How can policy-makers avoid large policy errors when they are uncertain about the true model of the economy? The author discusses some recent approaches that can be used for that purpose under two alternative scenarios: (i) the policy-maker has one reference model for choosing policy but cannot take a stand as to how that model is misspecified, and (ii) the policy-maker, being uncertain about the economy's true structure, entertains multiple distinct models of the economy. The author shows how these approaches can be implemented in practice using as benchmark models simplified versions of Fuhrer and Moore (1995) and Christiano, Eichenbaum, and Evans (2005).