That leaves two hypothesis: population evolution and cultural development.
Populations do evolve over long periods of time. We know that from physical differences in populations. Japanese women have smaller breasts than their Korean neighbors. Both groups have soft earwax and teeth that are larger than their jaws can handle. Many groups in the Asian geographic area have red spots on their lower spine as infants as do many American tribal natives who migrated from Asia. These suggest evolutionary practices over long periods.They do not suggest any way that higher IQ groups would be reproductively favored over lower IQ groups.
I cannot take the thinking any further and remain rigorous.
I know that the two cultures I have followed intimately, Japanese and Jews, have cultural attributes that seem to explain their resulting higher IQ testing. In Japanese culture intelligence, particularly logic and math skills are rewarded by selective adoption by higher status families. That adoption favors and rewards IQ attributes. Such family behavior in Japan is highly meritocratic and can resemble corporate behavior over centuries.
Jews have a high reward and prestige system in place for logical argument skills, linguistic talent and entrepreneurial facility.
That is what leads me to suggest cultural development as the source of ethnic differences in IQ tests.