Working time reduction: win-win and possibly win-win-win

Dean Baker counters American National Public Radio attacks on France’s 35-hour week and in doing so gives me an opportunity to write about something other than the interminable euro area crisis (and reflect on whether I shouldn’t try to leave the office at a reasonable hour tonight).

Dean is right. Those asserting that the working time reduction doesn’t create jobs, as its proponents claim, have to think about productivity. If average working time is reduced and there is no employment increase (as NPR and other critics assert), then productivity (specifically: hourly labour productivity) has risen in proportion to the cut in hours, to the extent that output has not fallen. (Dean is surely aware of the role of changes in output, but doesn’t explicitly mention it.)

This result comes from the fact that output can be decomposed into the product of hourly labour productivity, average working hours and the number of people in employment.

This means that introducing, say, a 35-hour in place of a 39-hour week leads to some combination of higher hourly labour productivity, lower output and higher employment. As Dean rightly says, if the only effect is higher productivity, working time reduction can be seen as a dream policy: material living standards and employment are unchanged and every full-time worker gets four hours more a week for living rather than working.

What do we know about the mix of outcomes in the case of the French 35 hour week? Although often conceived as a simple across-the-board reduction of four working hours per week, the policy, introduced via two laws in in 1998 and 2000), was highly complex. (Wikipedia en francais has a very good overview; there is also an English entry, but it is not accessible at the moment, as Wikipedia US is ‘on strike’!)

A considerable number of econometric studies were carried out. The problem is that isolating the effects of the measure is extremely difficult. Assumptions have to be made and behaviour modeled. Combine that with the highly controversial nature of the policy and you get a wide range of estimates that, for some reason, happen to coincide with the priors of the interest groups that commissioned them; again see the Wikipedia entry for details.

Without in any way claiming to supplant such studies, an intuitive guide to what happened in France and might be expected from similar initiatives elsewhere, can be relatively simply gleaned by comparing the development of the relevant productivity, growth and employment variables in France with a benchmark. In what follows I do this for the period 1995-2005, taking the 15 then Member States of the EU (EU15) as the comparator. Comparing the period before the reform we see that France largely moved in lock-step with the EU15. After the reform (which came on stream in phases) we see some interesting differences.

Until 1999 French hourly labour productivity moved in lock-step with that in the EU. From 2000 hourly productivity in France accelerated considerably; by 2005 it was around 2 percentage points (pp) higher.  (Fig. 1 below)

Similarly, output (real GDP) developed in parallel until 2000, when a gap gradually began to open up; by 2000 GDP was just under 2 pp. lower compared to the benchmark of the EU15. (Fig. 2 below)

Meanwhile employment growth was slower in France than in the rest of Europe, both before the reforms and at the end of the period (reflecting not least slower population growth; this is confirmed by looking at French and European employment rates, which follow a similar pattern). However, in the years following the reform French employment growth accelerated past that of  its European counterparts. (Fig. 3 below)

In other words we see exactly the mix of higher productivity, lower output and higher employment (with respect to trend) that was postulated above. You may well think that the gaps in productivity, output and employment are rather small given the 11% (4/39*100) cut in average working hours. The main reason is that average hours did not in fact fall by 11% due to, among other things, the substantial number of workers already working less than 39 hours, exemptions in the law, and the fact that some people continued to work 39 hours and received a higher wage).

Of course it is conceivable that the differential developments in France and the EU shown in the figures are due to another, unrelated development. I am not sure what that could plausibly be, however. (Suggestions anyone?) In the relatively short run the most important changes between countries’ developments result from different paths of aggregate demand, but they would push growth, employment and productivity in the same direction, not the pattern seen here. And of course even the most sophisticated econometric study has to construct some sort of counterfactual to analyse policy effects.

Let us then return to the normative debate around Dean Baker’s critique of NPR’s attack on working time reduction. Higher productivity is pretty uncontroversially a good thing: it is more for less. Higher employment is also generally accepted to be a good thing, at least in the context of high unemployment. That leaves output losses. And that is more tricky. A standard view would be that lower material living standards are a bad thing, and the question would be whether they are more or less offset by the increased leisure for workers and higher employment opportunities. In the context of climate change, however, and the challenge of decoupling emissions, output and welfare, a case can be made that, in wealthy countries, the output loss and associated reduction in emissions – in each case relative to trend, both are still rising in absolute terms but more slowly – is a good thing. It could be part of a successful transition strategy towards a sustainable economic model. (In addition to the quantity, there will also be qualitative changes in output, as workers with more leisure change their demand patterns.)

If you take that normative view – and personally I am inclined to – working time reduction, if intelligently introduced, would appear to offer the potential for a win-win-win strategy.

Reduction by how much? Full or partial wage compensation? Distributional issues? What balance between across-the-board compulsion and flexibility and individual choice?

But I really should try to get off earlier tonight….

Fig. 1 Hourly labour productivity, 1995 = 100

Fig. 2 Output (real GDP), 1995 = 100

Employment, 1995 = 100

No related posts.