Hiring Discrimination

I sometimes hear the argument that discrimination in hiring can’t exist because markets are too competitive. That seems rather spurious to me. It apparently attempts to invoke the efficient markets hypothesis, which would imply that the level of discrimination is optimal, not that it’s zero. In other words, it says that further reduction in discrimination would not pass a cost benefit analysis by these companies. It doesn’t say that discrimination is low in an absolute sense, let alone zero, nor that any part of the hiring process works well, let alone optimally. Dan Luu has an excellent post on that here.

To the extent that EMH is valid here, it means that either the current amount of discrimination is too small to address, or the cost of addressing it is too high relative to the benefit (to the companies, not the job applicants). To believe the former with any level of certainty, we would have to have cite some quantitative data. If it’s the latter, then it’s really a technological problem.

It’s bizarre for a tech CEO to believe that there’s an area in which technology simply cannot be advanced at all! Even moreso in an area where so little effort is made to track or improve performance. Again, you’d need to back that up with some data.

In many tech companies, most of the hiring/interview process is carried out by regular engineers and engineering managers. They rarely have any training on how to interview, the results are not tracked over the long term, and lessons are not learned and pushed back down to the interviewers in order to improve the process. There’s little reason to believe this is especially efficient at spotting and filtering for talent. Dan Luu has another relevant post here.

Perhaps it’s some glib belief in efficiency that allows companies to be so complacent about improving their hiring process? After all, if there were actually a better way, someone else would have found it, and would now be eating our lunch. The fact that we aren’t drowning is evidence that we are just as good at it as everyone else, therefore the market is efficient, and we have no need to improve.

Is there a name for this, where the belief that everything is fine leads to everything not being fine? Mid-Panglossianism? A self-defeating prophecy?

It might also be that companies believe there could be a better way but that if they discover it, others will quickly copy them and their advantage will be nullified. This is the arms race view, where the best way to avoid losing is to just not play. But that, too, is quite different from believing that there is zero discrimination.

To pile on a bit more: In what other area would any respectable tech company be so complacent as this? Does Meta assert that since social networking is so important to so many companies, the current state of the art must be efficient enough that no further development is warranted? Ridiculous!