I don't even care for the moment whether the causal link is true or not; for all I know, it may be. The broader point is that things can be real without you finding where they are in the brain. Moreover, you finding a correlation between something in the brain and something in the real world does not let you ascribe full causal force to the thing in the brain. And if you want to link the thing in the brain to future poverty, you might as well actually test that link, instead of saying "it hurts working memory by a little! Our work is done." Plus, how many interesting dynamic things are the in the world, where, say, poverty changes or income changes or something happens. This kind of stuff happens all of the time. Notice how you can explain exactly none of those changes using something that is a fixed human constant.
This stress thing is really popular among people who study healthcare too. But there's really nothing you can do about stress without radical social engineering. On the other hand, there are a million and one interventions you can think of on education and healthcare that can radically improve both. DC's voucher program, for instance, improves kids reading scores by a half grade. Except that program just got killed by Congress. Awesome.