This letter suggests, and describes through an example, a dominance approach to rational expectations as an alternative to the perfect foresight fixed point for closing a model. Dominance criteria have a long history in game theory [see, e.g., Harsanyi (1977, p. 107)]. It also has long been recognized that dominance is not a universally applicable model of behavior [see e.g., Harsanyi (1977, p....