In the aftermath of the French riots, French politicians of all hues quickly went back to business-as-usual: lofty talk on the need for a better “intégration au sein du pacte républicain” (whatever that means), plus more or less open attempts to shift blame, or even to blame fantasy figures such as the uncounted (and therefore no doubt uncountable) polygamous families living in crowded quarters.

As an economist, I will focus on what appears to me to be one of the root causes of the riots, namely the obscenely low employment rate of the young in France, and especially of low-skilled young men. While the rioters should not be categorized as “unemployed young foreigners”—a majority of them were French, and too young to be even looking for a job—it seems clear that they were reflecting perspectives that they saw around them. And while other approaches are clearly necessary to explain their rage, economics has an important role to play in explaining how so few unskilled young men can get a job in France.

Consider the following table (in French, avec toutes mes excuses), which uses OECD standardized data to compare France and the US. In the 25-54 age group, the employment rates of men in both countries have stayed very close; they have diverged a little for women, and the employment rate for people over 55 has dropped spectacularly in France. But it is the precipitous decline in the employment rate of young men that is the most striking feature of this table: 30 points, as opposed to…2 points in the US. Even if one focusses on the older half, the 20-24 age group, the difference remains striking: only 40% of these young men are employed in France, as compared to roughly 90% in the 30-49 age group.

The (relatively) generous welfare system in France probably played a limited role in the decrease in youth employment: its most discussed feature, the minimum income guarantee, may induce inactivity traps, but it does not cover the under-25. Of course, it may still have some indirect effects in inducing parents to stay away from employment; but it was only legislated in 1991, at a time when the unemployment rate of the young had already reached 20%. A spate of empirical studies, including those I have carried out with Guy Laroque1Some of our results are collected in our book, Institutions et emploi (Economica, 1999), where the empirical illustrations concern the female labor market., have shown that the labor demand problem is much more severe than anything linked to labor supply for this segment of the population. And the cost of the minimum wage, combined with a rigid education system that has not found a good way to accomodate the needs of the children of immigrants, is the main culprit.

Before I take on my mandate to give my opinion of the causes and possible remedies to this dreadful situation, I would like to comment on a statement made by Ezra Suleiman on this forum. Professor Suleiman deplores the paucity of well-informed analysis on the quartiers sensibles. It is true that these issues have not given rise to anything that approximates the monumental literature on minorities in the US; but a lot of data exists, and economists have started to use it. It is not even hard to find, as I will show in an example.

While I now teach economics in New York, I spent more than fifteen years working as a statistician in Paris; and during the riots, I was struck by the fact that most commentators were focussing on a 40% figure for youth unemployment in these quartiers. I am naturally skeptical of such a consensus, especially when it comes without a precise source or even a definition of the age group concerned. I also know that the unemployment rate is a very rough measure of the situation of these young people on the labor market, in particular because many of them are pursuing studies and others are not actively looking for work—not to mention the very large heterogeneity in the quartiers.

As it turns out, this 40% figure was attributed to a position paper written by a (respectable) CEO for the Institut Montaigne, a (respectable again) discussion group created by Claude Bébéar, one of the most influential businessmen in France. It is important to know exactly what this figure measures. The Institut Montaigne paper used the 1999 Census to compute the unemployment rate for the 15-24 in the ZUS (“zones urbaines sensibles”), a definition of the quartiers sensibles that is only loosely linked to objective criteria. Still, it is striking that about 7% of the total French population lives in neighborhoods where so many young men and women cannot get a job. And the map of the ZUS is rather reminiscent of that of the riots, so focusing on the ZUS does provide a useful starting point:

There is in fact much more descriptive statistical information in the 1999 Census, if one is willing to go to the Web site of Insee (http://www.recensement.insee .fr/RP99), a little quirky at times but a very rich source of data); and while it was carried out six years ago, the situation has probably not changed very much since then in its main structural features. A 2002 report2“Ségrégation urbaine et intégration sociale”, by Jean-Paul Fitoussi, Eloi Laurent and Joel Maurice. It is available (in French) at to the Conseil d’Analyse Economique, a sort of “wise men” advisory body set up in 1997 to inject economic expertise into the government decision-making process, used the census data to provide a fairly detailed picture of these areas. It compares in particular the characteristics of the ZUS to that of the urban areas to which they belong. The most striking results are that:

  • the percentage of foreigners is about twice as high (16% vs 8%, and less than 6% in the whole of France)
  • yet the average household is only slightly larger (2.6 members, vs 2.3)—not very strong evidence of polygamous, multi-children families
  • the total unemployment rate is much higher (25% vs 14%—10% in France)
  • so is the youth unemployment rate (40% indeed, vs 27%—20% in France).

The vexing issue of employment discrimination was at the center of the riots. The CAE report also examines the (few) studies that have tried to quantify it. The difficulty of course is to control for productivity, loosely defined. For instance, the report cites the fact that young Algerians who spent time in college are four times less likely to have an executive job than comparable French men; but we do not know how much of this difference remains after controlling for diploma at a finer level. It is particularly important in the context of the ZUS, since they concentrate the working-class and the low-education group. Indeed, in the 1999 Census 37.5% of workers living in a ZUS were blue-collar workers, as against 22.9% in their surrounding urban areas—a worrying signal in itself since blue-collar jobs are declining in France as in other rich countries. Moreover, the share of the population without any diploma, at 33% in the ZUS, is about twice higher than in the rest of France.

Given these observed differences, and controlling for the unobserved differences as best they could, the studies cited in the report give the impression that other things being equal, it is indeed more difficult for teenagers born from immigrants to obtain and keep a job—especially for children of immigrants from Africa (Maghreb included), and especially when they live in a ZUS. The difference is smaller than what one would expect, though: the order of magnitude is an increase in expected duration of unemployment of 10 to 20%.

This is of course not the end of the story. Employment outcomes are highly linked to education in France, as the following table shows—I replaced the French diploma names with rough US equivalents:

This is hardly surprising. In my work with Guy Laroque, I have argued that the high cost of the minimum wage for employers is a high hurdle for worker groups with low observable productivity characteristics (most notably education)—even higher of course if they are discriminated against. Consider the following figure, which shows the evolution of the ratio of the cost of the minimum hourly wage for an employer to the cost of the median hourly wage. While this may sound arcane, it is a natural and simple measure of the impact of the minimum wage on the French labor force.

In the United States, this ratio is about 40%. In France, it was high when the minimum wage was introduced in 1950; but inflation reduced its value considerably, to about 40%. After the wave of strikes in 1968, the value of the minimum wage was increased by 25%; and indexation on both inflation and on some measure of workers’ purchasing power was introduced. When the oil shocks came, the government pursued a policy of increasing welfare payments and therefore payroll taxes, along with occasional discretionary increases in the minimum wage (10% in 1981 for instance, when the Socialists came to power). As a consequence, unemployment increased sharply for the least-protected groups, and the unskilled young in the first place.

In the mid-eighties, the ratio peaked at 63%, more than 50% higher than in the United States. This coincides with a major change in economic policy in France, with more traditional concerns taking center stage. Starting in 1993, governments adopted a policy of reducing the payroll taxes on low wages; this contributed to a sharp decrease in our ratio—while preserving the net purchasing power of the minimum wage.

Unfortunately, the imposition of a 35-hour week by the Jospin government in 1997 had a very negative side-effect: since it was not possible to reduce the monthly wage of minimum wage-earners by 10% (the decrease in hours), the value of the hourly minimum wage had to increase proportionately… this, along with two discretionary increases in 1995 and 1997, explains why the ratio turned up, and a policy that was demonstrably showing sizable positive effects was jeopardized.

Given the striking decrease of the incidence of unemployment with education, the most important discrimination is that the young who live in the ZUS have a high risk of ending up on the wrong side of the education divide. The evidence is overwhelming here. An oft-quoted (and reliable) figure is that 90% of all children of immigrants in France study in only 10% of all French schools—and a lot of them are in the ZUS.

Despite the French obsession with equality of treatment (as opposed to equality of opportunity, a concept that is only now being seriously discussed), successive governments have tried to shift financial resources to schools that needed them most. Of every five teenagers who live in France, one now is in a school that is subsidized in some way, mostly through the ZEP (zone d’éducation prioritaire) program. More than 95% of all teenagers who live in a ZUS attend a school in a ZEP. This policy has been rather timid, though, as compared to charter schools in Britain, for instance: all schools are still rigidly controlled from the Ministry of Education in Paris; and the government only provides 10% more funding for a ZEP teenager. The few studies there are do not suggest that the return on this investment has been very high. A more drastic break with the standard practices seems to be in order; the Villepin government has announced a doubling of the ZEP money, which is welcome but must be followed by further increases.

Affirmative action à la francaise (the definition is still hazy) has been much discussed lately, unfortunately often on a polemical note. The American experience shows that such a policy can count notable successes (at least part of the emergence of a Black middle-class can be attributed to it), as well as some abuses. It also suggests that there is a very strong potential for a backlash—especially in France: with such a high unemployment for the young, French parents may not easily accept seeing their children crowded out of the race.

A further possible policy tool would be to try to stimulate job creation in these areas. Unfortunately, the zones franches (tax-free zones) policy initiated by the Juppé government in the mid-90s has been rather disappointing. Once again, evaluations are few; but they suggest that less than 10,000 (fixed-duration) jobs may have been created, at a high budgetary cost. Some lessons have been learned, and the program is now better focussed and presumably more efficient. But in any case, these policies seem to be more directed towards the symptoms than towards the causes; they satisfy the public’s cry for immediate action, but they do not attack the roots of the disease.

To conclude, let me give some figures drawn from the 1999 Census that concern the small town of Clichy-sous-Bois, where the riots erupted (again, these figures are available on the Web site of Insee: at such a local level, they of course give less detail than one would like). A big chunk of Clichy-sous-Bois is a ZUS. Of its 28,200 inhabitants of all ages, only 9,000 have a job. About 2,000 work in the public sector—high by American standards, but close to the French average; 4,900 work in private services, 1,200 in manufacturing. These jobs are unlikely to be in Clichy-sous-Bois, since only 4,000 persons work there in all. No fewer than 2,800 are unemployed, a very high figure if one compares it to the population of age 20-59, which numbers 7,200 men and 7,600 women.

About a third of the Clichéens are foreigners (9,300); many others probably are children of earlier immigrants, and became French at birth by jus soli. It is hard to know how many, since Insee is not allowed to ask the nationality of the parents: once a person becomes French, his or her past is supposed to vanish in a blissful mist. The non-student population is much less well educated than the average population of France:

The depiction in some American media of the young rioters as “Les Misérables” was somewhat exaggerated. But the contrast with Paris proper, where housing is more and more expensive and working-class inhabitants are more and more the exception, is striking given that Notre-Dame is only about ten miles from the Clichy-sous-Bois ZUS. In the 19th century, the bourgeois used to fret that the workers were “coming down from Belleville”, which is now within Paris. It is to be hoped that the “toiling classes, dangerous classes” attitude of the élite then does not turn into a “unemployed classes, dangerous classes” attitude that would adjourn any serious consideration of the strong remedies that French society must adopt. These, I have argued, must comprise a tenacious continuation of the policy of subsidizing low-skilled employment, along with policies that give children of immigrants a better chance of getting the diploma that still remains the key to employment in France.

Bernard Salanié is a professor of economics at Columbia University and Ecole Polytechnique. His latest book is The Economics of Taxation, MIT Press, 2003.


Some of our results are collected in our book, Institutions et emploi (Economica, 1999), where the empirical illustrations concern the female labor market.
“Ségrégation urbaine et intégration sociale”, by Jean-Paul Fitoussi, Eloi Laurent and Joel Maurice. It is available (in French) at