Discrepancy Does Not Imply Discrimination

In almost any discussion of political and social issues, someone will use the argument that a discrepancy between two groups is per se evidence of discrimination against one group. However, statistical analysis reveals that the latter is not necessarily true.

For example, it is true that women make less than men, but women also tend to go into professions that pay less than men. These professions may offer more flexibility along other margins than women prefer (e.g., more time with family, more safety, etc). Once adjusting for these factors, the pay gap almost disappears.

Could discrimination still be going on against women? Absolutely. For example:

• There may be social pressures that discourage women from pursuing certain professions. For example, the idea that women should be homemakers and men should be breadwinners encourages women to take on more flexible and secure work.
• There could also be phenomena like “tracking” by academic advisors: when my mother was growing up in the 60s, the school counselors would put each student on a “track” that would determine what classes they would take. Girls were exclusively tracked along fashion, home ec, and education, while boys tended to be tracked along science, mathematics, and athletics.
• There may also be unpleasant working conditions that discourage women from going into certain professions. I have seen students (both male and female) treat female professors horrifically compared to male professors.

In each of these cases, wages might be the same once adjusted for certain variables, but they are instances of discrimination.

This tells us that mere discrepancy is not sufficient evidence of discrimination in and of itself. It can only suggest that we gather more evidence.

The same holds in politics. Lately there have been claims that the unusual number of nationwide injunctions against executive orders are per se evidence of “lawfare” against the current administration, arguing that there is a judicial coup going on. However, for the same reason as just discussed, this argument needs more evidence.

There have been an unusual number of nationwide injunctions against the Trump administrations compared to other presidents. In his first term, there were some 84 injunctions issued against the administration, compared with 28 in the Biden administration, 12 in the Obama administration, and 6 under George W Bush’s administration. In the current term, there have been approximately 25 (source for these numbers).

But it is also true that President Trump has issued an unusually high number of executive orders. In his first term, he issued 220 orders, an average of 55 per year. That is the highest since Jimmy Carter. If the current administration keeps the pace set so far, it will issue an average of 1,312 orders per year. That’s the highest since FDR’s first term (source). With so many more executive orders, just by random chance we should expect more nationwide injunctions. Furthermore, many orders are extraordinarily broad and deal with topics that affect every American. The broad effects of these orders mean more people are likely to be affected, also increasing the odds of nationwide injunctions in response.

Discrimination is extraordinarily hard to prove. Basically, you’d need to show that an identical executive order before an identical judge would have gone the other way if a different administration had issued the order. That is hard to manage (albeit not impossible).

Just like the wage example above, discrimination may be a factor. But a discrepancy does not imply one is there. We need more evidence.

— PS: To head off an anticipated objection: The Supreme Court’s ruling in Trump v CASA does not provide evidence that courts were discriminating. The opinion does not address that issue. In fact, the opinion explicitly notes that the state of affairs for presidents of both parties is “where almost every major presidential act [is] immediately frozen by a federal district court” (see pgs 4–5).