Comment

Abstract

One of the recent major debates in macroeconomics concerns the role of technology shocks that play a major role in the standard real business cycle (RBC) model. One aspect of this debate concerns the response of hours to technology shocks. Galí’s (1999) empirical evidence concluded that the technology shock was not a key driver of the business cycle and cast doubt on the RBC modeling paradigm. Galí identified the effects of technology shocks using an estimated vector autoregression (VAR) that relied on the identification mechanism that only technology shocks have a permanent effect on labor productivity; this assumption is justified by the main RBC andNewKeynesian models. He finds that hours worked fall in response to a positive technology shock; the negative correlation between labor input and output contradicts the prediction of both the RBC model and the empirical data of a positive correlation. Since it was first published in 1999 (having been earlier released as a working paper in 1996), this “TFP ↑ ⇒ hours ↓” finding has created a great deal of discussion. Other authors such as Francis and Ramey (2005) and Basu, Fernald, and Kimball (2006) endorse the finding of Galí. The former paper carries out a number of robustness checks on the identified technology shocks, whereas the latter uses an entirely different methodology to identify the technology shocks. Both corroborate the TFP ↑ ⇒ hours ↓ finding of Galí. There have been twomain angles of attack on this finding: one concerns the low-frequency properties of the hours data series, and the other considers the identification assumption in the Galí work. This latter approach suggests that other factorsmay affect productivity in the long run. One example of the approach that questions the identification strategy is the article by Fisher (2006). Fisher allows for both traditional technological