Measuring and Explaining Economic Uncertainty Across US States and Metropolitan Areas

Abstract

This paper constructs novel measures of economic uncertainty at the state and extended metropolitan area (EMA) levels using forecast errors from predictive models of state/EMA growth. At the state level, we develop uncertainty measures based on sparse-group LASSO MIDAS regressions, incorporating structural adjustments for crisis periods. For Extended Metropolitan Areas (EMAs), we estimate a stacked sg-LASSO MIDAS fixed effects model and apply a forecast error decomposition to extract quarterly uncertainty shocks from annual GDP data. Our results show that overall EMA uncertainty follows similar patterns as state and macroeconomic uncertainty measures, but exhibits significant heterogeneity within periods as well as differentiated drivers. These results provide new insights into regional economic disparities.