This seems like a breakthrough study, and Sandhil Mullainathan’s name on the paper guarantees that it’s methodologically as tight as it could be. Note that one of the two studies is an experiment, while the other is a before-and-after on a fixed set of subjects,* eliminating any question about which way the causal arrow points.
Mullainathan comments that the 13-point measured IQ gap between Indian peasants under pre-harvest financial stress and the same people with a post-harvest cushion is about the same drop you’d get after missing a night’s sleep. It’s also about the same as the measured black-white IQ gap in the U.S.; in addition to having lower incomes, African-Americans also have much lower wealth (even controlling for income) and thus face much greater financial stress.
Kelly is right: The poor are different from the rest of us primarily because they have less money, and providing money, while not the full solution, is an essential piece.
*[Updated to correct an error spotted by Keith in comments.]