Comment by friendzis

If that was the supposed point of the excerpt then the article falls flat on this purpose. The supposed fallacy arises when AA is defined as a coordination problem axiomatically. This is "one more subsidy, bro" false cause tautology in itself. The only argument made is a reference to an article which more confirms the critique than opposes, yet no other parts of the supposed coordination puzzle are even as much as speculated.

This is evident in later paragraphs:

> Often it won't be obvious what issues need to be addressed in a coordination problem, which means despite our best attempts to find points of weaknesses while researching and designing a plan, the nature of a coordination problem is that missing one element can lead to failure. If we eliminate individual failed solutions as options it becomes impossible to find the successful coordinated solution.

A statement is made here that a failing individual solution can still be a part of a working coordinated solution, which is not inherently wrong in itself. However, another point raised in this paragraph is that it is supposedly impossible to evaluate suitability of an approach without finding a successful coordinated solution. This marks every failing policy as potentially part of a working coordinated solution and therefore a claim that a policy is part of such a solution inherently unfalsifiable.

> Coordination problems are a particular type of non-zero-sum game, and they are all around us. Until they are solved, they are very much a negative-sum game. The key to solving coordination problems, including affirmative action, is understanding all variables, designing a system-wide approach, and not letting a failure in one area doom the enterprise.

Here affirmative action is defined to be a coordination problem precisely over failure of existing, supposedly uncoordinated, approaches.