On Economists' Belief in the Law of Small Numbers