In the first week of November, tens of millions of Americans will vote for Donald Trump with a range of motivations. But two of those motivations have become a fascination of writers and analysts in recent weeks: whether Trump’s rise is the result of “racial anxiety” (e.g., a racist fear of Mexican, Muslims, blacks) or “economic anxiety” (e.g., financial angst among those without bright economic prospects).

Trump’s support is strongest among older men without a college degree, a once-dominant demographic that has been battered by globalization and technology in the last 40 years. This alone would make Trump’s rise seem like an economic phenomenon. But according to several analyses, the median Trump supporter makes an above-average wage and is far from indigent. And so, on Twitter, the ironic quip “Economic Anxiety!!” has become arch code for mocking the racist outbursts of Trumpians, a way of pointing out that their devotion to Trump has nothing to to do with their supposed financial plight.

But it’s really quite counterproductive to keep anti-minority sentiments and economic anxiety separate, for at least three reasons.

First, people tend to evaluate their status and happiness based on relative gains and losses, not absolute levels. Older non-college white men might not be desperately poor, but their economic fortunes are clearly in decline. They are watching their incomes stagnate, their children’s income fall behind, and their cultural status whither. Second, Trump is fanning their anger by scapegoating poor minorities, thus making many of his economic messages fundamentally racial.

Third, cultural anxiety and economic anxiety reinforce each other in ways that are invisible, yet obvious. For an analogy, consider the common cold. The rhinovirus can thrive in all sorts of climates. People get colds in the summer, spring, and fall, but they get sick more often in the winter. Why? Cold air cannot “give you” the common cold (an air temperature cannot conjure a virus), but the chain of causality is not complicated. Cold air weakens the immune system, allowing the virus to multiply in chillier nostrils. So, no, cold air does not, on its own, give people a virus, but it does create the conditions in which the virus can thrive.

Political science, like the natural sciences, has its share of nasty buggers that spread among people and infect the body politic. Trumpist racism toward Hispanics, Muslims, and blacks has bloomed in the last year into a potent political movement. Like the rhinovirus, racism has existed for a long time, but it also seems to thrive under certain economic conditions.

In 1995, the sociologist Lincoln Quillian published a widely cited paper on anti-immigrant and racial prejudice in Europe. He wrote that racism festers among the middle class majority (and not just among the poor) during periods of weak economic growth and a rising minority population. The majority, fearful of losing income and status, rallies around nationalism to protect what’s theirs.

Quillian’s model was prescient. Across Europe, anti-immigrant attitudes have increased among the working class in countries that have experienced slow growth, such as the U.K. before Brexit. In the U.S., too, resentment and anger grew among middle-class whites during the recession and weak recovery. In early 2011, with unemployment at 9 percent, the group with the darkest outlook on U.S growth was non-college white men. A weak economy, like a cold nose, is a cavity in which afflictions breed.

Perhaps this much is obvious: Recessions breed tribalism. But here’s where the research on prejudice and economics gets even more interesting—and, perhaps, more depressing for America’s future: Redistribution might breed tribalism, too. Just thinking about splitting the economic pie can make people more jealous about keeping slices out of the hands of other groups.

In their 2004 book Fighting Poverty in the US and Europe, the economists Alberto Alesina and Edward Glaeser identified a latent paradox of the welfare state. On the one hand, redistribution policies are often born out of trust in a culturally homogenous (i.e.: all white) country. On the other hand, when these places became more diverse, the welfare state can ironically lead to more negative attitudes toward immigrants. Foreigners receiving welfare are seen as an economic threat, or as competition for scarce economic resources that the old majority was hoping to keep for themselves. This somewhat describes the experience of the Twin Cities, for example, an area where quasi-socialist tax and housing policies passed several decades ago in periods of white homogeneity have been eroded as the metro area became more diverse.

In the debate over economic anxiety and Trump, some writers have asked: If economics is a factor in racism’s political ascent, how could one explain the popularity of George Wallace in the richer time, the late 1960s? The simple economic anxiety theory—recessions create racism—cannot explain Wallace’s rise. But what about the second theory, that redistribution politics can fan the flames of racism?

In his book Why Americans Hate Welfare, the Princeton professor of politics Martin Gilens wrote that American attitudes toward the poor are a paradox. The middle class supports helping the poor, in theory, but does not support welfare, in practice, and occasionally even supports candidates who would immiserate poor Americans. Why? Often it’s because white working class Americans equate poverty with blacks, whom they are more likely to consider lazy outsiders who are undeserving of government assistance.

Gilens found that national magazines’ and television channels’ portrayals of the poor took a sharp turn in the 1960s. Between 1950 and 1964, most stories and pictures about the poor featured white families. But in the late 1960s and early 1970s, during the rollout of the Great Society, blacks featured in at least 70 percent of media stories about poverty and welfare. Over the course of the decade, the media portrayal of welfare went from “a policy to help the poor” to “a policy to help blacks.” In 1968, in the middle of this period, the overtly racist candidate George Wallace won 13 percent of the vote and five southern states.

The moral of this story is not that redistribution is bad, or that diversity is doomed. In fact, the opposite is true. Diverse areas that share wealth and land tend to see the best income growth across generations. It’s the U.S. metros that refuse to share real estate and income between rich, middle class, and poor that impoverish themselves in the long run.

Research by Raj Chetty and Nathaniel Hendren on intergenerational mobility throughout the U.S. found that segregated neighborhoods in the southeast were among the worst areas for social mobility in the country—not only for blacks, but also for poor whites.

Why? One reasonable explanation is that in areas where richer whites don’t trust poor black families, there is less support for the very policies that would raise social mobility among the poor (including poor whites!), like business investment in low income areas, support for high quality schools, support for integrated housing, and support for public services in rich and poor parts of the metro area. And so, racial selfishness can ultimately punish everyone. Indeed, when Jonathan Rothwell, a Gallup researcher, isolated factors to try to explain nationalist support for Trump, he concluded that segregated areas with poor social mobility were particularly likely to support Trump.

The “economic anxiety” joke has its place; racism is truly at the heart of Trump’s appeal. But racism’s latent virus blooms into fullest view when majority groups fear the scarcity of their winnings. How fitting that when Trump is not fanning his supporters’ racism, he is reminding them that America’s winnings are scarce.

Trump is doomed. Trumpism is not. Sooner or later, the joke will be on everyone.