The Stationary of Productivity Shocks: Evidence from 25 OECD and Big 7 Countries

Abstract

We apply the panel covariate augmented Dickey-Fuller test to test stationarity of the productivity series for the OECD and Big-7 economies. The approach takes cross-sectional dependence into account. Using hours-worked per worker, we find that the series is non-stationary for the 25 OECD countries; but for the Big-7 the results are mixed. So, this paper achieves a battery of panel unit root tests to examine the stationarity properties of the series named hours worked per employee. The study period covers 1960-2012 for the OECD and the Big 7 countries. The tests we use account for cross sectional dependence and those that do not account for such dependence. Our results suggest that an analyst might infer that hours worked fall after a positive technology shock, when it may go up in a true data-generating process. The findings also suggest that although in a true data-generating process, the series may go up from a positive technology shock, analysts may infer a fall. The stationarity of the series is relevant in determining the effect of positive technology shock on productivity.

Keywords

• productivity shock
• panel unit root
• cross sectional dependence
• oecd and big-7 countries