His latest column is yet another broadside in the whole "does race explain the Republican realignment" argument, and as you might expect, it combines convincing specific examples of Republicans playing the race card in the South with totally unconvincing macro-level analysis. For instance, there's this:

... everyone knows that white men have turned away from the Democrats over God, guns, national security and so on. But what everyone knows isn’t true once you exclude the South from the picture. As the political scientist Larry Bartels points out, in the 1952 presidential election 40 percent of non-Southern white men voted Democratic; in 2004, that figure was virtually unchanged, at 39 percent.



First, as Matt has pointed out, the fact that the bulk of the white-male shift occurred in the South doesn't mean that white males were simply changing their party allegiance in response to GOP race-baiting. Most white Southerners were conservatives - on God, gays and guns, among many other issues - who happened to vote for the more liberal party in the '30s and '40s because it was the segregationist party, and once that issue receded, and the Republicans moved rightward, you would have expected them to shift to the more conservative party even in the absence of dog-whistle politics.

More importantly for the sake of this example, 1952 is a really poor baseline to use for comparisons to present-day politics, since it was an exceptional year - a Republican landslide in a Democratic era, created by Eisenhower's celebrity and ostentatious moderation, Truman's unpopularity and Stevenson's mediocrity as a candidate. Ike took 55 percent of the vote to Stevenson's 44 percent, meaning that the GOP vote was much higher than the FDR-to-LBJ norm in almost every demographic category - and for Bush to match Eisenhower's share of the non-Southern white-male vote fifty years later while winning only 51 percent of the vote to Kerry's 48 suggests that conservative have made gains between then and now in that demographic, rather than just treading water as Krugman suggests.

Moreover, even if the Republicans had merely tread water it would still be an impressive achievement, given that a rightward shift - all other things being equal, which they weren't - would have been expected to produce a 1964-style result, in which the GOP consolidated the South and lost ground everywhere else. Arthur Schlesinger famously announced that the results of '64 proved that "if the parties were realigned on an ideological basis ... the Democrats would win every election and the Republicans would lose every election." It was an entirely plausible contention at the time, and Krugman's "race explains everything" narrative doesn't explain why he was proven wrong.

Then there's this:

Ronald Reagan was among the “some” who tried to benefit from racial polarization ... True, he never used explicit racial rhetoric. Neither did Richard Nixon. As Thomas and Mary Edsall put it in their classic 1991 book, “Chain Reaction: The impact of race, rights and taxes on American politics,” “Reagan paralleled Nixon’s success in constructing a politics and a strategy of governing that attacked policies targeted toward blacks and other minorities without reference to race — a conservative politics that had the effect of polarizing the electorate along racial lines.”



This is one way to put it. Another way to put it is that Reagan paralleled Nixon's success in constructing a strategy of governing that attacked policies targeted toward minorities that were obviously failing - more obviously by the 1980s, certainly, than in 1968 - and deserved to be attacked. But to acknowledge this part of the story would be to get in the way of Krugman's conclusion:

Why does this history matter now? Because it tells why the vision of a permanent conservative majority, so widely accepted a few years ago, is wrong.

The point is that we have become a more diverse and less racist country over time. The “macaca” incident, in which Senator George Allen’s use of a racial insult led to his election defeat, epitomized the way in which America has changed for the better.

And because conservative ascendancy has depended so crucially on the racial backlash — a close look at voting data shows that religion and “values” issues have been far less important — I believe that the declining power of that backlash changes everything.



Memo to the Democrats: Please, please listen to this man. Convince yourself that the Republicans' current difficulties have everything to do with the waning of racism, and nothing to do with conservative victories on crime, welfare, and taxes and the declining salience of those issues in the public consciousness (not to mention the party's Bush-induced loss of foreign-policy credibility). Persuade yourselves that the liberalism of the 1960s and 1970s has absolutely nothing apologize for. I can think of no surer guarantee of a GOP comeback.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.