> Rather than having one intuitive factor (personal irresponsibility) which explains several highly correlated outputs (obesity, low income, low education), he wants to propose a more complicated theory which explains only one of these outputs.

I could go pick a bunch of correlated things and ascribe them all to "magic" or "Zeus" or what-have-you, but stuffing all the entropy into a black box like this shouldn't make it count as a "low-entropy theory".