Abstract:
This paper analyses the role of a news-based measure of economic policy uncertainty (EPU) in explaining time-varying co-movements in economic activity and volatility of 48 US states and 51 largest MSAs. In this regard, we, first, estimate a dynamic factor model with time-varying loadings and stochastic volatility (DFM-TV-SV). Then, in the second step, we use a quantile-on-quantile (QQ) predictive regression model to capture the effect of EPU on the common factor and stochastic volatility derived from the DFM-TV-SV for the states and MSAs. Our results show that EPU has a significant negative effect on the common economic activity of both the states and MASs, and it also significantly increases the common volatility. However, the impact of uncertainty varies substantially depending on the initial states (quantiles) of both common output or volatility and EPU. Thus, our results tend to suggest that policy design should be state-dependent.
