Pooling forecasts in linear rational expectations models |
| |
Authors: | Gregor W. Smith |
| |
Affiliation: | aDepartment of Economics, Queen's University, Kingston, Ontario, Canada K7L 3N6 |
| |
Abstract: | Estimating linear rational expectations models in a limited-information setting requires replacing the expectations of future, endogenous variables either with instrumented, actual values or with forecast survey data. Applying the method of Gottfries and Persson [Empirical examinations of the information sets of economic agents. Quarterly Journal of Economics 103, 251–259], I show how to augment these methods with actual, future values of the endogenous variables to improve statistical efficiency. The method is illustrated with an application to the US hybrid new Keynesian Phillips curve, where traditional, lagged instruments and the median forecast from the Survey of Professional Forecasters both appear to miss significant information used by price-setters, so that forecast pooling with actual values improves the statistical fit to inflation. |
| |
Keywords: | Forecast pooling Recursive projection New Keynesian Phillips curve |
本文献已被 ScienceDirect 等数据库收录! |
|