Thursday, March 26, 2009

Gromov Wins Abel Prize

Mikhail Gromov has won the Abel Prize recognizing superior work in mathematics. The Russia-born Gromov teaches at the Institute des Hautes Etudes Scientifiques in Bures-sur-Yvette. Just another of those "mediocre" French professors Sarkozy was blasting the other day.

Aubry's Papa Backs Fillon and Juppé for EU

Jacques Delors, the father of PS leader Martine Aubry, says he'd like to see François Fillon or Alain Juppé head the European Commission.

Martine had better hope she's a knockout on Drucker's show.

Racial Statistics

I see from the blog logs that a previous brief note on racial statistics has drawn attention from several readers at the Assemblée Nationale. My note merely referred to a lucid and excellent article by Etienne Wasmer. Perhaps the AN readers were more interested in the comment to my note by Jean Granville. I originally felt that M. Granville's comment was best left to speak for itself, without any gloss from me, but since it seems to be attracting readers in high places, perhaps a few remarks are warranted. M. Granville begins by saying:

This is outrageous, pure and simple.

First, Mr Wasner, we French are not the ones who have to "accept the other". It is "the other" who has to make the French accept him first. He is the one who decided to come, nobody forced him. So please, let's put our priorities in order.

Then, Mr Wasner has already decided that the problem to be solved is discrimination.

No. The main problem is an excessive immigration. Immigration is fine just as long as you can assimilate immigrants. When there are too many of them, that becomes impossible and problems occur. It happens to be the case that most recent immigrants are either arab or black, and therefore, if immgigrants don't integrate well, the ethnic statistics will tell us that Blacks and Arabs seem to be pushed aside. And no doubt, racism will be invoked as the primary factor here. Indeed, invoking any other factor will be deemed racist.

I see nothing in Wasmer's remarks to warrant this rejection, not to say "outrage." Wasmer does not speak of "acceptance" but simply of employment. But suppose we consider the issue of "accepting the other," which for M. Granville is paramount. Granville argues that "it is 'the other' who has to make the French accept him first." But what does that entail? Wasmer notes that the unemployment rate for youths with at least one Maghrébin parent is 1.5 times higher than for youths with two "French" parents, even when the level of educational attainment is the same. He doesn't say that "racism" is the only possible cause of this. Indeed, as an economist, he may well be thinking of the standard economic argument that a foreign name or "underprivileged" place of residence is taken by employers as a "signaling" device connoting other characteristics pertinent to the employment decision.

Whatever the cause, the social problem remains. But my question for M. Granville is, "How exactly has 'the other' failed in his duty to 'make the French accept him first?'" He has obtained the same diploma as the French job applicant; he has signaled his readiness to work more in order to earn more by applying for employment. And yet he is turned away at a higher rate than his competitor. Is M. Granville arguing that his preparation for employment is somehow deficient? Perhaps he has attended an inferior school. If so, that is a social problem, a challenge for the government, that more precise social statistics might help to solve. How can one answer the question of whether a high concentration of children of immigrants in a particular school district is a challenge to educators unless one knows where immigrant children are concentrated? Under current law, such judgments must be made by indirect inference. Wouldn't it be better to measure directly?

M. Granville also makes the following point:

And by the way, Jews probably have a better memory of what ethnic statistics can be used for. I don't think it a coincidence that Simone Veil opposed the modification of the constitution's preamble, and she is right, of course.

Having argued that it is best for purposes of lawmaking not to distinguish subgroups of the national population, M. Granville nevertheless finds it useful to distinguish the Jewish subgroup as one possessing a special memory in regard to the collection of "ethnic statistics." Indeed, the fichage of members of this or that category for purposes of exclusion, expropriation, expulsion, and extermination is what no one wishes to see repeated.

But why is this example the one for which opponents of improved statistical data reach immediately whenever the issue is debated? The data to be collected are to be kept anonymous. To ensure that this will be the case, surveys must be carefully vetted and approved. Personal identifying information will be removed from all records. There is simply nothing in common between the kinds of data to be collected by social scientists and the fiches used by Vichy's police. Since I am as Jewish as Mme Veil, I can say with utter confidence that not all Jews share the sentiments that M. Granville imputes to her and, by extension, "them." To say more about "the Jewish position" on the question, I would need to gather statistics using some reputable and recognized social scientific method. To rely on purely anecdotal evidence, as M. Granville does, is to open the door to the imputation of whatever "ideas" the commentator wishes to whatever group he or she chooses to characterize. This practice is precisely what the collection and analysis of data is meant to counter.

Finally, as to M. Granville's point that "excessive immigration" is the problem rather than discrimination, I can agree that it is difficult for any society to assimilate immigrants when the arrival rate exceeds a certain level. But what we don't know in the French case, precisely because we lack adequate statistics, is the degree to which high observed unemployment rates are distributed among different groups issus de l'immigration: Is high unemployment primarily a phenomenon of the first generation? To what extent do the second and third and fourth generations also exhibit differential unemployment and education rates when compared with les Français de souche? How do neighborhood patterns vary with assimilation? None of these questions can be answered by inference from data on last names and parental origin alone. That is why better data are needed. Because one cannot say what rate of immigration is "excessive" until one knows how well previous generations of immigrants have been integrated into the society, and in France, despite the excellent work of authors like Justin Vaïsse and Jonathan Laurence, we still do not have all the information we need.

UPDATE: roundup here.