The Implications of Information Lags for the Stabilization Bias and Optimal Delegation

Many papers for example Jensen (2002) and Walsh (2003) have shown that in a New Keynesian model with a significant degree of forward-looking behaviour, policy regimes that target either the change in the output-gap (speed limit targeting) or nominal income growth can considerably reduce the size of the stabilization bias - the inefficiency that arises when a central bank conducts policy under discretion as opposed to commitment. Inflation targeting can also reduce the size of the stabilization bias but unless inflation expectations in the model are predominantly backward-looking, this targeting regime does not perform as well as speed limit or nominal income growth targeting. Jensen (2002) and Walsh (2003) obtain their results using a New Keynesian model where changes in the policy rate affect macroeconomic variables immediately. In this paper, we compare the performance of several targeting regimes by using a New Keynesian model that includes a delayed response of monetary policy as a result of information lags. We find two results that are substantially different from Jensen (2002) and Walsh (2003). First the size of the stabilization bias is considerably reduced. Second, a regime that targets inflation outperforms a regime that targets either the change in the output-gap or the growth in nominal income even when inflation expectations are very forward-looking.