Ebonya pointed me to this paper by Jacob Vigdor which addresses a similar question to that of Huber and Stanig’s paper, “Why do the poor support right-wing parties?”. A natural answer to this sort of question is that people have economic ideologies–basically, you can support conservative economic policies if you believe they are good for the economy, or for moral reasons (for example, feeling that the flat tax is fairest). Huber and Stanig’s paper was interesting partly because it addressed variation among countries in these voting patterns.

Anyway, Vigdor has a slightly different perspective. Here’s his abstract:

Why do voters at the lower end of the socioeconomic spectrum support political candidates who generally disfavor redistributive policies? Existing explanations often presume that voters are explicitly acting in opposition to their economic self-interest, or that they hold persistently optimistic expectations regarding the probability of moving into the upper ranks of the income distribution. This paper provides an alternative economic explanation. When voters evaluate their well-being by making relative utility comparisons, support for redistribution depends not only on absolute income but on one’s status relative to a reference group. When reference groups are defined geographically, support depends on exposure to higher-income neighbors. The predictions of the model are supported by empirical evidence drawn from county-level election returns in 1980 and 2000, and by individual-level polling data following the 2000 election.

And now my thoughts:

1. First off, lots of research has found that people vote based on what they think is good for the country, not their personal pocketbook. In fact, it’s actually rational to vote for what you think is good for the country, and not rational to vote based on your perceived direct economic self-interest. (See our paper in Rationality and Society). The questions asked by Vigdor’s paper remain, but I’d like to frame them in terms of rationality rather than selfishness.

2. Regarding the quote at the top of page 1: Actually, it’s the rich who vote Republican and the lower income categories who vote for the Democrats. Given that, the puzzle stated in the introduction to the paper is much less puzzling.

This comes up again at the beginning of Section 6: I wonder if the author of this paper is trying to explain a phenomenon that’s not actually happening!

3. Regarding Figure 2: the relation between county income and Republican vote share shows strikingly different patterns in different states; see Figures 6,7,8,9 and the accompanying discussion in Section 3.4 of our red-blue paper. A lot is hidden by the aggregate pattern.

4. Following up on my comment 1 above, Vigdor refers to rich people voting “altruistically” if they vote to increase redistribution. I don’t like this characterization, since if you believe that redistribution is bad for the economy, it is more altruistic to oppose redistribution.

5. The paper discusses comparison to a reference group, which makes me think of geographic proximity or social networks, but then the analysis is based on people nearby on the income distribution (see page 15). I don’t see that this is appropriate, unless I’m missing something.

6. Also, on page 15 it says, “In the era of legal segregation, the reference groups of relatively poor white voters [in the south] may have consisted only of other whites.” Is this really true? I’d always heard the opposite, that whites in the south generally defined their status relative to blacks.

In any case, the ideas of this paper are interesting. I’d like to see them developed using geographic and social-network measures of proximity, and also explaining variations in the patterns of income and voting among states. (Income is a strong predictor of Republican voting in poor states, but not in rich states.)