A1176
Title: The speed of state-level recoveries
Authors: Luiggi Donayre - University of Minnesota - Duluth (United States) [presenting]
Irina Panovska - University of Texas at Dallas (United States)
Abstract: The abrupt decline in U.S. GDP in 2020:Q2 has reignited an interest in the shape of recessions. Aggregate models assume that business cycles are similar across states. However, several economic ideas suggest that recessions and their recoveries may be idiosyncratic to each state. For example, those triggered by financial crises could generate longer-lasting effects, indicating L-shaped recessions. Meanwhile, downward wage rigidity may generate contractions followed by recoveries of similar amplitude, implying U-shaped recessions. Since financial crises and labour market features are heterogeneous, the speed of recovery across states, and thus recession shapes, may vary widely. To study these issues, a Markov-switching model is estimated, augmented with a bounce-back effect, to U.S. state-level output growth and find large differences across states. At the aggregate level, the bounce-back parameter 0.32 suggests relatively slow recoveries. State-level bounce-back estimates range from 0.05 to 1.36, evidencing large differences in the speed of recovery. Southern and rust-belt states exhibit slow recoveries, suggesting more L-shaped recessions. Meanwhile, mountain and oil-producing states exhibit faster, U-shaped recoveries. Because economic policies are designed to smooth aggregate business cycle fluctuations, such policies' effects vary widely across states and regions.
