" Trend Breaks, Long-Run Restrictions, and the Contractionary Effects of Technology Improvements"

Abstract

Do hours worked rise or fall following an increase in the level of technology? Empirical results from structural VARs with long-run restrictions are often sensitive to whether hours worked enters the VAR in loglevels or in growth rates. However, this paper finds that once one allows for (statistically and economically plausible) trend breaks in labor productivity, the treatment of hours is relatively unimportant: Even with hours worked in levels, hours fall on impact following a technology improvement. Indeed, encompassing tests suggest that the specification with breaks can easily “explain” (or encompass) the large positive apparent response when the breaks are omitted; in contrast, the no-breaks specification has more difficulty explaining the negative response when breaks are included. The issue is the common low frequency high-low-high pattern of hours per capita and productivity growth in the post-world-war II period. Simple analytics plus simulations demonstrate that such low-frequency correlation almost inevitably implies a positive estimated impulse response, regardless of whether the low-frequency correlation is actually causal.This suggests the need for caution in interpreting the results of estimated responses with long-run restrictions, since results may be driven by effects other than those you think you are identifying. I thank Shanthi Ramnath, David Kang, Stephanie Wang, and Andrew McCallum for superb research assistance. I also thank Toni Braun, Susanto Basu, Jeff Campbell, Larry Christiano, Martin Eichenbaum, Jon Faust, Jonas Fisher, Neville Francis, Jordi Gali, Valerie Ramey, John Williams, and Rob Vigfusson for helpful conversations. I thank Valerie Ramey and Rob Vigfusson for providing computer code. The views expressed in this paper are my own and do not necessarily reflect the views of anyone else affiliated with the Federal Reserve System. JEL Codes E24, E32, O47, keywords Technology shocks, business cycles, structural change.