> I am deliberately applying an individual level test to the social level. This is not fuzzy utopian everything-should-be-perfect idealism though. Think of it as an extensive test suite to run against a government's policies.

OK, but that doesn't answer the question. What's the pass/fail criterion for your test suite? For China to be such a particularly bad case, it seems like you must be focusing on those particular groups and weighting them more highly than in the actual veil of ignorance test (where one assumes an equal chance of being anyone), no?