Working Paper: NBER ID: w1444
Authors: J. Bradford DeLong; Lawrence H. Summers
Abstract: This note shows that contrary to widespread belief there is little evidence that the business cycle is asymmetric. Using American data for the pre- and post-war periods and data on five other major OECD nations for the post-war period, we are unable to support the hypothesis that contractions are shorter and sharper than expansions. We conclude that there is not much basis for preferring some version of traditional cyclical techniques to more modern statistical methods.
Keywords: business cycles; asymmetry; GNP; industrial production; unemployment
JEL Codes: E32
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| economic downturns (F44) | brief and severe relative to trends (E32) |
| cyclical asymmetries (E32) | significant skewness in growth rate distributions (D39) |
| skewness coefficients (C46) | significant deviation from zero (C46) |
| median output growth rate (O40) | exceed mean by significant amount (Y60) |
| skewness in unemployment rates (J60) | strong general feature of business cycles (E32) |