Another Look At The Stock Market & The Business Cycle?

The stock market’s annual performance is comfortably in the black again. After a brief slump into negative territory on a year-over-year price basis in late-September and early October, the S&P 500 is higher by 4.9% through Oct. 21. Is that a sign that the economy will keep growing? History offers some evidence for responding with a cautious “yes.”

The usual caveats apply, of course, starting with Paul Samuelson’s famous remark that the stock market has predicted nine of the last five recessions. Perfection is not Mr. Market’s forte when it comes to discounting the future. The same is true for every other forecasting technique (or forecaster), which brings us back to square one: Does the stock market have any value for evaluating the macro trend? Yes, it does, or so the track record suggests.
The main glitch in looking to the equity market for a guesstimate of what’s coming is the habit of seeing recessions that never materialize. By contrast, the market’s record in forecasting growth boasts a better record. For example, if we use the S&P 500’s 12-month percentage change as a measure for anticipating economic performance for the year ahead, the historical record is encouraging.

Over the past half century of the American business cycle (according to NBER dates), only one of the previous eight recessions has struck without a decline in the S&P 500′s annual percentage change (either before or in the early stages of a downturn). The exception is the fleeting recession of 1980, which lasted a brief six months—the shortest on record.
Perhaps that record is anecdotal, which inspires looking for more quantitative evaluations. One possibility is analyzing the stock market’s annual performance relative to proxies of the business cycle, such as the 12-month percentage changes for industrial production and non-farm payrolls. Here too there’s reason to think that equity prices are a useful if flawed predictor of the broad cycle.
As an example, consider cross correlations for the annual changes of the S&P 500 and industrial production, as shown in the chart below. Correlation is a measure of statistical dependence with readings ranging from -1.0 (perfect negative correlation) to zero (no correlation) to 1.0 (perfect positive correlation). Correlation isn’t necessarily evidence of causation, of course, but it’s a useful way to begin looking for evidence (damning or otherwise) of relationships between data series. With that in mind, the cross correlations for the S&P and industrial production suggest that the equity market anticipates changes in industrial production at roughly six to eight months in advance. Or so one can infer from the peak in correlations for these series at roughly 0.55 at the -6 month level (see chart below).

The relationship isn’t as strong when we look at rolling 12-month changes for the equity market vs. the equivalent for non-farm payrolls. But it’s also hard not to notice that the correlations for this pair also show a relatively strong relationship at the -6 to -8 month readings. The implication: stock prices anticipate changes in the labor market by around six to eight months in advance.

Like every other predictor, you can’t count on the stock market for flawless forecasts. But as the charts above suggest, you shouldn’t dismiss Mr. Market’s implied predictions either. No one should use this or any other predictor in a vacuum. That said, the stock market appears to be voting in favor of growth these days, if only moderately. Yes, it’s the worst forecasting tool we have… except in comparison to most of the alternative techniques for peering into the future.

One Response to Another Look At The Stock Market & The Business Cycle?

  1. EconomicsLOL says:

    Interesting piece. I wonder if shock events, a la Lehman (or Euromaggedon) would change results?

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>