The comparative forecast performance of univariate and multivariate models: an application to real interest rate forecasting

Does the use of information on the past history of the nominal interest rat es and inflation entail improvement in forecasts of the ex ante real intere st rate over its forecasts obtained from using just the past history of the realized real interest rates? To answer this question we set up a univaria te unobserved components model for the realized real interest rates and a b ivariate model for the nominal rate and inflation which imposes cointegrati on restrictions between them. The two models are estimated under normality with the Kalman filter. It is found that the error-correction model provide s more accurate one-period ahead forecasts of the real rate within the esti mation sample whereas the unobserved components model yields forecasts with smaller forecast variances. In the post-sample period, the forecasts from the bivariate model are not only more accurate but also have tighter confid ence bounds than the forecasts from the unobserved components model. (C) 19 98 Elsevier Science B.V. All rights reserved.