Nonlinearity in Nexus between Working Hours and Productivity

Average weekly working hours in most industrialized economies have displayed steady declines since the 1970s. Changes in working hours may have two contrasting effects on hourly productivity: a "fatigue effect" and a "learning effect." An increase in working hours may lead to the accumulation of a worker's proficiency and skill in his or her job, while it may at the same time cause the level of the worker's fatigue to increase. Estimation of the Cobb-Douglas and translog production functions with Hansen's (1999) threshold regression methods shows that there exist multiple (at most four) thresholds in the linkage between working time and productivity, supporting the existence of a learning effect as well as a fatigue effect from an extension in working hours. The results of our study provide some implications for the recent discussion on reform of the working time system: (i) a reduction in working hours, on one hand, may increase productivity by reducing fatigue and allowing more leisure, but on the other hand, it may hinder workers from accumulating proficiency and skills, thereby reducing their productivity; (ii) the overall effects of changes in working time depend upon a variety of factors, such as the initial level of working hours and the features of the industry concerned