Irrational Expectations and Econometric Practice∗

Athanasios Orphanides and John C. Williams’ excellent conference paper, “Inflation Scares and Forecast-Based Monetary Policy,” contributes importantly to the new and rapidly growing branch of the literature on bounded rationality and learning in macroeconomics. Their paper, like many others, derives interesting and useful theoretical results that show how the introduction of bounded rationality and learning impacts on the effects of monetary policy shocks and the characteristics of optimal monetary policy rules. This note suggests that some additional empirical work–some “irrational expectations econometrics,” if you will–might serve to make these purely theoretical results seem more relevant and convincing. ∗Discussant’s comments on Athanasios Orphanides and John C. Williams’ “Inflation Scares and ForecastBased Monetary Policy,” prepared for the March 2003 Federal Reserve Bank of Atlanta Conference on Monetary Policy and Learning. †Please address correspondence to: Peter N. Ireland, Boston College, Department of Economics, 140 Commonwealth Avenue, Chestnut Hill, MA 02467 USA. Tel: (617) 552-3687. Fax: (617) 552-2308. Email: [email protected]. http://www2.bc.edu/~irelandp. 1 Irrational Expectations ... Athanasios Orphanides and John C. Williams’ excellent conference paper, “Inflation Scares and Forecast-Based Monetary Policy,” takes as one of its starting points the observation that many central banks around the world devote considerable resources towards producing their own, internal macroeconomic forecasts and towards monitoring private, external macroeconomic forecasts. And, moreover, monetary policy actions taken by central banks around the world often appear to be driven, at least in part, by changing internal and private macroeconomic forecasts. Within the Federal Reserve System, for example, the Greenbook–which presents the macroeconomic forecasts generated by the Federal Reserve Board staff–serves as one of the principal documents guiding the policy deliberations at each meeting of the Federal Open Market Committee. Orphanides and Williams’ paper takes its second starting point another observation: that most contemporary models of the monetary business cycle attach no special importance to macroeconomic forecasts in the design of monetary policy rules. To see where this result comes from, and why it holds true so generally, consider a very simple model in which, for some reason, the central bank decides to set its policy instrument, the short-term nominal interest rate rt at time t, as a linear function of expected or forecasted output y t+1 and inflation πt+1 at time t+ 1, according to the policy rule rt = αyy e t+1 + αππ e t+1, (1) where αy and απ are coefficients chosen by the central bank that measure the sensitivity of the interest rate response to movements in expected output and inflation. Next, suppose that the structure of this simple model implies that the forecasts y t+1 and π e t+1 of output