Forecasting with Bayesian vector autoregressions: An application to post-liberalization Turkey, 1980-1991.

Title: Forecasting with Bayesian vector autoregressions: An application to post-liberalization Turkey, 1980-1991.
Identifier: AAI9224856
identifier: 9224856
Creator: Selcuk, Faruk.
Contributor: Adviser: Salih N. Neftci
Date: 1992
Language: English
Publisher: City University of New York.
Subject: Economics, General | Economics, Theory
Abstract: The study compares the forecast performances of a univariate and a vector autoregressive (VAR) model to two bayesian vector autoregressive (BVAR1, BVAR2) models in connection with nine major macroeconomic variables selected from the post-liberalization Turkish economy.;After the investigation of time-series properties of the variables, the study concentrates on the specification of time-varying parameter estimation of BVAR models using the Kalman filter. With regard to the out-of-sample forecasts, the BVAR models appear to have more accurate forecasts over long horizons than the univariate and the simple VAR model for most of the variables. The results indicate the superiority of the BVAR model in capturing the important long-run interactions between the variables.
Type: dissertation
Source: PQT Legacy CUNY.xlsx
degree: Ph.D.

Media: Forecasting with Bayesian vector autoregressions: An application to post-liberalization Turkey, 1980-1991.