An Index with Attitude

“The Political (and Economic) Origins of Consumer Confidence” by Suzanna De Boef and Paul M. Kellstedt, in American Journal of Political Science (Oct. 2004), 350 Main St., Malden, Mass. 02148.

On the 10th day of every month, Wall Street anxiously awaits the release of the University of Michigan’s Index of Consumer Sentiment, a key measure of what Americans think about the state of the economy—and how inclined they might be to open their wallets. In a campaign year, political candidates are also keenly interested in the news from Ann Arbor. What’s rarely appreciated is that the index is not a purely economic indicator. Political feelings play a significant role in how people respond to the University of Michigan pollsters.

Shortly before the 1992 presidential election, for example, 49 percent of Republicans said they believed the economy would improve during the next year, while only 19 percent of Democrats did. A month after the election, Republican sentiment was about the same, but Democrats’ confidence in the economic future shot up to 62 percent.

Using a variety of statistical techniques, De Boef and Kellstedt, who are political scientists at Pennsylvania State University and Texas A&M University, respectively, zoom in for a closer look at what shapes consumer confidence. They estimate that about 75 percent of the index is actually determined by economic factors, with the rest being influenced by politics and other perhaps “irrational” factors.

Over the long term, according to De Boef and Kellstedt, consumer sentiment does track changes in actual economic conditions, but in the short term it’s subject to strong influence by shocks such as the Enron scandal (whose effects faded after four months) and, more significantly, by changes in public opinion about the president’s ability to manage the economy. Examining the period from 1981 to 2000, the authors conclude, “For every five percentage point gain in [the president’s] economic approval ratings . . . consumer sentiment goes up an average of one point.”

What about media coverage of the economy? Only when the sources cited in news stories are “nonpolitical” does there appear to be any effect, and even then it’s indirect. Such stories seem to influence the public’s approval ratings of the president, not its view of the economy itself. And presidents may as well forget about “talking up” the economy. According to De Boef and Kellstedt, such happy talk has no impact on consumer confidence at all.