The aim of this paper is to estimate the effects of a technology shock in the euro area within a structural VAR framework. Since the impact of these shocks on labor use is a controversial issue in the related literature, we give particular attention to it. Given that the estimated effects of a technology shock are quite sensible to the low-frequency properties of the labor input measure, we resort to an extensive statistical analysis to investigate whether hours worked are better characterized as stationary or difference stationary. We conduct a battery of classical unit root and stationary tests, analyze the small-sample properties of some of the tests-statistics, explore encompassing tests and Bayesian odds ratios to ascertain if the more appropriate VAR model is the one in which hours per capita enter in levels or first-differences. The evidence gathered is in support of hours being stationary, which leads to the conclusion that per capita hours worked rise after a technology shock in the euro area.
As for the responses of the remaining variables, our results are in line with the bulk of the literature. |