The State-Contingent Link between Market Illiquidity and Economic Growth

Abstract

I examine the ability of U.S. equity and bond market illiquidity to predict U.S. macroeconomic variables, between 1946 and 2010. In contrast to existing studies, I allow for illiquidity's predictive ability to be state contingent, using a Markov regime switching model. I uncover strong evidence that the predictive power of equity market illiquidity is state-contingent, with much higher predictability in states associated with historical periods of economic and financial stress. Furthermore, economic growth forecasts from Markov regime switching models that include market liquidity in the set of predictor variables are statistically better than forecasts from Markov switching models that exclude liquidity.

Keywords: autoregression, bid-ask spread, economic growth, forecasts, regime switching, trading volume

JEL Classification: G12, G14, G17