Thanks for this nice analysis. But I don't really understand why this result is shocking or alarming. Let's take the exmple of primary school completion. Assume an (I would say realistic) situation where children don't complete primary school due to factors correlated with income, and family income needs to be above some threshold for children to complete school. In this case, if income is rising equally among all quintiles, you would expect the richer students to have higher growth in completion rates, as they are closer to the threshold. This of course only holds for families that start below the threshold, but in countries that start with very low education rates, this could mean large growth among students in the upper quintiles. If this effect is driving the patterns above in some countries, it doesn't really seem like something that needs to be fixed.