Thinking Twice

In a blog post with which I think I mostly agree, Scott Sumner writes something that I'm not yet ready to agree with. (And hey - for a change it's not about NGDP growth level targeting!) Sumner echoes some claims made elsewhere by Kevin Williamson and Matthew Yglesias, to which Sumner himself links and quotes liberally. In doing so, however, my guess is that he may have overreached.

Here's the offending quote:
Like the Occupy Wall Street movement, Sanders's [sic] supporters are mostly white.
In a literal sense, this may be true, but that sense is not very interesting. In a majority-white country like the United States, virtually every popular politician is bound to have a set of followers that skews white. That's about as interesting as discovering that the Government of Uruguay is "run by Latinos." If this is the sense in which Sumner meant that Sanders' "supporters are mostly white," then the statement seems regrettable in that it suggests that this is not true of other politicians.

A more charitable reading of Sumner is that he meant that supporters of Bernie Sanders skew more white than supporters of, say, "the average" politician. In other words, it could be that Sumner is suggesting that Sanders' supporters are comprised of a larger majority of whites than the majority of whites who support the average politician.

That statement is certainly more rhetorically defensible, but unfortunately I can't find any empirical facts that support this claim. The piece by Williamson mentions in passing that Vermont, the state Sanders represents, is the "second-whitest" state in the United States. If this is what caused Sumner to suggest that Sanders' supporters are "mostly white," then so be it. However, it simply does not follow that, because a presidential candidate comes from a majority-white state, therefore that candidate's supporters are disproportionately white.

As much as I'd like to agree with Scott Sumner on this one, I don't see any good evidence suggesting that Bernie Sanders' supporters are any whiter than anyone else's supporters. Neither Sumner, Williamson, nor Yglesias write anything that would support such a claim.

I did some poking around, and according to recent Gallup polling, Sanders' supporters are more white and male at least compared to supporters of Hillary Clinton. So there is at least some evidence in Sumner's favor here, but he does not supply a link to buttress his claim.

Conclusion: Sumner's point stands, at least for now.