Does the GOP nominee’s ideology matter?
posted at 11:35 am on November 17, 2011 by Karl
Having critiqued Nate Silver’s model for calculating the odds of various GOP candidates winning the 2012 presidential election — particularly his inclusion and assessment of candidate ideology — I was curious as to what political scientists would make of his model.
Brendan Nyhan and Jacob Montgomery do the most extensive deconstruction of Silver’s model, pointing out that it performs worse than Douglas Hibbs’ “Bread and Peace” model (which Silver criticized) on out-of-sample data, and has a larger mean average error than all of the most well-known election forecasting models.
Regarding candidate ideology, Nyhan and Montgomery add:
First, when the economy is growing and presidential approval is high, strong moderate candidates may be scared off from entering the race, leaving only ideologues. A similar effect has been shown when one party has held the presidency for a long period of time. When this happens, the opposition tends to perform better due to the perception that is “time for a change”, and opposition parties are likely to nominate more moderate candidates in the hopes of regaining control of the White House at the expense of ideological purity.
Second, the estimates of challenger ideology that Silver uses are primarily drawn from voter perceptions of the candidates. However, these perceptions are driven by the content of the campaign, which is itself shaped by the economic context. Candidates who appear extreme in one era may seem less so in the next (consider the changing perceptions of Ronald Reagan between 1976 and 1980, for instance). For all of these reasons, Silver’s estimates of the effects of challenger ideology and election outcomes are likely to be significantly exaggerated.
Similarly, Seth Masket notes:
[P]erceptions of the Republican nominee’s ideological stances may well change by next year. It’s very hard to make realistic projections of Cain’s governing ideology since he’s never governed before. Perry would be facing a more liberal electorate than he’s ever faced, and Romney would be facing a more conservative one. Plus, given Romney’s history, there should be substantially large error bars on either side of his line.
Indeed, while I cannot find a good link at the moment, some political scientists argue that voters’ perception of ideology would be a better variable for forecasting.
Alan Abramowitz also has several problems with Silver’s model. Of course, Abramowitz has his own “Time For a Change” model, which gives Obama a good chance of winning a second term even with fairly modest economic growth next year and an approval rating in the low- to mid-forties. It’s worth noting the “Time For a Change” model has over-predicted the vote of the incumbent candidate by at least 1.85% in each of the last four presidential elections.
For more criticism of Abramowitz’s model — and of election forecasting models generally — see Sean Trende. Although I’m a fan of Trende’s work, his criticism of these models has its own weaknesses. In particular, Trende does not really acknowledge political scientists admit the limitations of such models. Nyhan among others stresses the problem inherent in a small data set, and the risk of overfitting models to conform to past results. Hibbs will admit his model only accounts for approximately 77% of election results (iirc). James E. Campbell, creator of the “Trial Heat and Economy” forecasting model would be the first to admit that model blew up (.pdf) in 2008 due to the intervention of the financial crisis. That a model does not forecast unknown unknowns is not a strong criticism (even though some foresaw the financial panic, few would have pinpointed its eruption to the month).
But I digress. The takeaway here is that election forecasting models are admittedly limited attempts to quantify the basic factors on which elections usually turn — peace and prosperity. Other factors may matter, but the challenger’s ideology likely does not matter more than a point or two — which is within the margin of error for even the best models. Silver’s model was probably a nice traffic driver for the New York Times, but it likely overestimates the effect of challenger ideology. Nyhan and Montgomery did not directly test whether adding estimates of challenger ideology to existing forecasting models would improve their performance, but the Silver model’s large mean average error compared to others suggests an answer.