European Economic Perspectives 25

The Thirty Nine Hours

France has introduced a mandatory reduction in the working week. Will this reduce unemployment?

In 1998, the French government passed legislation stipulating a maximum working week of 35 hours by the year 2000. French firms are now in the process of negotiating the timetable for reducing hours with their unions or personnel delegates, starting from the current 39-hour week. Similar projects have been envisaged in other European countries, notably Italy and Spain, in the hope that cuts in working time will prove to be an efficient policy for reducing unemployment.

But while the idea of mandatory reductions in weekly hours is becoming increasingly popular among European politicians and union leaders, there have been very few empirical assessments of the effects of such regulation. This is surprising since France has already carried out such an experiment.

On 1 February 1982, a few months after François Mitterrand became President, the government decreed a shortening of the working week from 40 to 39 hours. This policy had been promised in the socialists’ election manifesto but its prompt implementation was still largely unexpected. At the same time, the government mandated stable monthly earnings for workers on the minimum wage, which had itself been raised by 5% in July 1981. It also recommended stabilization of monthly earnings for other workers, advice followed by 90% of firms.

A recent CEPR Discussion Paper by Bruno Crépon and Francis Kramarz investigates the impact of the 1982 reduction in working hours on workers’ transitions between employment and unemployment. The analysis uses longitudinal data on individual wages, hours and employment derived from the French Labour Force Survey (Enquête Emploi) for the period 1977–87.

The researchers take two different approaches to the data, based on two natural experiments associated with the cut in hours. In the first, they compare workers who worked between 36 and 39 hours before 1982 with observationally identical workers who worked exactly 40 hours. The data reveal that workers directly affected by the changes (those working 40 hours in March 1981) were more likely to lose their jobs between 1981–2 than workers not affected by the changes (those working 36–39 hours in March 1981). Employment fell by between 2% and 4%.

The second experiment takes advantage of a specific feature of the implementation process of the reduction in hours: the fact that some firms were surprised by the 1 February 1982 decree enforcing the new standard. Hence, by April 1982, when that year’s French Labour Force Survey took place, a sizeable share of firms had not had time to complete negotiations with their employees and many workers were still on 40-hour weeks. The research indicates that these workers were more strongly affected by the reduction in standard hours, with induced job losses of at least 4.1%.

Why did the maximum hours legislation lead to these large employment losses, an outcome so contrary to the goal of the policy? What seems to have happened is that with workers’ monthly earnings typically stable, the reduction in hours effectively increased the hourly pay of those working more than 39 hours. Any gains in workers’ hourly productivity seem to have been insufficient to compensate employers for this increase and they responded by firing people.

The research also shows that not only did employment fall as a consequence of the hours reduction, but lower-wage workers suffered disproportionately. Indeed, minimum wage workers were the group most affected by the changes, with job losses of more than 8%.

This result supports the view that the combination of reduced hours and mandatory (or at least, recommended) earnings stability led to an increase in hourly pay that was unacceptable to many firms. Wage rigidity was most binding for low-wage workers, particularly after the 5% increase in the minimum wage, and the most effective cost-minimization strategy for firms was to fire some of these workers and hire new ones. Such replacements could still be paid the minimum wage rate but they had monthly earnings that were based on actual hours worked.

As the researchers acknowledge, the 35-hour working week currently being introduced may not have the same effects as the 39-hour week because its arrival has been flagged well in advance and is not a surprise. But the 1982 experience suggests that cuts in working time are by no means a sure route to higher employment.

This article discusses research reported in ‘Employed 40 Hours or Not-Employed 39: Lessons from the 1982 Mandatory Reduction of the Workweek’, CEPR Discussion Paper No. 2358 (January 2000) by Bruno Crépon and Francis Kramarz. The authors are both at INSEE-CREST; and Kramarz is a Research Fellow in CEPR’s Labour Economics programme.

Download PDF

Return to contents

 Browse Archives