A SMARTer way to forecast

Abstract

In this paper we describe the newly developed System for Model Analysis in Real Time (SMART) used for forecasting and model analysis in Norges Bank. While the long-term goal is to include all empirical models used in forecasting in Norges Bank, the emphasis in this paper will be on the empirical model systems for inflation and GDP. SMART builds on Norges Bank’s previous System for Averaging short-term Models (SAM), but with greater flexibility and a richer set of models. In addition, SMART contains a real-time database with a wide-ranging set of historical data, forecasts from empirical models, Norges Bank’s forecasts from Monetary Policy Reports (MPR) and forecasts from other institutions (e.g. Statistics Norway). Overall, SMART seems to provide good forecasts and will be a useful tool in the monetary policy process.