Letter from France — ‘Paris brûle mais Marseille est calme’

Alan Kirman, GREQAM Marseille, is currently enjoying a spell as Richard B Gibson member of the Institute for Advanced Study Princeton. In this contribution, he reflects on recent disturbances in France.

Sitting peacefully in Princeton, I was reflecting on the appropriate subject for this year’s letter, when I saw that the front page photo in the New York Times was of cars burning in the streets in the Paris banlieux. Despite the suspicion that these events were being overplayed by the US media, they were undoubtedly the worst troubles since 1968 and lasted over two weeks with over 8000 vehicles burned and a number of public buildings, schools in particular, damaged and even destroyed. Now that calm has returned at the price of invoking an emergency law from the time of Algerian independence, it is worth reflecting on the origins and consequences of these riots.

Paradoxically, when I was a graduate student here, in 1967, there were major riots in several American cities. Detroit is the outstanding example, and a number of historians have drawn parallels between those events and the current riots in France. Both the similarities and the differences have been widely discussed. This has not been well received in the French press which feels that France has nothing to learn from the US.

Parallels with the USA
Tom Sugrue (2005) a historian and sociologist, argues that the parallels between the 60s in the US and the riots in France are uncomfortably close. He depicts the riots in Detroit and other Northern US cities as involving, to a large extent, second and third generation black immigrants from the South, just as the current riots in France involved the children and grandchildren of immigrants mainly from the Maghreb and from West Africa. At the time in the US there was the feeling that the idea of treating individuals as equal regardless of race or colour would triumph over the sectarian ideas that were so widespread. However, as Sugrue points out, the notion of universal colour-blindness evoked by Martin Luther King was not enough to eliminate the real and experienced discrimination that confined blacks to poor and over crowded housing nor to prevent them from experiencing an unemployment rate twice that of the local average. As he says, ‘Colourblindness as a principle did not eliminate the practice of discrimination or the institutionalisation of racial and ethnic difference in labour and housing markets.’ It will not help he argues, in France, to continue the denial of these differences in the interest of maintaining the myth of ‘universalism’ which characterizes much of French thought. In particular, he argues, the official policy of not gathering statistics on ethnicity is harmful. For, without such statistics the existence of discrimination is difficult to demonstrate and the measuring of its consequences impossible.

Joan Scott (2005) a prominent historian who has recently written a book on equality of opportunity in France, deals tellingly with the differences between the US and France. One thing that she brings out is the asymmetry of the position of the political parties. In France the defence of ‘universalism’, the refusal to accept that groups within a nation have separate identities, is essentially associated with the left. The idea is that the individual is an abstraction and that any attempt to give him an identity means that he is no longer equal before the law and is, in particular, vulnerable to persecution. The opposite association can be seen in the US where the left has typically been associated with the identifying of differences and policies to remedy the inequalities between groups and where the right maintains that any attempt to classify and favour groups is inequitable and penalises the majority.

However, the universalist position has suffered a blow in France with the recent riots and the reason is not difficult to identify. As Scott points out the rioters were almost universally referred to as ‘immigrés’ although very few, if any, of them actually were immigrants. This identification is reinforced by the recourse to an old law from 1955 at the time of the Algerian crisis which allows the deportation of foreigners guilty of crimes in France. This, the government has said, will be invoked to remove those ‘illegal’ immigrants who committed crimes during the riots. The not very subtle implication is that those responsible for the riots were ‘immigrants’ and some of these did not have the proper papers. The facts are different and can be appreciated by looking at the identity of those arrested. They were overwhelmingly French citizens and often second or third generation French. The fact that they are referred to as ‘immigrants’ reveals the emptiness of the claim of universal equality. Furthermore, it reveals the idea that it is the fault of these people who have not been able to assimilate properly the universal republican values that are associated with being French. Yet, as Alain Touraine has suggested, ‘It is no longer acceptable to think and act as if France was the trustee of Universal Values and, in that role, has the right to treat as inferior all those who do not correspond to this ideal… that the French have created for themselves’.

Here we come to the heart of the problem, the rioters feel that they are excluded from French society because of their identity but French society denies that identity and worse, suggests that, even were it to exist, it should be eliminated.

Questioning the universalist ideal
Joan Scott, suggests that the very French notion of universalism will, and probably should be put into question. This would suggest that the demand that immigrants or those of ‘immigrant origin’ should be fully integrated or should show that they have become enracinés, (rooted in French society) should be weakened. In a sense people whose identity results in discrimination should be allowed to have their identity recognised and used to their advantage. This change is opposed by the left, on the grounds that it will lead to more and easier negative discrimination. It is also rejected by the right who argue that if there is a problem it is the fault of those of ‘immigrant origin’ who are not willing to integrate. The latter is contested by some sociologists (see e.g., Brouard and Tiber (2005)) who suggest that, from the point of view of those who protest, the subjective integration has already happened. The only problem is not that the population involved does not want to be fully integrated but rather sees itself as being prevented from doing so. In this view, were the doors somehow to be opened, the problem would disappear.

Indeed, the official police report on the riots says that the identity that united the rioters was based as much on the feeling that they were being excluded from French society as on their social or ethnic origins. The youngsters involved in pillaging have, the report says, the impression that they were penalised by the absence of any future, by the impossibility of becoming involved in the labour force and quite simply by their poverty, the colour of their skins and their names. Everything, it goes on to say, happened as if any confidence that the rioters had had in national institutions and even in the private sector which many view with jealousy but, aspire to as a source of economic integration had been lost.

Why riots?
We know too little about the dynamics of riots which, after all should be of interest to economists since they can cause significant disruption and cost. In the case in question the damage is estimated at around 250 million euros. What causes a protest to become a riot is a complex problem. The French authorities seem to have real difficulties themselves in understanding the evolution of these events. Nicolas Sarkozy, the minister of the interior and obviously, from his name, of immigrant origin, said several days after the beginning of the riots that they were ‘in no way spontaneous’ and were ‘perfectly organised’. The official police report published several weeks after the events and mentioned above, said that France had experienced a non-organised insurrection, that Islam and Islamic organisations had played no part in it and there was no linkage between the cities involved.

For a riot to develop there has to be some sort of solidarity between the participants, some sort of common identity. The police report suggests that the solidarity that the young who took part in the riots felt was due to the feeling of having been penalised by their names, the colour of their skins and by their poverty. Those it says, who ransacked the cities were those who had no prospects and no investment through their work in the French system. This has led many to suggest that those involved had already abandoned the system and that they were collectively doing what they had done as individuals before. If this were so it would show up in the statistics concerning the participants. If this was a correct view of what was going on, one would expect to find many repeat offenders, some who had taken up illegal activities, ‘faute de mieux’ and some who had fallen into the hands of Islamic extremists.

Yet, all of this is contradicted by the official figures. Most of the minors who appeared before the judges had no previous police record which seems to contradict the assertion by Nicolas Sarkozy, the Minister for the Interior, that the individuals involved were the ‘racaille’ or scum of the generation. No known drug dealer nor any known militant Islamist was taken into custody. Neither Sarkozy’s remarks nor the policy adopted towards first time offenders seem to have been oriented towards calming the situation. From the 29th of October till the 18th of November, 3,101 individuals were detained as a direct result of the urban riots, 562 adults were convicted of whom 422 were given prison sentences and 577 minors were brought before childrens’ courts and of these 118 were held in custody. These are very high proportions and the magistrates involved have been seriously criticized for the over-rapid justice meted out. Cited in evidence of this is a circular dated November 9th, shortly after the beginning of the riots, from the Ministry of the Interior, exhorting judges to be firm. Human rights groups claim that this sort of procedure will only serve to exacerbate tensions and will alienate a group that was previously not involved in any sort of criminal activity. This should be coupled with the fact that the accused had no idea of the mechanics of judicial procedure and, in large part, considered that their appeal could not succeed. However, being condemned has drastic consequences, for getting a job, once they have a police record, is well-nigh impossible. Thus, a considerable number of individuals were pushed into the camp of the truly excluded.

The fact that there were so many for whom this was their first brush with the law, suggests that the riots took on a life of their own and attracted many who alone would not have contemplated any sort of violence against public or private property. Of course, this does not contradict the fact that there was deep-seated dissatisfaction with the system and that this was a combination of racial, social and economic problems. The French system and its more ‘enlightened’ members would, despite all that has been said, still like to deny that the basic problem is racial. They would suggest that it was indeed social but this is ill defined and they would argue that it was certainly economic. So, let me look at this seemingly least controversial aspect, the economic problem.

The simple economic explanation
To what extent is the simplistic analysis favoured by many economists justified? This consists in saying that the rioters are the disaffected with no real hope of employment and they have just become desperate. To paraphrase The Economist giving these people jobs and homes would solve the problem, (shades of Marie Antoinette). A few figures suggest that the correlation between unemployment figures and the places in which the worst riots took place is far from simple. In Toulouse unemployment has steadily declined from nearly 13 per cent in 1998 to 10.5 per cent today. St Denis is bad, having started in 1998 with 14.6 per cent and settled at 14.1 per cent today despite having fallen to 11 per cent in 2000. Lyon started in 1998 with 10.5 per cent went down to around 7 per cent in 2000 and has climbed back to 9.5 per cent today, still below the national average. Strasbourg has moved from 8.5 per cent in 1998 to 10.2 per cent today. Marseille has gone from 19 per cent in 1998 to 14.1 per cent today, a big improvement but still far worse than the national average. Ranking the cities by their unemployment levels or trends would be a poor indicator of the severity of the riots. Now, of course, the important figures are those affecting the areas of these cities with large low-income housing developments. The figures however suggest that almost uniformly, the unemployment, in these poor areas is roughly twice the overall figure for a given city. So there is little in these figures to explain why the rioting was much worse in Toulouse and in Lyon than in Marseille which was hardly affected.

As usual the reality is more complex. Let me take the example of Marseille. A number of explanations for the relative calm in Marseille have been advanced. One argument is that Marseille has a well developed ‘parallel economy’ and the local figures who operate it did not want it disturbed by massive police intervention. Another more optimistic view is that advanced by people such as Claire Duport, a sociologist. She points out that a very highly developed network of public sporting and social facilities has been created in Marseille and that this investment has been productive. She notes that in the four northern ‘arrondissements’ of Marseille which number, in total, 225,000 inhabitants there are seventy two public facilities of this type. In Toulouse in the Murail district where there are 50,000 inhabitants there are only two such facilities. One for 3000 people in Marseille, one for 25,000 in Toulouse. As a result the administration and the running of the activities in these public facilities has created more than 3,500 jobs in the Northern districts of Marseille, which amounts to 4 per cent of the active population and 25 per cent of the tertiary sector employment. 74 per cent of the jobs in these ‘socio-culturelles’ facilities are given to ‘animateurs’ who run the courses and sports that use the infrastructure. Almost without exception these jobs are held by people aged between 17 and 24. On this point it is worth noting that the number of these jobs has increased by 661 per cent over the last twenty years.

Now add to this the fact that a ‘zone franche’, which carries large tax advantages, has been set up in the Northern part of Marseille and that some 200 firms have set up in consequence. The overall job creation has, it is claimed, reduced the unemployment level from 37 per cent to 16 per cent in the last few years. All of this seems to offer some guidelines as to how to reduce the tensions in large cities but the most important problem remains the barriers to entry to the labour market. Here the French system is characterised by an astonishing adherence to specific diplomas for specific jobs. This is supposed to level the playing field and insure that there is no favouritism. Yet for this to work there must be equality of access to education and this brings me to the policies that have been adopted with a view to improving the educational performance of the underprivileged.

The role of education
The most important effort to combat the handicap experienced by young people in the poorer areas in France has been the creation of the so-called ZEP, (zones d’éducation prioritaire). The idea is rather simple, to increase the percentage of those who manage to leave school with some sort of diploma, the possession of such a diploma being a sine qua non in France for obtaining a job. The basic feature of the ZEP is that they are entitled to funds over and above those given to other schools. To what extent has this policy been successful? There have been two apparently contrasting studies by prominent economists. Roland Benabou and his co-authors, (Benabou et al. (2005)) made a careful study of the effects of the ZEP on performance in school exams and came to the depressing conclusion that they had essentially no impact on the performance of the children in question. Notice that this study separates out carefully the selection bias due to the nature of the population in the ZEP before the latter were created. This allows them to get an accurate estimate of the ZEP effect. The authors show three things. Firstly, although the total investments in the ZEP were significant they were thinly spread and little went into concrete measures to improve quality such as a reduction in class size. Secondly, much but apparently not enough went to bonuses to teachers to teach in the ZEP. This was intended to reduce the mobility of ZEP teachers and stabilize the teaching population. It did not succeed. Thirdly the examination performance of the pupils was not improved.

In apparent contrast to this rather negative judgement, Thomas Piketty (2004), also in a careful econometric study, arrived at the conclusion that the smaller class-size at schools in ZEP even though the difference from non-ZEP schools is minimal, had a significant effect on exam results. His results suggest that a reduction of class size from 22 to 18 could reduce the discrepancy between ZEP and non ZEP performance by 40 per cent. The current size of classes in a ZEP is on average 22 as opposed to 23 in non-ZEP but Piketty estimates that even this differential has reduced the gap in performance between ZEP and non-ZEP by 10 per cent.

How can these two apparently contradictory results be reconciled? Firstly, they are based on data for different periods. Piketty used a data set from one year, 1997, whilst Benabou et al. used data from a five year period from 1992 to 1997. Secondly, whilst Benabou et al. used data from colleges, (the equivalent of American junior high schools), Piketty looked at primary schools. One has the impression that the effect of putting resources into a ZEP is not monotonic but rather more U—shaped. The initial effect of being declared a ZEP is to frighten off parents of children who might have gone to schools in the area. This may be through the choice of residence in the case of new arrivals or through manipulation of the mechanism which allocates children from certain areas to certain schools. Thus a simple ‘announcement effect’ may worsen the performance of these schools. However, if significant funding is actually translated into concrete measures then the Piketty effect can kick in and these schools become more, rather than less, attractive. In other words the worst of all possible worlds is to be declared a ZEP and to receive only very limited additional funding. In fact the funds involved have to be substantial to achieve even a one child reduction in the size of an average class. Another alternative would be to increase the average size of classes outside the ZEP to 24.2 and this would be enough to obtain the reduction to 18 within the ZEP. There are now nearly 700 ZEP in France and the schools concerned take over 12 per cent (nearly 18 per cent if we count the priority networks, REP, as well) of children of primary school age and over 15 per cent, (20 per cent with REP), of older children. These numbers reflect the politics, some ZEP have only 44 per cent of children from poor backgrounds and are situated in the bucolic heartland of France, others are in the poor areas of large cities and have over 80 per cent of such children. Were the criterion of ethnicity to be permitted such discrepancies would be even more evident.

What might be done to improve the situation? Firstly, it seems clear that the best results are obtained by investing resources in very young children, and this bears out the results of Jim Heckman who came to this conclusion for the US. Secondly, it would be best to concentrate the resources on the really needy rather than scatter them thinly across the educational landscape and indeed Benabou et al. make that point. However, this is not politically feasible in France. The creation of the ZEP was fiercely opposed when it happened in 1982. Yet to move even one zone from ZEP to non-ZEP would create an outcry today. This sort of reform stickiness is characteristic of the evolution of the French system. The solution proposed by the current government is to have a three tier system of ZEP with the highest priority ZEPs receiving the bulk of the resources. In addition, 1000 new teachers are to be appointed to the ZEP and apparently these jobs are to be concentrated in the high priority areas, and this should help reduce class size. How the ranking of the ZEP is to be done has not been made explicit, and it is unlikely that it will be as radical as the Dutch system which awards points to schools for each pupil, with those from poor or underprivileged backgrounds counting for more.

The reform has been greeted by protests from many seeing this as a convenient way for the government to reduce the overall credits given to the ZEP. In fact, one of the conclusions that might be drawn from the negative results of Benabou et al. is that the government simply did not put enough resources into the ZEP for them to have any effect. Worse, as the number of ZEP expanded and funds did not increase in parallel it could be argued that they became even less effective. Perhaps more to the point as those authors remark is to make the distinction between being labelled ZEP and what is actually done with the resources made available. They see neither the label nor the use to which the funds have been put as having had much effect.

It is difficult to believe that France will be able to maintain its stance on identity and at some point there will have to be explicit recognition of the need to help racial minorities directly. Although this goes against the grain of French political thought and might seem to play into the hands of the national ‘purists’ such as Le Pen, in the end it seems to be the only solution to the vicious circle in which people feel themselves excluded by their ethnicity but have no means to fight against that discrimination. The rigid insistence on a certain type of diploma for each type of job might seem to be egalitarian and just but, in fact, simply succeeds in preserving the status quo, in spite of the extra investments by the government in education. Recognising the ethnicity of school populations would provide one way of insuring that funds get channeled to where they are really needed. Another simple channel for positive discrimination would be to reduce the perceived tension between the individuals and the state by increasing the number of police officers whose family’s origins were North African or West African. In a number of countries this has proved to be effective and this together with the election of at least one or two members of parliament with these origins would surely provide a strong signal of non-exclusion. The only way to achieve the second in a democracy with fairly large geographical constituencies would be for the major parties themselves to do some positive discrimination. As long as they deny the need for this it will not happen.


Bénabou R, Francis Kramarz and Corinne Prost (2005) ‘Zones d’éducation prioritaire: quels moyens pour quels résultats ?: Une évaluation sur la période 1982-1992’. Economie et Statistique, N.° 380, pp.3-34

Brouard S and V Tiber (2005), Français comme les autres?, Paris: Presses de la Fondation Nationale des Sciences Politiques.

Piketty T (2004) ‘L’Impact de la taille des classes et de la ségrégation sociale sur la réussite scolaire dans les écoles françaises : une estimation à partir du panel primaire 1997’, mimeo DELTA, EHESS, Paris.

Page Options