Yeah, I spent a few minutes going over the last sentence in tengbretson's comment. I would actually recharacterize Piketty's (and the author of this chart) arg

Yeah, I spent a few minutes going over the last sentence in tengbretson's comment. I would actually recharacterize Piketty's (and the author of this chart) argument as: "X increasing is a symptom that Y is increasing. Let's implement policy A. This may temporarily increase X, but it will stop Y from increasing, which will in turn stop X from increasing in the long run."