The effects of reducing working hours

The debate on if and how to shorten the length of the working week has strengthened in the past years, especially because of the COVID-19 pandemic. The number of working hours can have a direct impact on economic outcomes such as labor cost, employment, and productivity, but also workers’ health and welfare. Since 1995, there have been five national reforms reducing standard working time in European countries. Aside from the (in)famous French 35-hour reform, an interesting yet understudied case study is that of the 1996 Portuguese reform.

Portugal began significantly reforming the length of its working week in 1991. At the same time, a non-binding agreement was made to further cut hours to 40 hours by 1995 through collective agreements, but the gradual reduction occurred only partially. In 1996, national legislation cut the number of working hours to 40 (from about 45 hours across the various collective agreements). This reform provides an ideal case study: i) firms had to significantly adjust their standard hours of work; ii) the standard day stayed at 8 hours per week but the number of work days went from 6 to 5 per week; iii) contrary to the French reform, it did not entail any compensating measures.

This paper shows that as expected working hours decreased significantly in affected firms (relative to firms whose collective agreement already specified 40 hours). Not surprisingly, giving downside wage rigidity, wages per hour increased significantly, but overall monthly salaries did not adjust. Theory would argue that an increase in the labor cost would lead to employment loss, yet employment was not significantly affected. We show that more affected sectors experienced significantly higher price growth, suggesting that the higher labor cost was shifted onto consumers. The policy implication is that stricter (and costlier) regulations for firms do not have to be negative for employment insofar as firms have other margins of adjustment. However, this raises other questions, such as the distributional and potentially negative effects for consumers.

Click here to go to the paper by Marta C. Lopes and Alessandro Tondini.