Something certainly feels right about the thesis of David Brooks’ latest NYT column. I almost found myself agreeing with it. But then I thought about it for a little while, and concluded that Brooks badly misses the mark. He writes:

You can see why, in the disrupted landscape depicted in [Richard Linklater’s film, Boyhood], people would form the sort of partisan attachments that are common today. Today, partisanship for many people is not about which party has the better policies, as it was, say, in the days of Eisenhower and Kennedy. It’s not even about which party has the better philosophy, as it was in the Reagan era. These days, partisanship is often totalistic. People often use partisan identity to fill the void left when their other attachments wither away — religious, ethnic, communal and familial.

Last week my colleague Thomas Edsall quoted a political scientist, Alex Theodoridis, who noted this phenomenon: “Partisanship for many Americans today takes the form of a visceral, even subconscious, attachment to a party group. Our party becomes a part of our self-concept in deep and meaningful ways.”

When politics is used as a cure for spiritual and social loneliness, it’s harder to win people over with policy or philosophical arguments. Everything is shaped on a deeper level, through the parables, fables and myths that our most fundamental groups use to define themselves.

This is a neat spin on a hoary conservative trope and perennial David Brooks theme: deracination leads to nihilism.

In this version, the loss of traditional forms of “thick” collective identity leads to a sense of emptiness, which people then seek to expiate by digging into partisan identity. It’s true, as Alex Theodoridis says, that Americans have become more partisan, but it’s not so clear to me that this is due to people compensating for a loss of other forms of group identity.

The “group theory of democracy” (which I discuss at some length in this post) suggests that partisan attachments are driven by other, prior group identities. If the group theory of democracy is correct, and I think it is, you shouldn’t expect the weakening of other group identities to harden partisan identity. You should expect the reverse.

The group theory says that the group membership with which you most closely personally identify pretty well determines your partisan affiliation. Are you a member of a union? Odds are, you’re a Democrat. Are you an Evangelical Christian? You’re probably Republican. Are you African-American? Democrat. Hispanic? Democrat. Got lots of graduate education? Democrat. A white male? Republican. Small town white person? Republican. Etc. The fact that identity is intersectional, that we belong to and identify with multiple groups simultaneously, complicates this, but not very much.

This wouldn’t work as well it does if our “other attachments” are, as Brooks says, withering away.

Michigan State political scientist (and Niskanen senior fellow) Matt Grossman has observed that there’s an important difference between the Republican and Democratic parties. The Democrats are a diverse coalition of interest groups each pushing for fairly concrete policy changes, while the Republicans are a relatively homogeneous group unified around an ideological vision.

Trump has shown how easily the unifying “small government” ideology of a Republican Party that is 75% white and Christian can morph into an ethno-cultural group identity ideology. The fact that it’s been about 100 years since the foreign-born share of the population was as high as it is now, and the fact that white Christians are no longer a majority, has strengthened white Christian group identity by threatening its centrality to American national identity. And thanks to the knock-on effects of the 1964 Civil Rights act and de-unionization, which has broken the link between working class whites and the Democratic Party, millions of white Christians have sorted out of the Democratic Party and into the GOP.

This has made Republican partisan identity a better and better proxy for white Christian identity. But because white Christians are historically America’s politically and culturally dominant ethno-cultural block, and remain the biggest one, even if they no longer constitute a majority, members of this group prefer to cast their ethnic and religious group interests as the national interest, and to think of themselves not as one group among many, but as the real Americans, alone fully vested with moral citizenship.

That’s how hyperventilating jingoism has become a marker of white Christian identity politics, which is centered on countering the multicultural threat to white, Christian sociopolitical dominance. Trump latched onto and encouraged this dynamic to winning effect, which is turning what was once the coded ethno-cultural subtext of the GOP’s unifying ideology into the more or less explicit main event. This, in turn, has activated the group identities of those most viscerally threatened by an ethno-nationalist Republican agenda—African-Americans, Muslims, well-educated women, Hispanics, Asians, and others who identify with recent immigrant communities—cementing their loyalty to the Democratic Party.

Whatever this is, it’s not the withering away of ethnic and religious attachments. On the contrary, it seems to me that the hardening of partisan identity is a predictable side-effect of the increased salience and significance of some of the group memberships Brooks suggests have weakened in importance, and of the long-term sorting of Southern and working-class whites into the GOP.

So Brooks’ thesis is wrong. Partisanship has become harder and hotter because it has become more cleanly expressive of deeper ethno-cultural identities, not because we’ve lost them and have latched onto partisanship as a substitute.

Will Wilkinson is Vice President for Policy at the Niskanen Center