How We Monitor the Economy
July 5, 2017
Monitoring the business cycle is of paramount importance. Since 1965, two-year rolling losses of 20% or more in the S&P 500 Composite Index have always been preceded by a business cycle peak. Equity markets tend to be more volatile following business cycle peaks than during periods of economic expansion.
To monitor the business cycle, we developed our Economic Composite. The cornerstone of the Economic Composite is the term structure. Research by Estrella and Mishkin (1996, 1998) and Wright (2006) concludes that the term structure (i.e., three-month yields less ten-year yields) is a viable tool for predicting business cycle peaks and has a longer lead time than other economic variables. Our research concurs with these studies.
Our model calculates the current term structure’s decile rank relative to historical spreads. Other economic variables such as the ISM survey, capacity utilization and the unemployment rate are also converted to decile ranks relative to their historical values. The decile ranks of each variable are combined into one master rank. When the decile rank of the master rank records a value of one (i.e., the bottom decile), then an Economic Composite business cycle peak warning signal is triggered.
In our historical study of U.S. business cycles, Economic Composite business cycle peak warnings preceded actual business cycle peak dates (as defined by the National Bureau of Economic Research, or NBER) by an average of 289 days. Since 1960, the average decline for the S&P 500 Index following an Economic Composite warning signal for the U.S. economy to the subsequent S&P 500 Index low was 26.1%. The latest Economic Composite warning signal for the U.S. economy occurred on October 27, 2006, 430 days before the NBER peak date of December 2007.
Figure 1 shows the historical NBER cycle peak dates, historical Economic Composite warning signals, and the historical trading price of the S&P 500 Index (log scale).
Estrella, Arturo and Frederic S. Mishkin (1998), “The Yield Curve as a Predictor of U.S. Recessions,” Current Issues in Economics and Finance, 2.
Wright, Jonathan H. (2006), “The Yield Curve and Predicting Recessions,” Finance and Economics Discussion Series Working Paper No. 2006-7.