He seems to hint at some awareness of this in Part II, but doesn't really address it explicitly. It seems more likely to me that the transition was due to existing (but very real) discrimination in the higher-status jobs combined with automation (in the form of compilers) of the highly feminised lower-status jobs.

The author's thesis, if I understand correctly, is that there was some sort of new discrimination that drove women out of the industry. I'm more inclined to think that there was never any particular shortage of discrimination.