“…A policymaker that is confident about the reference model will typically choose a high θ as he does not want to pay a high price in terms of social costs by deviating too much from the optimal policy in a model he believes in. The problem solves for the 3 For similar approaches in dealing with model uncertainty, see Hansen and Sargent (2003a,b), Giordani and Söderlind (2004), , Kilponen (2003Kilponen ( , 2004, , Söderström (2004, 2005), Walsh (2004Walsh ( , 2005b, Onatski and Williams (2003) and Tetlow and Muehlen (2001). 4 In setting up the state space form of the model we are assuming that policy is implemented using the short-term interest rate, i t .…”