Helping Central Banks Measure Unobserved Variables Using Real-time Forecasts

Central banks rely on estimates of unobserved variables to set interest rates. Currently, policymakers draw on evidence from a variety of models, using implicit and informal methods. In partnership with several central banks, we apply statistical methods for combining forecasts to structural measurement issues. The estimate of the unobserved variable reflects its performance in forecasting inflation in real time, consistent with the inflation targeting regime common in central banks. The project will provide central banks with a better indication of the structure of the underlying economy and will facilitate transparency about monetary policy decisions.