Quick Hit: Falling Auto Sales Don’t Foretell Crashes

You can't gauge stocks' direction based on the number of deals made in the showroom. Photo by leventince/iStockPhoto.

The latest indicator media would like you to believe magically signals trouble for the bull market is a three-month-long dip in US auto sales. The theory says such big-ticket purchases are uniquely forward-looking, as it requires strong confidence to make such a large commitment. Confidence is good, so falling sales signaling falling confidence would be bad. Perhaps the theory is true, but if so, someone forgot to tell markets. History shows auto sales do not predict turning points for stocks.

It is always a fallacy to think one indicator is the forecasting silver bullet rendering all others obsolete. But this is particularly true when we're discussing measures of past economic activity. Stocks are forward-looking economic indicators-they assess expected conditions over roughly the next three to thirty months, focusing most on the next year or so. Economic data are backward-looking. There is nothing special about larger purchases or auto purchases to make those data somehow more forward-looking than purchases of any other widget.

Don't take our word for it-there is ample data showing US light vehicle sales simply do not reliably foretell equity market inflection points. Exhibit 1 plots the seasonally adjusted annual rate of US light vehicle sales (in units) from the US Bureau of Economic Analysis against S&P 500 historical bear markets since 1976. (That is the BEA's full run of car sales data.) Exhibit 2 shows the same, only the sales data are smoothed using a six-month moving average to facilitate trend identification.

Exhibit 1: US Light Vehicle Sales and S&P 500 Bear Markets

Source: FactSet, as of 4/6/2017. Seasonally adjusted annual rate of light vehicle sales, January 1976 - March 2017.

Exhibit 2: Six-Month Moving Average, US Light Vehicle Sales and S&P 500 Bear Markets

Source: FactSet, as of 4/6/2017. Six-month moving average of seasonally adjusted annual rate of light vehicle sales, January 1976 - March 2017.

Note the steep decline in the late 1970s began more than two years before the S&P 500 entered a bear market. Similarly, car sales were rocky for many months before August 1987's peak. Heck, auto sales fell during basically all of the brief December 1987 - June 1990 bull. And the ensuing bear. And they were flattish for long periods during the mid-1990s and flat-to-down during the 2000s bull. About the only time falling auto sales correlated with a market peak in the history of the series is 2000. One correct read. A slew of falsehoods.

The theory car sales declines foretell big equity market problems is a clunker.