• the best comparative statistic is the effective annual duration of work, excluding holidays and sick leave (the OFCE authors agree)
• the data to base the calculation on should come from the Labour Force Survey, not from employers' official payroll returns
• the population to assess should be that of fulltime salaried workers only, excluding part-time workers and non-salaried workers

The OFCE authors (Eric Heyer and Mathieu Plane) point out that the Labour Force Survey (LFS), based on declarations to national statistics personnel of a sample of the working-age population, though (supposedly) run according to Eurostat standards, notoriously produces variable overstatements of work duration, introducing a systematic bias both in time series and country series, that is not homogenous and cannot be controlled for. (Note by afew: One might also wonder why a think tank close to bosses' unions should not trust employers' official declarations of hours worked...).

The choice of fulltime salaried workers only, say Heyer and Plane, leads to the conclusion that effective working hours are longer in Germany than in France (whether one looks at the LFS or payroll declarations). But the inclusion of part-time workers produces a different conclusion, since short-hours part-time work occupies a much larger place in German employment than French: the LFS still gives a longer annual duration in Germany, but the employers' payroll declarations give the reverse.

Which, though Heyer and Plane don't say it in so many words, points up the cherry-picking in the Rexecode report. Run all the numbers, identify the restrictions and exclusions that allow you to reach the desired conclusion, then – this is Rexecode's "seriousness" guarantee – order up the statistical work from Eurostat according to specs.

Conclusion, insofar as there is one: trans-country comparisons are in fact difficult to establish; there is no clear evidence that the French work less than the Germans, or indeed than a number of other Europeans.

Heyer and Plane then look at Rexecode's contention that working hours have gone down faster over the past decade in France than elsewhere in Europe. Once again, Rexecode use the LFS, and fail to communicate a major problem: French statistics agency INSEE changed methodology in 2003, from an annual to a continuous survey, resulting in a considerable break in continuity.

Dynamic analysis over the whole period is fallacious – yet Rexecode go ahead and count the decline in French working hours from 1998 to 2010 without considering the methodological break. As Heyer and Plane point out, annual working hours have declined much faster in Germany than in France since 2003.

Next on the list: employment. Heyer and Plane take Rexecode's data, based on the LFS (more dependable on jobs than on real work hours) and show that they demonstrate two things: one, that job creation in France has been more dynamic since 1999 than in most major European countries, (twice as dynamic as Germany), and that this creation has mostly been of fulltime jobs – whereas in Germany (and the Netherlands), job creation has been heavily dependent on part-time jobs:

1. The right-hand column shows the contribution to total employment increase, of part-time jobs occupied by people who say (LFS) they really want a fulltime job.
2. The greater than 100% contribution of part-time (second column) in Germany (Allemagne) and the Netherlands (Pays-Bas) reflects a decrease in fulltime jobs (!) of -11.3% in the Netherlands and -2.4% in Germany.
3. Meanwhile, France has created more fulltime posts and depends much less on part-time, and particularly unwanted, part-time.

GDP growth: the Rexecode report does not underline that total GDP growth over the 1999-2010 period was 13.1% for Germany, 16.1% for France – but prefers to concentrate on GDP per capita, +13.5% in Germany, +7.3% in France. Conclusion, German success story.

But the difference lies in the demographics: German total population decreased over the period by 0.4% while French rose by 7.3%. And also (German success story?) in mercantilism:

La période 2006-2010 a surtout été celle de l’arrivée à maturité des politiques de désinflation compétitive mises en place en Allemagne (réformes Hartz, TVA sociale…) dont les effets ont joué à plein de 2006 à 2008. Cette politique allemande a entraîné une baisse des parts de marché des autres grands pays de la zone euro au profit de l’Allemagne. Cette politique « non coopérative » a entraîné une baisse relative du PIB par tête relatif de la France, de l’Espagne et l’Italie. (Heyer et Plane, p.5)

(The 2006-2010 period saw above all the ripening of the competitive deflation policies applied in Germany (Hartz reforms, VAT increase...), the effects of which came fully into play between 2006 and 2008. This German policy resulted in a decrease in market share of the other major eurozone countries in favour of Germany. This "non-cooperative" policy caused a relative fall in GDP per capita in France, Spain and Italy. )

Finally, Rexecode contend that German success is demonstrated by a rise in individual purchasing power, equal to the rise in GDP per capita. Heyer and Plane point out that there is no such equivalence.

Ce raisonnement ne tient pas compte de l’évolution du partage de la valeur ajoutée. Or, la politique allemande, en compressant les coûts salariaux, a permis d’augmenter fortement les taux de marge des entreprises allemandes. Les gains de compétitivité des entreprises allemandes se sont fait au détriment de la dynamique des salaires et du pouvoir d’achat des ménages, ce qui n’a pas été le cas en France où le partage de la valeur ajoutée est resté relativement stable au cours de la dernière décennie.

(This reasoning ignores change in the share of added value. But German policy, by compressing wage costs, enabled a strong increase in German businesses' margins. The competitiveness gains of German businesses were accomplished to the detriment of wage dynamics and household purchasing power, which was not the case in France where the share of added value has remained relatively stable over the past decade.)

In all, a good rebuttal of conservative spin (Coe-Rexecode made a similar effort before the presidential elections in 2007), but more than that an interesting peek beneath the skirts of German "success".