Statistical Modeling, Causal Inference, and Social Science: Larry Bartels on income, voting, and the economy

Larry Bartels on income, voting, and the economy

Larry Bartels spoke in our seminar the other day and talked about this paper on Democrats, Republicans, and the economy. It started with this graph, which showed that incomes have grown faster under Democratic presidents, especially on the low end of the scale:

He looked at it in a number of ways, and the evidence seemed convincing that, at least in the short term, the Democrats were better than Republicans for the economy. This is consistent with Democrats' general policies of lowering unemployment, as compared to Republicans lowering inflation, and, by comparing first-term to second-term presidents, he found that the result couldn't simply be explained as a rebound or alternation pattern.

But then, he asked, why have the Republicans won so many elections? Why aren't the Democrats consistently dominating? Non-economic issues are part of the story, of course, but lots of evidence shows the economy to be a key concern for voters, so it's still hard to see how, with a pattern such as shown above, the Republicans could keep winning.

Larry had two explanations.

Ian

Disclaimer: this is all based on a VERY quick reading of the paper linked to under Bartels' homepage.

A couple of issues: there doesn't seem to be much if any consideration for the changing policies under the different presidents. Kennedy, for instance, adopted tax cut policies that might be more in line with what is considered traditionally Republican today. And Carter finally got serious about fighting inflation (again, something apparently ceded here to Republicans in the "One side is anti-poverty, the other is anti-inflation" divide) with the elevation of Volker to the Chairman of the Fed. (For that matter, the role of the Fed Reserve Chairman is left absent, with neither Volker, Greenspan, or their predecessors included. This would seem relevant to me in a sort of principle-agent structure since Republicans might have been systematically more inclined to allow the Fed to be the defacto leaders on some policiesl or the Democrats systematically less likely, etc.) This paper seems to simply assume that all democrat administrations acted alike, as did all Republican ones. A poor assumption, if that is the case (see disclaimer -- VERY quick read) since Kennedy had policies that seemed out of line with later Dems (cuts, defense spending, etc), since Johnson inherited a certain amount of his policies (should he be considered a less typical Dem like Kennedy, or did he do enough to individuate himself later on to be considered one or two terms?), since Carter as admitted by the study doesn't actually preside over a reduction in inequality, since Nixon ended reliance on the gold standard (a considerable economic shift that at least might warrant some form of extra consideration), and so on. Of the 11 presidents since 1948, several seem to have particular considerations that argue for not assuming uniformity of behavior. And 3 or 4 out of 11 strikes me as significant.

Ian

Disclaimer: this is all based on a VERY quick reading of the paper linked to under Bartels' homepage.

Ian

Count me as agreeing that it might be useful (actually more useful than this current study) to look at actual policies.

Disclaimer: this is all based on a VERY quick reading of the paper linked to under Bartels' homepage.

I think it's the general categorization of Dem v. Rep as starkly different from each other and entirely similar in their group behaviour that strikes me as the most limiting, and ultimately the most crippling. Kennedy, for instance, adopted tax cut policies that might be more in line with what is considered traditionally Republican today. And Carter finally got serious about fighting inflation (again, something apparently ceded here to Republicans in the "One side is anti-poverty, the other is anti-inflation" divide) with the elevation of Volker to the Chairman of the Fed. (For that matter, the role of the Fed Reserve Chairman is left absent, with neither Volker, Greenspan, or their predecessors included. This would seem relevant to me in a sort of principle-agent structure since Republicans might have been systematically more inclined to allow the Fed to be the defacto leaders on some policiesl or the Democrats systematically less likely, etc.) This paper seems to simply assume that all democrat administrations acted alike, as did all Republican ones. A poor assumption, if that is the case (see disclaimer -- VERY quick read) since Kennedy had policies that seemed out of line with later Dems (cuts, defense spending, etc), since Johnson inherited a certain amount of his policies (should he be considered a less typical Dem like Kennedy, or did he do enough to individuate himself later on to be considered one or two terms?), since Carter as admitted by the study doesn't actually preside over a reduction in inequality, since Nixon ended reliance on the gold standard (a considerable economic shift that at least might warrant some form of extra consideration), and so on. Of the 11 presidents since 1948, several seem to have particular considerations that argue for not assuming uniformity of behavior. And 3 or 4 out of 11 strikes me as significant.

TheOneEyedMan

There a cohort problem here. There is some income mobility (at least some) in the US, so we really have to see if the same people are doing better under different sorts of administrations.

It could easily be that different sorts of people within each starting income level have different outcomes after a switch in administration.

Koray Can

Disclaimer: I didn't read the paper. However, if it is true that only years 2 & 3 count towards what a government wants to do in general as opposed to what it needs to do in order to do well at the elections, then it doesn't really matter one way or another.

Andrew Gelman

Ian,

I agree that each presidential term is idiosyncratic. Nonetheless, the patterns are pretty striking, and they are consistent with the consistent pattern over the decades of Republicans being more pro-business and the Democrats being more pro-labor. As noted, the analysis is inherently about short-term effects, but when you put it all together it's a compelling average pattern.

OneEyedMan,

I agree about the cohort problem. This would be harder to study using cross-sectional data but would certainly be another interesting thing to look at.

TCO

1. There is a lot of evidence that voters act on more then just the economy.
2. Economy is not purely a function of president.
3. There is a tiny data set to get any kind of trend out of (how voters act, how economy does under different parties, etc.)