Prior Selection for Vector Autoregressions

By Domenico Giannone, M. Lenza, Giorgio E. Primiceri. Review of Economics and Statistics, 2015

705 citations recorded by the source corpus.

Abstract 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 solution to this problem is to use informative priors in order to shrink the richly parameterized unrestricted model toward a parsimonious naıve 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, theoretically grounded and easy to implement, greatly reduces the number and importance of subjective choices in the setting of the prior. Moreover, it performs very well in terms of both out-of-sample forecasting—as well as factor models—and accuracy in the estimation of impulse response functions.

Subfields: Economics and Econometrics, Management Science and Operations Research

View Paper: https://www.semanticscholar.org/paper/14e47842a32a7190d110aeb1cf8a24434236b7b8

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Economics and Econometrics

In Management Science and Operations Research

Citations in the corpus

What's Up with the Phillips Curve? (2020, positive): “The tightness of the prior is chosen based on the data-driven method described in Giannone et al. (2015).”
Real-Time Forecasting With a Mixed-Frequency VAR (2015, neutral): “…toward the prior mean and we choose the hyperparameter to maximize the log MDD. MDDbased hyperparameter selection has been discussed, for instance, in Phillips (1996), in Del Negro and Schorfheide (2004), Del Negro et al. (2007), and, most recently, studied in Giannone, Lenza, and Primiceri (2014).”
If the Fed Sneezes, Who Catches a Cold? (2017, background): “We conclude from this exercise that this prior choice strikes a reasonable balance between making the US estimates based on (2) close to their BVAR counterparts, a minimal consistency requirement, and imposing too tight a constraint on the posterior inference.(18) The flexibility of this approach re”
Mending the broken link: Heterogeneous bank lending rates and monetary policy pass-through (2020, background): “The vector of hyper-parameters controlling the informativeness of the prior, is treated as random, to account for the uncertainty surrounding the specification of the prior, and the posterior of ( ) is jointly computed (see Giannone et al. 2015).”
How Does Monetary Policy Affect Income and Wealth Inequality? Evidence from Quantitative Easing in the Euro Area (2018, methodology): “We adopt the Bayesian estimation method described in Giannone et al. (2015).”
A quasi-Bayesian local likelihood approach to time varying parameter VAR models (2019, methodology): “Giannone et al. (2015) use a hierarchical prior for the overall tightness parameter ; p ( ) which then delivers a posterior”
Energy Markets and Global Economic Conditions (2020, methodology): “As in Baumeister and Kilian (2012), we rely on the data-based procedure proposed in Giannone, Lenza, and Primiceri (2015) for selecting the optimal degree of shrinkage in our recursively estimated Bayesian VARs (BVARs) based on the marginal data density.”
What are the Macroeconomic Effects of Asset Purchases (2016, methodology): “When our model was estimated with six lags, it was, as a result of the short time-series, necessary to use a Litterman (1986) prior, with the hyper-parameters estimated from the data following the approach in Giannone et al.”
Macroeconomic forecast accuracy in a data‐rich environment (2019, background): “See Banbura et al. (2010), Koop (2013), Carriero et al. (2015) and Giannone et al. (2015), among others.”