Why The Reuters/University Of Michigan Consumer Poll Is Hogwash

The Reuters/University of Michigan Surveys of Consumers said the final reading in May for its index of confidence fell to 59.8 from April’s 62.6. According to Reuters, "May’s reading was the lowest since 58.7 in June 1980."

A look at the methodology makes the value of the results look very thin, even though they are picked up by thousands of media outlets and often move the stock market up or down.

During the survey, which is based on only 500 or so interviews, those being polled are asked 50 core questions. One has to ask what happens to the people who become frustrated and drop off the phone call, or those who rush though the later part of the survey to get on about their daily business.

The people surveyed are asked about personal finances, business conditions, and buying conditions. They are also asked about "expected" income changes, a very subjective assessment. Each subject polled is ask about their overall opinions and "in their own words their reasons for holding those views." In other words, intelligent people’s assessments are lumped in with buffoons.

The PR on the survey does indicate that those polled often hold opinions about the future of the economy which are confirmed by later events, but some of the deviations are meaningful. especially when it comes to views of changes in the prime rate and housing.

Much of what the survey’s researchers "sell" as the strengths of the poll is predictive. It is hardly clear that the information gathered is very good at reading the crystal ball.

Douglas A. McIntyre