radar as a serious candidate. He would never win his party’s nomination, let alone the general election.

Ta-Nehisi Coates makes this case in his latest Atlantic cover story, “The First White President,” to demonstrate that race matters a lot more in American politics than most pundits and politicians seem willing to acknowledge.

He goes on to highlight how Americans across the socio-economic spectrum often point to the plight of the “white working class”—-a plight that white elites are themselves are often responsible for creating—in justifying policies that disproportionately harm blacks and other minority groups. Meanwhile, social problems like drug addiction, which have long plagued minority communities, only seem to grow worthy of an urgent and compassionate response when “working class” whites become affected.

Coates’s exploration of these topics is powerful, as is his condemnation of the President’s inadequate response to the recent violence in Charlottesville. However, a severe paradox emerges with regards to his central thesis:

Coates complains that it is reductive and misleading to “blame” Clinton’s loss on the “white working class” (as many have done) given that Trump decisively won among all whites—across the income and education spectrum, across gender and geographic lines, etc. True enough. But then Coates puts forward an alternative frame which turns out to be no less problematic than the one he is critiquing: Trump was elected primarily because of racial resentment, and he maintains his hold on power by playing to Americans’ latent sympathies with white supremacists. 

As an African American and a Muslim, like Coates I often find myself disturbed by Trump’s rhetoric and some of his policies. However, as a social researcher I have also been consistently troubled by the near-total lack of engagement among pundits and scholars with the pretty robust data confounding the ‘white supremacy’ theory of Trump’s success. For instance:

What Was the 2016 Election Really About?

In Coates’ depiction, white Americans rose up in 2016 to dismantle the legacy of America’s first black president, Barack Obama—-and to reclaim what is rightfully “theirs.” 

Let’s set aside the fact that many of the white voters who proved most decisive for Trump’s election actually voted for Obama in 2008 and 2012. It turns out that Trump did not mobilize or energize whites towards the ballot box either: their participation rate was roughly equivalent to 2012 (and lower than in 2008). In fact, whites actually made up a smaller share of the electorate than they did in previous cycles, while Hispanics, Asians and racial “others” comprised larger shares than they had in 2012 or 2008. 

This wouldn’t, in itself, be devastating for the “white supremacy” hypothesis had Trump won some kind of unprecedented share of the whites who did turn out to vote. He did not. Trump didn’t even exceed Romney’s 2012 numbers with whites overall. However, he did outperform his predecessor with blacks and Hispanics.

These are very inconvenient truths: If Trump was the white supremacy candidate, why was such a pivotal component of his base drawn from those who previously voted for Obama (often twice)? Why didn’t far greater numbers of whites turn out for Trump than Romney? Why did fewer whites vote for the Republican in this cycle than they did in the last? Conversely, why did Trump win a larger share of blacks than any Republican since 2004? Why did he outperform Romney among Asians and Hispanics as well?

To understand these realities it is important to look at the larger picture: there has been a broad and consistent trend away from Democrats, cutting across virtually every imaginable demographic category, for the better part of a decade. From this perspective, 2016 was not so much an outlier as the culmination of long-running trends. In other words, the most recent election might have been more about Americans rejecting the Democratic Party and their candidate rather than embracing the Republicans or Trump.

This is where arguments like Mark Lilla’s become important (despite their limitations):

While it is true, as Coates and others have asserted, that “all politics are identity politics,” it is also clear that the Democrats’ current approach is not working: the party cannot seize control of the government without support from more whites—particularly from those outside of metropolitan and coastal areas. And if Democrats cannot be voted into office in sufficient numbers, they will remain unable to realize policies they believe will benefit racial minorities, sexual and gender minorities, etc. To the extent that his opposition refuses come to terms with this predicament, Trump will only be further empowered.

The Generalizability Problem

Was race a critical factor in the 2016 election? Undoubtedly. But this can be said of virtually every cycle in American history—including the 2008 and 2012 races which resulted in the election and re-election of Barack Obama. The specific ways race influences politics are more complex and unpredictable than the analyses of Coates and others seem to suggest—even among those who voted for Trump.

In fact, looking at whites’ stagnant turnout in 2016, and Republicans’ lowered vote share among whites as compared to 2012, it could be that Trump’s inflammatory rhetoric was, on balance, a drag on his candidacy among white voters rather than the key to his electoral success. In order to explore this possibility, researchers would have to turn their attention away from the motivations of whites who did vote for Trump and towards the large majority of eligible non-Hispanic white voters who did not.

Indeed, even if we assumed for the sake of argument that 100 percent of whites who voted for Trump were motivated primarily by racial resentment—-a claim which has no factual basis, to be clear—-one still would not be able to make sweeping generalizations about white people from those voters. Nearly as many chose to abstain (36 percent) as those who voted for Trump (37 percent)—a fact which should raise grave concerns about participation bias right off the bat. An additional 27 percent explicitly rejected the man by casting their votes for other candidates (Clinton, Johnson, Stein). Social scientists and journalists would not tolerate any other group being maligned in this way on the basis of such a dubiously representative sample. The double-standard here is perplexing.

Here’s one thing I do understand: if the left wants to spend the next four years denigrating whites as irredeemable racists because they didn’t vote for Hillary Clinton in sufficient numbers in 2016—-and then come, hat-in-hand, asking for a larger share of their votes in 2020—-they should count on Trump being a two-term president.

Musa al-Gharbi is a Paul F. Lazarsfeld Fellow in Sociology at Columbia University. Readers can connect to his research and social media via his website: Fiat Sophia.

***Editor’s Note: This piece has been updated from its original version.