Structural breaks and nonlinearities in hours worked: are they really nonstationary?

This paper applies univariate unit root tests to annual data of 26 countries to examine the stationary properties of hours worked over the period 1950-2010. To this end, both conventional unit root tests and recent unit root tests allowing for structural breaks and nonlinearities in data generating process are utilized. The main finding is that when allowing for two structural breaks and nonlinearities in data, there is evidence that hours worked in some of the countries is stationary.