Working Paper: NBER ID: w2229
Authors: Robert King; Charles Plosser; James Stock; Mark Watson
Abstract: Recent developments in macroeconomic theory emphasize that transient economic fluctuations can arise as responses to changes in long run factors -- in particular, technological improvements -- rather than short run factors. This contrasts with the view that short run fluctuations and shifts in long run trends are largely unrelated. We examine empirically the effect of shifts in stochastic trends that are common to several macroeconomic series. Using a linear time series model related to a VAR, we consider first a system with GNP, consumption and investment with a single common stochastic trend; we then examine this system augmented by money and prices and an additional stochastic trend. Our results suggest that movements in the "real" stochastic trend account for one-half to two-thirds of the variation in postwar U.S. GNP.
Keywords: stochastic trends; economic fluctuations; VAR models; cointegration; macroeconomic analysis
JEL Codes: E32; C32
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| movements in the real stochastic trend (C22) | variation in postwar U.S. GNP (N12) |
| innovations in the real permanent component of GNP (O39) | output (C67) |
| innovations in the real permanent component of GNP (O39) | consumption (E21) |
| innovations in the real permanent component of GNP (O39) | investment (G31) |
| innovations in the permanent component (O39) | fluctuations in GNP at the 14-quarter horizon (F44) |
| innovations in the permanent component (O39) | fluctuations in GNP at the two-year horizon (F44) |
| innovations in the permanent component (O39) | fluctuations in GNP at the four-year horizon (F44) |