The impact of disaggregated oil shocks on state-level consumption of the United States

ABSTRACT

We analyse the impact of oil supply, global economic activity, oil-specific consumption demand and oil inventory demand shocks on state-level consumption of the United States (U.S.) over the period of 1975:Q1 to 2012:Q2. We find that positive economic activity shocks and oil production shocks (associated with increase and decrease in oil prices, respectively) increase consumption growth. At the same time, oil-specific consumption and inventory demand shocks raise oil prices and reduce the growth rate of state-level consumption. Across the shocks, the strongest effect originates from the global demand shock. In addition, our above observations are virtually invariant to the degree of oil dependency (oil consumed minus oil produced as a ratio of oil consumed) of the states. Our results have important policy implications.

KEYWORDS:

• Oil shocks
• state-level consumption
• oil dependency
• local projection model
• impulse response functions

JEL CLASSIFICATION:

• C23
• E21
• Q41

Acknowledgments

We would like to thank an anonymous referee for many helpful comments. However, any remaining errors are solely ours.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ The study also found that shocks pushing up oil prices are followed by a national decrease in car registrations, with this decline being larger than the increase in registrations following a comparable oil-price drop.

² For a detailed review of literature on alternative specification of the consumption function, along with variables included in the model, the reader is referred to Coskun et al. (2018).

³ The data set is available from Dr. Guo Li (at the Fannie Mae) upon request.

⁴ The data is available for download from the research segment of the website of Professor Christiane Baumeister at: https://sites.google.com/site/cjsbaumeister/research. Note that the oil-shocks data are monthly, but are converted into quarterly using three-month averages to match our consumption and income data.

⁵ Besides the issue of identification, with the IRFs being independent of the data generating process, the LP approach allows us to estimate the iRFs on a variable-by-variable basis, and hence overcomes the degrees of freedom problem associated with standard VARs.

⁶ The maximum length of forecast horizons is set to 12, which corresponds to 12-quarter forecast horizons. ${\gamma }_s\left(L\right)$ is a polynomial of order 4, which corresponds to 4-quarter lags for control variables.

⁷ We use the Akaike Information Criterion (AIC) and select a maximum of 4 lags (1 year) for automatic lag detection. The optimal lags for the dependent variable and the dynamic regressor are 3 and 1, respectively. The long-run relationship indicated a positive and statistically impact of income on consumption. Complete details of the estimation results are available upon request from the authors.

⁸ Further details about LP IRF techniques can be found in Jordà (2005).

⁹ Figure A1 and A2 in the Appendix A of the article shows the effect on consumption growth (based on the LP model in Equation (1)(1) $\mathrm{\Delta }{C}_{i,t+s}={\alpha }_{i,s}+{\beta }_{s}OilShoc{k}_{i,t}+{\gamma }_s\left(L\right)X_{i,t-1}+{\delta }_{s}EC{M}_{i,t-1}+{\epsilon }_{i,t+s},\mathrm{f}\mathrm{o}\mathrm{r}s=0,1,2,\dots S,$(1) ) from alternative approaches of identifying oil supply and oil demand shocks by Caldara, Cavallo, and Iacoviello (2019) and oil supply shock by Kilian (2008) respectively. Note that Caldara, Cavallo, and Iacoviello (2019) combine narrative analysis, as in Kilian (2008) who used country-specific episodes of exogenous disruptions in oil production, with a panel of observations on country-specific oil production and consumption to estimate oil supply and demand elasticities. They then embed these elasticities in a VAR to identify the oil supply and demand equations and, consequently, the associated oil supply and oil demand shocks. As can be seen from the IRFs, our basic results are robust to these other measures of oil supply and oil demand shocks, with a positive supply shock (i.e., increase in oil production) increasing consumption growth, and a positive demand shock (i.e., an increase in oil-price due to speculative or precautionary reasons) negatively impacting consumption growth.

¹⁰ The impact across the two regimes were also statistically significant, and the IRFs for the high and low-oil dependence regimes with their respective 95% confidence bands are available upon request from the authors.