Crises and Recoveries in an Empirical Model of Consumption Disasters

Working Paper: NBER ID: w15920

Authors: Emi Nakamura; J. N. Steinsson; Robert Barro; JosĂ© UrsĂșa

Abstract: We estimate an empirical model of consumption disasters using a new panel data set on personal consumer expenditure for 24 countries and more than 100 years, and study its implications for asset prices. The model allows for permanent and transitory effects of disasters that unfold over multiple years. It also allows the timing of disasters to be correlated across countries. Our estimates imply that the average disaster reaches its trough after 6 years, with a peak-to-trough drop in consumption of about 30%, but that roughly half of this decline is reversed in a subsequent recovery. Uncertainty about consumption growth increases dramatically during disasters. Our estimated model generates a sizable equity premium from disaster risk, but one that is substantially smaller than in models in which disasters are permanent and instantaneous. It yields new predictions for the dynamics of risk-free interest rates, the term structure of interest rates, and the pricing of short-term versus long-term risky assets. The persistence of consumption declines in our model implies that a large value of the intertemporal elasticity of substitution is necessary to explain stock-market crashes at the onset of disasters.

Keywords: Consumption disasters; Equity premium; Asset pricing; Bayesian estimation

JEL Codes: E21; G12


Causal Claims Network Graph

Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.


Causal Claims

CauseEffect
disasters (H84)consumption (E21)
disasters (H84)recovery effect on consumption (E21)
disasters (H84)uncertainty about future consumption growth (D15)
disasters (H84)equity premium from disaster risk (G52)
disasters (H84)stock market crashes (G01)
disasters (H84)dynamics of risk-free interest rates (E43)
disasters (H84)term structure of interest rates (E43)

Back to index