The Economy, and Voters’ Perception of It, Are Two Very Different Things

by Lynn Vavreck

We learned on Thursday that the United States economy expanded at a healthy rate this summer, growing 3.5 percent from July through September. This growth — as robust as it is — is unlikely, however, to influence how the president’s party fares this week at the polls. Midterm election results are not as directly related to objective economic conditions as presidential election outcomes are, something we discussed at The Upshot.

But the two types of elections have one thing in common when it comes to the nation’s economy: People’s assessments of how the economy is doing are equally related to vote choice in both midterm and presidential elections, even though objective economic conditions correlate much more strongly to presidential outcomes.

If you’re scratching your head trying to figure this out, you’re not alone. The only way this can be true is if something is mucking up the relationship between the actual economy and people’s judgments about the economy. And that’s exactly what is happening.

To understand how opinions about the economy and the actual economy might affect vote decisions in different ways, I looked for patterns of association between the real economy (growth) and election outcomes (seats in midterm years and vote share in presidential years) over the period from 1948 to 2012.

The correlation between growth in gross domestic product in the first six months of a presidential election year and two-party vote share for president is 0.64, and it’s even higher for another common objective economic indicator, real disposable income. But in midterm elections over the same period, the correlation between G.D.P. growth and the number of seats the president’s party gains or loses in the Senate is 0.03. In other words, there is a strong relationship between the nation’s growth rate and presidential vote share, but no relationship between the nation’s growth rate and Senate (or House) seat loss in midterm years.

Apparently, the objective state of the economy comes to bear on people’s choice for president in a way that it does not in their choice of representative or senator.

Where people’s assessments of the economy are concerned, however, it’s a different story. The American National Election Studies first asked people to evaluate the nation’s economic performance in 1980. Since then, the correlation betweeen the share of people who think the economy has gotten better and vote outcomes in presidential elections is just 0.4. In midterm years, the correlation between sanguine assessments of the economy in the A.N.E.S. and Senate seat loss is also 0.4. (It is less — 0.2 — in the House.)

It turns out that what’s mucking up the relationship between the actual economy and people’s opinions about it is partisanship.

For example, data from the most recent wave of the The New York Times/CBS News Battleground Tracker reveals that, on average, 24 percent of Americans think the nation’s economy has gotten better over the last year. But when that number is broken out by party, a revealing pattern emerges: Among self-described Democrats, 43 percent think things have gotten better this past year, but among repugicans, only 8 percent think so. Independents fall in the middle at 20 percent. It’s the same economy, but people view it differently, a result that is not particular to 2014.

Partisanship functions as a lens through which perceptions of the state of the nation’s economy are filtered. If you’re in the incumbent president’s party, your lenses are more likely to be rose-colored, and if you’re not, your lenses tend to be very dark — regardless of how the economy is actually performing! People answer survey questions about the economy as if they are being asked about their party identification. For some reason, asking people to “think back over the last year” and tell us whether they think “the economy has gotten better, stayed the same, or gotten worse,” actually activates their party identification in a way that perhaps even they don’t appreciate.

Despite this pattern of partisan conditioning, voters nonetheless manage to bring the actual state of the nation’s economy to bear on choices in the voting booth, but only in presidential election years. In midterm elections, the economy matters only through attitudes constructed through the lens of party identification — and remotely, through presidential approval.

The next time you think you’re judging an objective reality, stop and ask yourself whether your partisan lenses are on — or off.