Krugman:

Mr. Obama’s white support continues to be concentrated among the highly educated; there was little in Tuesday’s results to suggest that his problems with working-class whites have significantly diminished. ... Discussions of how and why Mr. Obama’s support narrowed over time have a Rashomon-like quality: different observers see very different truths.

Ah yes, he asks how and why without asking whether or if. Perfect.

The problem here is with Krugman's assumption -- that Barack Obama's support from whites is whithering.

On the surface, Krugman's assumption seems like it might be correct -- in February, Obama won 42% of the white vote and from March to May he won just 38%.

But Krugman doesn't take regional differences into account, nor does he take the influx of pro-McCain voters into the primary into account. After adjusting for those voters and eliminating home states (IL and NY) as well as western states (AZ, CA, and UT), none of which had primaries in March-May, the "withering" shrinks. Taking these factors into account, Obama won 40% of the white vote in February and 39% from March to May.

Narrowing it down further, in Ohio and Mississippi, Obama won 35% of the white vote (40% if you include Texas). In Pennsylvania, he won 37%. And in Indiana and North Carolina, he won 41%.

So this "withering" narrative is manifestly false (especially with respect to Obama's position in national polls circa November/December 2007).

What is true is that Barack Obama does not receive as much support from whites as does Hillary Clinton. It turns out that white women are the biggest reason for this. In North Carolina, she won 65% of white women. In Mississippi she won 71%. In Ohio, 67%. In Pennsylvania, 68%. Even in Indiana, she won 61%. Typically, white women outnumber white men by about 30% in Democratic primaries.

So Obama's support from whites is not withering, and it does not suggest the fundamental weakness that Krugman seems to believe.

But in the end, the fundamental mistake Krugman and others make is obsessing about the Democratic primary as if it were a predictor of how general election votes will break down.

They need look no further than the most recent Pew Survey to get a handle on whether Obama has got "white problem."

As you can see, it doesn't exist. He does exactly the same as Hillary Clinton relative to John McCain -- 45% to 50%.

To put that in perspective, Bill Clinton won 43% of the white vote in 1996. Kerry won 41% in 2004. Probably, Obama would win the presidency with 42% of the white vote in 2008. So 45% is a damn good starting spot.

When you look at that table you do see one big problem jump out at you though: black voters do not support Hillary Clinton nearly as strongly as they support Barack Obama. The gap in this poll -- 24% -- could easily mean 2 or 3 points nationally.

To the extent that we see a "race-based" problem with one of the Democratic candidates, it isn't Barack Obama's problem. The sooner Paul Krugman and people who agree with him wake up to these realities, the better off we'll be.