Estimating U.S. Housing Price Network Connectedness: Evidence from Dynamic Elastic Net, Lasso, and Ridge Vector Autoregressive Models

Abstract

This paper investigates the dynamic connectedness of random shocks to housing prices between the 50 U.S. states and the District of Columbia. The paper implements a standard vector autoregressive (VAR) model as well as three VAR models with shrinkage effects - Elastic Net, Lasso, and Ridge VAR models. The transmission of random shocks on a regional basis flows from Southern states to Western states to Midwestern states to Northeastern states. Since VAR models generally confront parameter values between zero and one, the Elastic Net and Lasso VAR models perform the best since the penalty involves the absolute value rather than he squared value as in the Ridge VAR model. Our results have important implications for investors and policymakers.