Ever since Pierre Bourdieu stigmatized the “reproduction of elites”, these elites have felt guilty. That their children’s achievements compare to theirs is perceived as a sign of unfair privilege. And prominent members of those elites do not miss an opportunity to publicly complain about “reproduction” and lack of social mobility, even though privately they spare no money, effort, time and connections to lift their progeny as high as possible in the social ladder.
Reproduction is what life is made of. That social structures reproduce themselves should therefore come as no surprise. Parents transmit genetic, human, financial, and social capital to their kids; this is not only a natural “default” outcome but for many people such transmission is the most important purpose in their life.
For Marxists and partisans of “social justice”, this is unfair because you do not choose your parents. Some kids are lucky to be born in an educated, wealthy family; others are unlucky.
Traditionally this problem had been corrected by putting in place a public education system which was supposed to give everybody the ability to acquire human capital and to progress in society despite an unfavorable initial environment. This system was based on strict meritocratic criteria and was meant as giving opportunities to those who had the will and capacity to seize them; it was not meant to be evaluated on the basis of statistical data regarding the relative outcomes of various social groups.
The system was deemed fair because it was meritocratic, regardless of its outcomes. If indeed the elites reproduced themselves, this was just tough luck for the non-elites who had been on average incapable of seizing their opportunities. Whether or not the system is fair depends on its design and not on its outcomes.
In the era of political correctness, this perception is no longer tolerated. The system has to deliver “equality of outcome”, otherwise it is considered as biased. Furthermore, any person who would claim that the system is fair could be cornered into admitting that members of those groups who do comparatively worse are less deserving, and from them easily accused of racism, sexism, and so forth. This, despite that it is generally the Marxists, not the conservatives, who insist on categorizing individuals by sex, ethnicity, class and other collective characteristics.
As a result the guilty elites are gradually eroding the meritocratic system that brought them to the top, by introducing arbitrary criteria meant to promote “diversity” (a conveniently vague concept) in the recruitment process for elite schools and positions.
So what does it mean to promote “diversity”? To answer that question, we need to note that the criteria by which the system is being evaluated have changed. In 7th century China, participants in the Mandarinate contest had their exams copied by a bureaucrat, so as to make sure that the graders could not recognize the handwriting of the candidates and indulge in favoritism. In the 21th century West, instead, elite educational institutions boast of the proportions of various “disadvantaged groups” in their recruitment, while relying on increasingly opaque and arbitrary procedures.
The two processes go hand in hand: if I am targeting a given statistical distribution for the personal characteristics of my students, I cannot at the same time abide by strict rules that apply to all individuals equally.
The most transparent I could get is to have a segmented recruitment process, by which there would be a fixed number of slots for each group. In such a situation, though, it would be all too obvious that those who are admitted to school X in capacity of their belonging to some anatomical group, are not in the same category as the others. The equality of outcome agenda would simply defeat itself if it were to use such obvious means. Instead, it has to rely on opaque means in order to preserve the illusion that the preferred groups are thriving in a process which does not systematically favor them, but instead relies on criteria that are supposed to have less of a disparate impact on the disadvantaged.
These techniques range from having an admission meeting in order to demote members of the non-preferred groups who would have made it on meritocratic criteria, so as to make room for members of the preferred groups who would not have made it (up to the point where the statistical targets are met), to getting rid of some parts of an entrance exam on the grounds of their alleged disparate impact, and replace them by tests that leave considerably more discretion to the admission committee.
As an example of the first method, I once briefly participated in an NSF-style body in the French university system which was in charge of allocating an important set of grants. After discussing the academic merits of the candidates and ranking them, we then counted the number of people who resided outside Paris and the number of women. If the result was not deemed acceptable by the president of the jury, then some men and some Parisians were demoted from their ranking and replaced by provincials and women. Since I was very uncomfortable in contributing to a process that I do not approve of, I did not last long in that jury, especially given that the president greeted me and the other members by complaining that there were not enough women in the jury (I guess they appointed me just to let me know). This Darwinian elimination process guarantees that the jury will eventually be mostly made of yes-men (and women) who will never challenge its non-meritocratic criteria.
As an example of the second method, the French elite school Sciences-Po has decided to withdraw its general culture test from its entrance exam, on the grounds that “disadvantaged groups” — like recent immigrants — would perform poorly because their background made them less acquainted with mainstream higher French culture (similarly, Pierre Bourdieu advocated that selection at school should emphasize mathematics, which is less culturally loaded than humanities). There were also talks of getting rid of the English language test, on similar grounds that the disadvantaged groups were less proficient in foreign languages, having fewer opportunities to live and vacation abroad. Somebody must have pointed out that English is used to communicate in the modern professional world, and that maybe, just maybe, social mobility would not improve if the Sciences Po graduates, regardless of their family background, were incapable of speaking English. So the English test was finally maintained, but the general culture exam was suppressed.
Which brings the following interesting question: How long can an elite survive, if it recruits its members so as to get rid of any of the characteristics that make it legitimate as an elite? If these people are not more knowledgeable, more proficient in English, nor better at logical reasoning than the average Joe, on what grounds do they hold privileged positions in society? This is of course exactly the question that the eighteenth century enlightened liberals were asking on the eve of the French Revolution. We may speculate that competition in labor markets will do to educational institutions that abandon meritocracy what the French Revolution did to the aristocratic system.
The new criteria that Sciences Po uses heavily favor those applicants who have an “interesting” and “diverse” profile. The fair exam principles borrowed from the Chinese Mandarinate system were well received in a Catholic country for which salvation is a reward for good actions (the selective exams reward hard work, and all candidates who were admitted had “suffered” in preparing the exam; therefore they tend to believe that their suffering was rewarded). By contrast, recruiting “interesting people” is a neo-Calvinist concept borrowed from U.S. universities. Salvation is now an outcome of pre-destination, not of your actions. In fact Sciences Po is remarkably opaque in disclosing how you become an interesting person, because they do not want people to develop a fake personality in order to make it to the school. As a result of the new system, some 40 % of a class had to take no written exam  and was admitted on the grounds of a bogus motivation letter which was at best written by their parents, and a 20 minute interview on no specific topic.
The important point here, though, is that these loose criteria, while contributing to the goals of apparent equality of outcomes, at the same time provide the oligarchy with considerable discretion in order to co-opt its members. It is very easy to decide that members of influential networks (financial contributors, political acquaintances, colleagues’ children, media pundits…) just happen to have kids whose profile is wonderfully interesting and diverse. After all, nobody can disprove you and it may even be true! It is easy to imagine that a family located at the center of power has more opportunities for a challenging, original and diverse experience than the children of a regular electrical engineer or manager of a medium-size supermarket in some dull provincial city. And, when one compares these boring middle-class people, whose only claim to upward mobility is hard work and academic excellence, to the Chosen who cannot be bothered being asked demonstrating their skills, all talk of the elite reproducing itself suddenly vanishes . One only opens Bourdieu’s grave when it is convenient.
NB:  This ignores specific procedures for foreigners and applicants from “disadvantaged neighborhoods” who also waive any written exam.
 The trick is not to distinguish, in the statistics and in the rhetoric, between relatively high skilled workers earning a fair wage on their human capital, and the actual oligarchy in control of power. The dismantling of meritocracy benefits the latter at the expense of the former.