Simplified Regional-National DSGE Models for U.S. State-Level Analysis

This thesis develops a series of simplified Dynamic Stochastic General Equilibrium (DSGE) model variants to analyze macroeconomic dynamics at the U.S. state level. Using South Dakota as a case study, I extend the Smets and Wouters (2007) mediumscale framework by implementing state-specific output, labor, consumption, investment, and wages into a dual-economy structure, where the state is modeled as a closed economy influenced by national-level shocks. The models are estimated using Bayesian techniques on quarterly data from 2007 - 2020. Model performance is evaluated through impulse response functions and variance decompositions. A key result is that South Dakota exhibits firm habit formation in consumption behavior, leading to more gradual and dampened responses to shocks than the national economy. This shock-resistance has implications for state-level policy design and highlights the need for tailored responses in the presence of national economic disturbances.