34 Pages Posted: 20 Jan 2012
Date Written: January 2012
Vector autoregressions (VARs) are flexible time series models that can capture complex dynamic interrelationships among macroeconomic variables. However, their dense parameterization leads to unstable inference and inaccurate out-of-sample forecasts, particularly for models with many variables. A potential solution to this problem is to use informative priors, in order to shrink the richly parameterized unrestricted model towards a parsimonious naïve benchmark, and thus reduce estimation uncertainty. This paper studies the optimal choice of the informativeness of these priors, which we treat as additional parameters, in the spirit of hierarchical modeling. This approach is theoretically grounded, easy to implement, and greatly reduces the number and importance of subjective choices in the setting of the prior. Moreover, it performs very well both in terms of out-of-sample forecasting, and accuracy in the estimation of impulse response functions.
Keywords: Bayesian Methods, Forecasting, Hierarchical Modeling, Impulse Responses, Marginal Likelihood
JEL Classification: C11, C32, C52, E37
Suggested Citation: Suggested Citation
Giannone, Domenico and Lenza, Michele and Primiceri, Giorgio E., Prior Selection for Vector Autoregressions (January 2012). CEPR Discussion Paper No. DP8755. Available at SSRN: https://ssrn.com/abstract=1988686
By Todd Clark
This is a CEPR Discussion Paper. CEPR charges a fee of $5.00 for this paper.Login using your CEPR Personal Profile
File name: DP8755.
If you wish to purchase the right to make copies of this paper for distribution to others, please select the quantity.