You speaking of rational responses. You likely refer to the rational expectations hypothesis in economics. The default assumption is of course that everyone makes rational decisions, including the poor.

Since they are making decisions that you don't think are rational ("it's probably rare for the optimal solution ..."), you conclude that the reason is that the poor people are less well-informed.

This makes sense, so long as everyone has equal access to the same services, or at least access to the same services you do, and that there is competition among those services.

However, as the author points out, the poor areas are under-served in terms of banking choices, and banks use "private databases like ChexSystems that currently keep more than a million low income Americans from being able to open accounts."

Because of the obvious market inequality, you cannot conclude as you did that the people who use RiteCheck are making either an irrational decision or un-informed decisions.