Large Bayesian VARs for Binary and Censored Variables

TL;DR

This paper introduces an extended Bayesian VAR model capable of jointly handling binary, censored, and continuous variables, with efficient estimation suitable for high-dimensional data, and demonstrates its effectiveness in economic forecasting and structural analysis.

Contribution

The paper develops a scalable Bayesian VAR framework for mixed data types, enabling more comprehensive economic modeling and forecasting.

Findings

Accurately forecasts recessions and interest rates

Effectively models the impact of financial shocks on recession probabilities

Demonstrates utility across diverse empirical applications

Abstract

We extend the standard VAR to jointly model the dynamics of binary, censored and continuous variables, and develop an efficient estimation approach that scales well to high-dimensional settings. In an out-of-sample forecasting exercise, we show that the proposed VARs forecast recessions and short-term interest rates well. We demonstrate the utility of the proposed framework using a wide rage of empirical applications, including conditional forecasting and a structural analysis that examines the dynamic effects of a financial shock on recession probabilities.