There are two ways to make causal claims: by empirics and by theory. That is a theoretical claim, not an empirical one. Ideally we should have both.