“What people say, what people do, and what they say they do are entirely different things.” Margaret Mead

Judging by the reaction of equity markets and news media to the monthly release of consumer sentiment measures, which can easily send the Dow up or down 100 points, you would have to conclude that economists and stock traders haven’t learned the basic lesson of watching people’s actions, not listening to their words. Each month the University of Michigan (UM) conducts a survey asking individual households how they view their personal finances, business conditions and buying conditions. From this data, UM calculates the consumer sentiment index, which many investors and financial analysts use to gauge the strength of the economy and the markets. The belief is, when people are feeling bad about the economy, whether it is due to instability in the Middle East or an increase in the unemployment rate, they will cut their spending to reflect their weariness about the future. Because consumer spending represents a full 70% of all economic activity, forecasting consumer behavior by this or any other measure should provide a forecast of changes in the economy and stock market. While we agree that forecasting consumer behavior is the key economic variable, we find absolutely no support for the notion that sentiment surveys help in estimating what consumers will do next.

Try comparing almost thirty years of monthly returns for the UM Consumer Sentiment Survey index and the S&P 500, looking for a relationship that, according to many economists, should exist. The result? Changes in the Consumer Sentiment Index and the S&P 500 are effectively uncorrelated. That is to say, at a correlation coefficient of 0.195, the correlation is too small to be economically useful, meaning there is no significant relationship in the movement of the two indices over time. If they moved in the same direction, the correlation would have to be above 0.50 to be significant, and above 0.75 to be really meaningful or truly useful. If they moved in opposite directions, the measure would have to be -0.50 and -0.75 to be significant and meaningful, respectively. If we adjust the data by lagging the S&P 500 by a month – supposing that how consumers feel today will affect their future behavior - the correlation falls to almost exactly zero. While there are times when the two measures move in sync, there are also times when they move in opposite directions. This of course is what leads to their lack of meaningful correlation.

Another index used to measure consumer sentiment is the RBC CASH Index, created in 2002, which is no better a gauge of consumer behavior than the UM measure. The correlation between the RBC CASH Index and the S&P 500 (both lagged and un-lagged) is less than 0.20.

The evidence illustrates Americans don’t change their spending habits based on how they feel in any given month. Why? Americans are more concerned about maintaining their standard of living than worrying about the ‘what if’s’ of tomorrow’s economy. Age and stage of life are the best predictors of spending, not surveys. As people age, they move through very predictable productivity, earnings and spending patterns. The consumer life cycle affects everything from the demand of potato chips to cycles of innovation and economic growth. Dad may gripe about interest rates or the price of oil, but that doesn’t stop him from buying Junior a new bike. And it certainly doesn’t affect his regular monthly payments on things like the mortgage, the car, or cable TV.