Design of Modern Forecasting and Policy Analysis Systems at Central Banks

In the last 15 years, a growing number of central banks have adopted a new type of monetary policy regime: inflation targeting. Unlike exchange rate management, money targeting, or other regimes, inflation targeting requires deeper understanding of economic forces that determine inflation, output, and other macroeconomic variables, and (no less importantly) new attitudes to public communication.

To facilitate economic analysis and make monetary policy decisions more transparent and accountable, many central banks have built new model-based forecasting and policy analysis systems. These involve computationally intensive tasks and require powerful software, such as MATLAB, and the domain-specific community-developed IRIS Toolbox.

This presentation describes key components of such a forecasting and policy analysis system, including the core macroeconomic model; explains why forecasters and policy-makers need to distinguish between short-term conjunctural forecasts and medium-term structural projections; and demonstrates live computer-aided experiments to illustrate typical modeling and forecasting tasks. These experiments examine, for example, the properties of a stochastic economic model, using the model to analyze recent data outturns, build consistent projections and policy scenarios, and condition model simulations upon judgmental adjustments.