How Much Should Monetary Policy Take Output

I would like to thank the organizers for the privilege of discussing two of the fine papers presented here today. My assignment is to discuss Gaspar and Smets (2002) and Wolfers (2003), as well as to give my own view on how much monetary policy should take output into account and the best way of doing so. Let me start with the Gaspar and Smets paper. I find it very interesting and well worth studying. Let me start from the standard objective for monetary policy that involves stabilizing inflation around an inflation target and stabilizing output around an output target equal to potential output. This can be formulated as an intertemporal loss function to be minimized, (1) E t ∑ ∞ = + − 0) 1 (τ τ τ δ δ t L , with the quadratic period loss function (2) 2 2) () (t t t t y y L − + − = λ π π. Here, E t denotes expectations conditional on information available in period t, δ is the discount factor and fulfills 0 < δ < 1, π t is inflation, π is the inflation target, y t is output, t y is potential output, and t t y y − * Presented at the Council on Foreign Relations/International Finance conference " Stabilizing the Economy, " New York, July 11, 2002, and revised in February 2004. I thank Kathleen Hurley for editorial and secretarial assistance. Remaining errors and expressed views are my own.