Savings is used because looking at per capita income is a disingenuous metric. Since making 10 times more than someone doesn't get you very far if you live in

Savings is used because looking at per capita income is a disingenuous metric. Since making 10 times more than someone doesn't get you very far if you live in a place that's 20 times more expensive, savings rate was used as a rough proxy for "how much can an average person set aside for this considering all other factors". You could certainly factor in GINI coefficients and quintile distributions for a more accurate analysis, but I don't think it will change the fundamental nature of the results: The capital a parent earmarks for their child's college education takes more time to accumulate and is more quickly exhausted in the United States than in many other countries. This means that the United States parents may be less able to afford a secondary program for their child in the United States than their counterparts abroad who didn't have to bare the same initial cost burdens.