The COVID-19 pandemic and the consumption of nondurables and services

ABSTRACT

Contrary to the Great and other past US Recessions, the reduction in services consumption exceeds the decline in nondurables consumption during the COVID-19 pandemic. We study the drivers of this unprecedented phenomenon through the lens of an estimated multi-sector Dynamic Stochastic General Equilibrium (DSGE) model that distinguishes between nondurables and services sectors. We find that economic uncertainty is once again important, but it does not generate sectoral heterogeneity. Demand-side factors reallocating consumption across sectors and proxying for voluntary and regulatory social distancing measures, as well as the lack of wage adjustments in services despite plummeting employment, became influential during the pandemic.

KEYWORDS:

• COVID-19
• services
• nondurables
• DSGE

JEL CLASSIFICATION:

• E31
• E32
• E24

Acknowledgments

We are deeply grateful to the editor, David Peel, an anonymous referee, and seminar participants at the 6th International Symposium in Computational Economics and Finance for their helpful comments and suggestions.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Baker et al. (2020) emphasize voluntary social distancing and nonpharmaceutical policy interventions on commercial activity in services as the reasons behind the large stock market reaction to COVID-19; they mention: ‘By late March, nearly 100% of U.S. residents lived in counties where state or local officials had closed schools and dine-in restaurants, roughly 70% lived in counties with mandatory closures of non-essential businesses, and roughly 90% were subject to stay-at-home orders and bans on public gatherings’. Gupta et al. (2020) and Maloney and Taskin (2020) study the determinants of social distancing using microdata.

² Atalay (2017) finds that sectoral shocks have a pronounced influence on the business cycle in a multi-industry model.

³ Had we relaxed this assumption, we could model a distinct investment price. Given the lack of investment data across nondurables and services, we abstract from this channel.

⁴ The price markup is set at 10%, the discount factor, $\beta$, at 0.98, the depreciation rate, $\delta$, at 2.5%, and the government-spending-to-output ratio at 18%. The steady-state share of services consumption, $\tau$, is fixed at its sample average (0.74).

⁵ During the period of the binding zero lower bound, we substitute the FFR by the shadow rate of Wu and Xia (2016). All other data series are obtained from FRED.

⁶ Differences in the treatment of labour trends blur this comparison. Moura (2018) removes sector-specific quadratic trends. In contrast, we treat aggregate labour as in Smets and Wouters (2007), and demean our sectoral labour series. Iacoviello and Neri (2010) estimate $\varphi$ at 0.66 and 0.97 for patient and impatient agents, respectively, using rather tight prior distributions centred at 1 with 0.1 standard deviation. Cardi, Restout, and Claeys (2020) and Artuc, Chaudhuri, and McLaren (2010) find barriers to labour reallocation across non-traded and traded goods using a VAR model and the CPS survey, respectively.