Working Paper: NBER ID: w10631
Authors: Neville Francis; Valerie A. Ramey
Abstract: This paper investigates the source of historical fluctuations in annual US data extending back to the late 19th century. Long-run identifying restrictions are used to decompose productivity, hours, and output into technology shocks and non-technology shocks. A variety of models with differing auxiliary assumptions are investigated. The preferred model suggests that the Great Depression was a period in which both types of shocks were very negative. On the other hand, our estimates support the microeconomic evidence of historically large positive technology shocks from 1934 to 1936. Finally, both types of shocks are responsible for the reduction in the variance of output in the post-WWII period.
Keywords: No keywords provided
JEL Codes: E2; E3
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Technology shocks (O33) | Variance in productivity (O49) |
| Technology shocks (O33) | Economic performance during the Great Depression (N12) |
| Non-technology shocks (E39) | Economic performance during the Great Depression (N12) |
| Technology shocks (1934-1936) (N13) | Economic performance (P17) |
| Reduction in variance of output (post-WWII) (C29) | Technology shocks (O33) |
| Positive technology shock (O49) | Productivity in early period (O49) |
| Positive technology shock (O49) | Labor input (post-WWII) (J29) |