299.Pooling-Based Data Interpolation and Backdatingby Massimiliano Marcellino
Pooling forecasts obtained from different procedures typically reducesthe mean square forecast error and more generally improves the qualityof the forecast. In this paper we evaluate whether pooling interpolatedor backdated time series obtained from different procedures can alsoimprove the quality of the generated data. Both simulation resultsand empirical analyses with macroeconomic time series indicate thatpooling plays a positive and important role also in this context.
Key words: Pooling, Interpolation, Factor Model, Kalman Filter, Spline
JEL Classification: C32, C43, C82.
Download PDF Working Paper
IGIER - Università Bocconi