The Use and Abuse of Taylor Rules: How precisely can we estimate them?

This paper surveys the economic literature on simple policy rules and analyzes econometric methods used to estimate them, emphasizing effects of model misspecification. We draw attention to inconsistencies in evaluation of the rules and implications for policy advice, which is commonly done based on benchmark rules that could be improperly estimated, or selected for a wrong reason. We simulate a simple macroeconomic model with an interest rate obtained from a simple policy reaction function similar to Taylor (1993). We estimate different versions of the simple policy rule, using the simulated data. Estimations document illusionary presence of extra variables, such as lagged interest rate, output gap growth, and inflation differential; or claim the policy function to be forward looking. Length of the sample or ignorance of real time data errors do not seem to have significant impact on the results.