After doing some searching, it appears that period life expectancy is a perfectly reasonable metric for life expectancy. It looks to be a clever way of capturing a sort of “instantaneous” life expectancy for a given year. In a given year, you can track how many people of each age range die, but that only tells you death rates, not life expectancy. And you can’t just average the ages of everyone living because you don’t know when they’ll die. So what they do is they take the death rates for age ranges within a given year and they calculate the expected value of an infant’s death age, assuming that infant would live through the same death rates that were tracked that year. In my opinion, this seems like a natural choice.

That said, there’s definitely other bullshittery going on.