Working Hours of Part-timers and the Measurement of Firm-level Productivity

This paper empirically quantifies what effect the data availability on part-timers' firm-level working hours may have on the accuracy of productivity measurement by using a data set on a large number of Japanese firms. Despite the practical importance of part-time workers in productivity measurement, this issue has not been considered seriously in past empirical studies. According to the analysis of this paper: (1) firm-level working hours of part-timers are quite heterogeneous even within the same industry; (2) when using industry average working hours, the bias of measured productivity is around 4% at the sample mean and from 1% to 2% at the sample median. The biases are especially large for service industries such as restaurant, hotel, and retail trade, where the part-time ratio is high; (3) however, the correlation between measured productivities using industry average working hours and those using firm-level hours is very high. This suggests there is only a small mismeasurement when using industry aggregate data in analyzing effects of firm characteristics or policy measures on productivity; (4) it is desirable to calculate full-time and part-time hours separately in productivity analyses covering service industries. In considering the importance of planning a valid economic growth strategy, enriching firm-level statistics is a cost-effective investment.