Working Paper: NBER ID: w29304
Authors: Martin Beraja; Christian K. Wolf
Abstract: We argue that recoveries from demand-driven recessions with expenditure cuts concentrated in services or non-durables will tend to be weaker than recoveries from recessions more biased towards durables. Intuitively, the smaller the bias towards more durable goods, the less the recovery is buffeted by pent-up demand. We show that, in a standard multi-sector business-cycle model, this prediction holds if and only if, following an aggregate demand shock to all categories of spending (e.g., a monetary shock), expenditure on more durable goods reverts back faster. This testable condition receives ample support in U.S. data. We then use (i) a semi-structural shift-share and (ii) a structural model to quantify this effect of varying demand composition on recovery dynamics, and find it to be large. We also discuss implications for optimal stabilization policy.
Keywords: demand composition; recoveries; monetary policy; pent-up demand
JEL Codes: E32; E52
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Demand-driven recessions with spending cuts in services or nondurables (H69) | Weaker recoveries (E65) |
| Spending cuts concentrated in services or nondurables (H69) | Recovery dynamics (C69) |
| Aggregate demand shock (E00) | Faster reversion of expenditures on durable goods (E21) |
| Durable expenditures (E20) | Stronger overshoot compared to services (L89) |
| Normalized cumulative impulse response (NCIR) of output in service-heavy recession (E19) | Larger than in durables-led recession (E20) |
| Differences in demand composition (R22) | Implications for optimal stabilization policy (E63) |