Working Paper: NBER ID: w27733
Authors: Qian Chen; Christoffer Koch; Padma Sharma; Gary Richardson
Abstract: Banking-system shutdowns during contractions scar economies. Four times in the last forty years, governors suspended payments from state-insured depository institutions. Suspensions of payments in Nebraska (1983), Ohio (1985), and Maryland (1985), which were short and occurred during expansions, had little measurable impact on macroeconomic aggregates. Rhode Island’s payments crisis (1991), which was prolonged and occurred during a recession, lengthened and deepened the downturn. Unemployment increased. Output declined, possibly permanently relative to what might have been. We document these effects using a novel Bayesian method for synthetic control that characterizes the principal types of uncertainty in this form of analysis. Our findings suggest policies that ensure banks continue to process payments during contractions – including the bailouts of financial institutions in 2008 and the unprecedented support of the financial system during the COVID crisis – have substantial value.
Keywords: payments crises; macroeconomic aggregates; Bayesian synthetic control
JEL Codes: C01; C11; E02; E32; E44; E58; G01; G2; G21; G28
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Unemployment increase (J64) | Labor force participation decline (J21) |
| Payment suspensions in Rhode Island (J65) | Unemployment increase (J64) |
| Payment suspensions in Rhode Island (J65) | Labor force participation decline (J21) |
| Payment suspensions in Rhode Island (J65) | Gross state product (GSP) decline (P24) |
| Payment suspensions in Rhode Island (J65) | GSP decline (F62) |