The Banality of Elections

No amount of good messaging can trump the effects of the economy

This is the second election cycle that political junkies have had Nate R. Silver’s election models, which use reams of polling and demographic data to make predictions that proved exactly right in all Senate contests and off in only two states in the presidential race. Given that track record, Silver’s current projections do not look good for Democrats. As of this writing, Silver sees Democrats losing the House, with the GOP taking a 230 to 205 majority and keeping the Senate with a bare 52 to 48 majority.

But as fascinating as Silver’s model is, it tells us very little about why these swings are occurring. Silver’s polling and demographic data do not say much about voters’ motives, and he does not correlate his results to other factors to see what is driving voter sentiment. Luckily, there is a large literature in economics and political science tackling just this question, and its conclusions boil down to a confirmation of James Carville’s famous epigram: “It’s the economy, stupid.”

Economic models for predicting election results took off in 1978, after the publication of Yale Economics Professor Ray C. Fair’s “The Effect of Economic Events on Votes for President”. Fair has updated his model for each election since and has concluded that three economic factors can produce a reliable prediction of election outcomes: per capita economic growth, inflation, and the number of quarters during the current president’s term that per capita economic growth has topped 3.2 percent. Since Fair started making predictions, he has only missed one election, in 1992. On average, the model is off by only 2.5 percentage points. In 2008, it missed Obama’s vote share by only 1.5 points.

Applying these methods to Congressional elections is controversial, but Fair and others insist it can be done. Fair has found that the same economic factors that he uses to predict presidential elections can be used to predict on-year Congressional elections, and even fewer economic criteria are needed to predict midterms. Fair’s equation worked the last go-around, coming within a percentage point of Democrats’ actual share of the vote.

Fair’s final prediction for 2010 has yet to be released, and his model only produces a vote share estimate, not how many actual seats each party is expected to receive. However, the political scientist Douglas A. Hibbs Jr. has produced his own economic model relying on changes in disposable income rather than Fair’s three metrics and predicts that Democrats will win 211 seats and Republicans 224. This estimate is only a few seats off from Silver’s prediction, despite not using a single piece of polling data.

Some object that this way of analyzing elections is reductive and soulless. Deborah Solomon admonished Fair during an 2004 interview with him for the New York Times Magazine, saying, “It saddens me that you teach this to students at Yale, who could be thinking about society in complex and meaningful ways.” It should sadden all of us that a journalist at one of the most influential newspapers in the world thinks complexity and meaning are incompatible with social scientific rigor.

Far from making elections uninteresting, economic models, and their success, serve as evidence that voters know what they are doing.  Ensuring good economic performance is perhaps the most important task of a government. How a government handles unemployment, inflation, and wage growth has a more powerful effect on the real lives of voters than its conduct in any other policy realm. It’s perfectly rational, then, for voters to throw the bums out when the government isn’t ameliorating economic pain or to keep incumbents in office when things are going well.

Or it would be rational if the bums being thrown out of the American government were more unified. Between our bicameral legislature, our separate executive branch, the independent Federal Reserve, and the Senate’s de facto supermajority requirement, it is hard to say that the Democrats set to lose reelection are truly responsible for high unemployment. Indeed, I would argue that Senate Republicans who have blocked jobs bill after jobs bill, and forced the White House to shrink the stimulus package, bear more responsibility. What’s more, this was likely intentional. Why would Congressional Republicans support policies leading to economic growth when further stagnation would lead to gains in the next election?

Compare this to what’s happening in the U.K. now. The Conservative-Liberal Democrat coalition has absolute control over fiscal policy and is using it to issue sweeping spending cuts and tax hikes. Governing is so easy that the coalition implemented a carbon tax almost by accident. Like many others, I doubt the cuts and hikes will work, but their effects, whatever they are, will belong solely to the governing coalition. If they fail and the coalition is booted from office due to poor economic conditions, voters’ economic motivations will serve to punish the guilty party. By giving total power to the winning party, the U.K.’s system makes that party truly accountable.

Sadly, an American parliament isn’t in the offing anytime soon. But at the very least, amending Senate rules to abolish the filibuster and eliminate “holds”—which allow a single Senator to block legislation—would give the majority more power, and thus more accountability.

If voters are going to judge Congress on economic conditions, Congress should have a greater ability to change them.

Dylan R. Matthews ’12, a Crimson editorial writer, is a social studies concentrator in Kirkland House. His column appears on alternate Tuesdays.