Modeling the employment–oil price nexus: A non-linear cointegration analysis for the U.S. market

ABSTRACT

Theoretically, fluctuations in oil prices are expected to affect production costs and may force businesses to delay their investment decisions, triggering pressures on employment. Following these theoretical notions, this paper investigates the asymmetric impact of oil prices on employment (measured as total employment, male employment, and female employment), in a nonlinear cointegration structure for the U.S. market. In doing so, this paper adopts the nonlinear autoregressive distributed lags (NARDL) model to shed light on such asymmetric association, as the NARDL model recently emerged as a new direction in examining nonlinear cointegration and asymmetry. The empirical findings document a long-run asymmetric effect in case of total employment and male employment only. Furthermore, the short-run asymmetric effect was detected for all three employment categories. As a final point, the Granger Causality test documents a unidirectional causality running from oil price decrease to both total employment and male employment.

KEYWORDS:

• Employment
• oil price
• asymmetry
• NARDL

JEL CLASSIFICATIONS:

• E24
• J21
• Q40
• Q49

Acknowledgments

The authors wish to express their gratitude to two referees and the editor of this journal whose comments and suggestions improved the merit of this work. Needless to say, the usual disclaimer applies.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Amjad M. Kisswani http://orcid.org/0000-0003-3259-4969

Notes

1 The effect found in case of male employment but not in case of female employment, could be due to the differences in the labor participation rates, where male have higher rates compared to female participation rate.

2 See Carruth, Hooker, and Oswald (1998) for more details.

3 In this paper, we used the EViews package for our analysis. Furthermore, although EViews allows for linear ARDL estimation, non-linear ARDL is only obtained by using Richard Olayeni’s 2017 ‘Add-in’ for EViews.

4 See Bahmani-Oskooee and Ghodsi (2016) for more details.