Do oil price changes have symmetric or asymmetric effects on the unemployment rate?: Empirical evidence from Alaska

ABSTRACT

The literature on the oil price–unemployment nexus assumes that the effects of changes in oil prices are symmetric. In this paper, we use the nonlinear autoregressive distributed lag methodology and test the symmetric versus asymmetric effects of oil price changes on unemployment in the case of Alaska. Using quarterly data over the period 1987–2014, the results show that indeed the effects are asymmetric in the short run; Alaska’s unemployment rate is more responsive to oil price increases than to oil price decreases. However, there is little evidence that the asymmetric short-run effects last into the long run.

KEYWORDS:

• Alaska
• asymmetry
• nonlinear ARDL
• oil prices
• unemployment

Notes

¹ These studies are part of a larger literature on the so-called oil price–macroeconomics nexus. Jones at al. (2004) and later Hamdi and Sbia (2013) provide an extended review of the relevant literature.

² Another advantage of the ARDL method is that, unlike the standard cointegration methods, the short- and long-run coefficients in Eq. (6) can be estimated in a single step.

³ Since the gross state product for Alaska is not available on a quarterly basis, personal income is used in our model. A strong correlation between these two variables measured on an annual basis (correlation coefficient = 0.97) shows that personal income also should be relevant in explaining unemployment in Alaska.

⁴ This model specification outperforms an alternative model in which the interest rate is in logarithm in terms of expected signs and number of significant coefficients.

⁵ These findings are consistent when using the Brent crude oil prices in estimating Eq. (9). Interested readers can contact the authors for the results.