Gasoline Prices, Discretionary Spending Don't Mix

A guy walks into a restaurant and asks what's good. The waiter tells him the soup du jour was great yesterday. How different, really, is this from our task as market analysts?

Last week's great concern was that consumers were not taking the $10 or $20 they might be saving at the gasoline pump to spend $100 at the mall. I regard this as prima facie evidence of sanity, but clearly I'm not in tune with the times.

Never Has Been, Probably Never Will Be

Our good friends at the American Automobile Association keep a time series of national average retail gasoline prices, which I suppose is harmless enough. How is this related to the total returns for the Russell 3000 consumer discretionary sector index? The answer, in two words, is amazingly poorly.

First, let's map the two series on a common logarithmic scale and a one-year rolling correlation of returns between them. I use the one-year period to sidestep the effects of seasonality for the simple reason there are two things you don't want to step in, and seasonality is one of them. Please note how I have led the consumer discretionary sector's total return by 35 days; more on that below.

Source: Bloomberg

Also, please note how weak and irregular the correlation of returns is. It has been negative throughout 2015 with the exception of a few days in January as the weak uptrend in equities has intersected with a strong downtrend in gasoline prices. The correlation level last week was a statistically insignificant negative 0.021. For most of the history of the two series, the correlation of returns has been strongly positive; that is, rising gasoline prices and rising consumer discretionary returns are lying together like the lion and the lamb.

I mentioned above gasoline prices lead consumer discretionary returns by 35 days on average. If we construct a two-day correlogram of the two series, we see the leading effect of gasoline expands going back to that lead-time. Consumer discretionary sector returns have a very weak leading relationship to gasoline prices peaking at 19 days. Restated, the anticipation of higher consumer spending does not transmit strongly into more gasoline consumption.

Source: Bloomberg

I might add the relationship seen above for the consumer discretionary sector applies to the broad market as measured by the Russell 3000 index as well. Maybe there was a negative relationship back in the 1970s or early 1980s, but people have changed their behavior enough over the years so that higher gasoline prices simply do not sink the stock market. And, as we should have learned last week, lower gasoline prices do not induce behavior seen in drunken sailors or normal politicians.

At the time of publication, Howard Simons had no positions in the stocks mentioned.