Comment by dan-robertson

I think the point is that they are good for comparing different cohorts and different countries, though the article makes some claims about how the scores tend to correlate with other desirable outcomes. Naturally one should try to avoid Goodhart’s law here.