Niall Ferguson vs. political science

by Andrew Gelman on September 13, 2012 · 10 comments

in Campaigns and elections,Media

The Harvard-affiliated historian writes:

It’s a paradox. The economy is in the doldrums. Yet the incumbent is ahead in the polls. According to a huge body of research by political scientists, this is not supposed to happen. On the other side of the Atlantic, it hardly ever does. But in America today, the law of political gravity has been suspended.

Data (from Doug Hibbs):

hibbsnew.png

Take a look at 2% growth on the above chart. Then get back to me. Sometimes the incumbent party won with 2% growth, other times they lost. It’s “hardly a law of political gravity” that the incumbent will lose.

To be fair, Doug Hibbs recently wrote that his model gives Obama a forecast vote share of 47%. Hibbs also writes that his standard error is 2.2%, so an Obama win is hardly out of the question “even if my [Hibbs’s] guess of a 1-2 percent average per capita real income growth over the last 4 months of the term turns out to be accurate . . .”

So I see no basis for Ferguson’s claim that “according to a huge body of research by political scientists, this is not supposed to happen.” A more accurate statement is that voters appear to respond to recent changes in the economy (hence Obama not being hurt by the full four years of recession), 2% growth is not historically so bad, and, sure, maybe some of the discrepancy between model and polls is coming from the campaign and personal characteristics.

But to say “it’s about personal likability” . . . no way. This is the same sort of self-deluded excuse that liberals were giving when Ronald Reagan was winning elections, and it doesn’t look any better when it’s coming from a conservative.

Romney may yet win. If he loses, I recommend to Ferguson move to the reasonable backup position that Obama got lucky on the timing of the economy. That’s gonna make Ferguson look a lot less

Memo to Newsweek: Since you’re running columns about political science, I suggest you ask my colleague Robert Erikson to write for you. He knows a lot, and I think he won’t embarrass you.

P.S. Just for a change of pace, I have no reason to think Ferguson is being a hack in this case. I’m guessing it’s just a combination of four factors:

1. He doesn’t know the literature.

2. He’s so convinced that Obama is a disaster that he can’t understand why anyone would vote for him (recall Ferguson’s earlier claim, reminiscent of Thomas Frank, that “If young Americans knew what was good for them, they would all be in the Tea Party.”).

3. He’s overconfident: once something comes to his mind, he doesn’t check it.

4. Nobody knowledgeable is editing his column.

I think there’s enough in 1-4 above to explain Ferguson’s error, without recourse to the “he’s just a hack” explanation.

P.P.S. No, Ferguson doesn’t really deserve all this attention. But this is a political science blog, and Ferguson did refer to “a huge body of research by political scientists.” I guess we should just be glad that this time it’s a historian (rather than, say, an economist or a physicist) getting confused about political science.

{ 10 comments }

Geof Gee September 13, 2012 at 10:06 am

Interesting. Hasn’t Ferguson suffered from a terrible case of foot-in-mouth disease lately?

http://www.theatlantic.com/business/archive/2012/08/a-full-fact-check-of-niall-fergusons-very-bad-argument-against-obama/261306/

Vladimir September 13, 2012 at 10:45 am

“(hence Obama not being hurt by the full four years of recession)” I think the NBER Business Cycle Dating folks may quibble with that statement. Anyhow, as Krugman pointed out the trend in Ohio and certainly Michigan, in output and the unemployment rate is certainly favorable to Obama. I wonder how these findings hold in particular states (especially competitive states) where other factors such as demography or “values” are stressed for their impact on the vote.

Mikhail September 13, 2012 at 11:15 am

I feel like I should defend the historical profession by pointing out that most of us don’t like Ferguson very much either.

Procopius September 14, 2012 at 6:44 am

If it helps, I am neither a historian nor a political scientist, and I think most people like me who know the name think that Niall Ferguson is a novelist, not a historian. Certainly the “counterfactual” that made his reputation should be considered fiction.

PJR September 13, 2012 at 12:24 pm

Incumbent Obama’s economic growth leading to the 2012 election most resembles what incumbent Ike had in 1956, right? This is ignored by non-experts (and, when convenient, politicians challenging incumbents) who prefer to give great weight to unemployment statistics when judging how the economy should affect election outcomes. Yet over the past several decades, it would seem that political science empirical studies have rather consistently found that unemployment metrics are poor-to-lousy predictors of presidential election outcomes. Is this poli sci finding so unbelievable that it should be ignored?

matt w September 13, 2012 at 2:11 pm

That’s a pretty spectacular graph. One thing that strikes me is that the two cases in which the incumbent party markedly underperformed the apparent trend are the two cases in which an eligible incumbent president dropped out of the election after the New Hampshire primary, due in part to an unpopular war. — I mean, I’m sure political scientists have already noticed this, but it’s pretty striking even to a layperson.

Joe in MD September 14, 2012 at 7:52 am

Since the worst unemployment is in the local and state government people, I suspect the models are wrong by assuming a broader distribution. Also the polarization present today is unprecedented, also messing with the models.

Kate September 14, 2012 at 10:45 am

Perhaps this has something to do with: (1) the economic downturn was caused by the Financial Crisis; and (2) Romney’s past is closely associated with the cause of that crisis, namely, Bain Capital.

Not all downturns are alike. So perhaps historic data is not as determinative as it might otherwise be.

Barry September 14, 2012 at 12:14 pm

Andrew, I’d point out that your four explanations for what Niall wrote don’t speak highly to his scholarly judgement:

“1. He doesn’t know the literature.”

And didn’t seem to even bother to ask anybody with a clue.

“2. He’s so convinced that Obama is a disaster ”

IOW, gets his news from Fox News (and certainly doesn’t show the sort of judgement which a history professor should show.

“3. He’s overconfident: once something comes to his mind, he doesn’t check it.”

Even when it’s out of his field of expertise, and seems to be a strange outcome, which should lead him to question snap judgements.

“4. Nobody knowledgeable is editing his column.”

See point 1.

Gnash Equilibrium September 14, 2012 at 1:44 pm

Regarding your P.P.S., I followed the “economist” link and it seems the problem was not with what Krugman wrote, but with your failure to understand irony. Your criticism makes sense only if Krugman actually thought those hypothetical percentages were realistic, but that is not a sensible way to interpret what he wrote.

Comments on this entry are closed.

Previous post:

Next post: