There's a danger here, of which I imagine they're aware - higher-order effects are very hard to estimate in chance and magnitude, or in how your actions specifically contributed to them. This makes them perfect for justifying whatever you want, intentionally or accidentally. Overemphasize some positive second-order effects, fail to notice some negative second-order effects, and suddenly your first-order selfish choice looks like a charity. Societies and markets are dynamic systems, so predicting outcomes is less like predicting path of a rocket from Newton's laws, and more like predicting the weather.