Thursday, August 28, 2008

What does Consumer Confidence really tell us about the US economy?

The consumer confidence survey, published by the Conference Board, gets a lot of press. A Google search of consumer confidence gets 5,410,000 hits and the first media title is a Reuters article from the survey’s August release, where confidence rose 5 points to 56.9, Consumer confidence bounces.

My impression of the consumer confidence measure is that it is just a poor measure of gas and home prices.

Confidence dropped in the early 1980’s following the oil shocks of 1973 and 1979, in 1991 with the Gulf War, and again over the last year when oil topped $145/barrel in July. Notice how confidence turned up in August, just as oil and gas prices began to subside. Confidence also is very susceptible to the housing market (as it should be). Further, consumer confidence hit its lowest levels during the last two housing corrections, in 1992 and in 2008.

Upon first glance, consumers are very good at calling a recession. However, there are plenty of mid-cycle dips (circled areas) in consumer confidence, where the economy is not in a recession. So I have to ask myself, do consumers really know what they are talking about?

The Confidence Board surveys 5,000 individuals for the following purpose: “The Consumer Confidence Survey™ reflects prevailing business conditions and likely developments for the months ahead. This monthly report details consumer attitudes and buying intentions, with data available by age, income, and region.”

Certainly it details consumer attitudes and intentions. The August report indicates that current business conditions have not changed, the labor market is worse, and the economy is now more likely to get better over the next 12 months (although consumers are still quite negative about the outlook). This is certainly consistent with the consensus outlook on Wall Street and not new information. But do these consumers really know anything about business conditions?

This graph reinforces the fact that consumers are really not the best judge of economic conditions. Standard banking practice sets interest rates according to inflation expectations. If banks expect inflation to rise over the next year, interest rates (mortgage, car loan, savings, etc) will rise with the expected supply of money. In order to maintain the same real return, nominal rates must rise.

Consumers are obviously unaware of this simple fact; there is absolutely no correlation between consumer inflation expectations and the interest rate.
  • During the 1990’s, consumers were fairly confident that inflation was going to be stable in the oncoming 12 months, while at the same time, interest rates were expected to be quite volatile.

  • Since October 2007, inflation expectations have surged and the number of people that believe interest rates will be higher over the next 12-months has been falling.

So what is driving interest rates? Apparently it is Fed policy.

Interest rate expectations appear to be highly correlated with Fed policy. Why wouldn’t they? That is the goal of Fed policy: target lower interest rates in order to drive down long-term interest rates (mortgage, car loan, savings, etc). However, long-term interest rates don’t always follow the Fed’s lead. Mortgage rates – which have stabilized over the last three weeks – have risen since May, but the Fed funds target fell 3.25% since last year.

If one wants data on gas prices, it is better to download it directly from the Energy Information Administration’s website, rather than use the Confidence Board’s faulty measure of consumer confidence.

Please leave comments. Best, Nontruths


  1. We just can't be trusted, can we? Consumer confidence is a subjective measure so I guess is supposed to have little basis in reality. Maybe it can be used as the contrary indication to what is statistically going on.

  2. What surprises me as soon as I take more than a mere first glance at nearly every survey I read on cnbc, msn, or many of the major news outlets, is these surveys always seem to be conducted among a minimal amount of people! Considering the US population, 5,000 people is hardly a suburb.

    It's no surprise to me that a survey of 5,000 people (who are these people by the way, normal citizens, experts) could not possibly and properly correlate consumer inflation and interest rates.

    I will conquer that consumer confidence does usually correlate with weak consumer spending and similar reports.