Uncertainty and energy extraction

Abstract

We examine the effects of uncertainty by investigating the effects of oil price uncertainty on employment in domestic oil and gas extraction. This industry is interesting to examine for two reasons. First, employment in this industry should be closely linked to oil prices, facilitating identification of more subtle economic relations related to uncertainty. For example, recent research finds that uncertainty about oil prices tends to diminish drilling activity in a small sample of drilling rigs in Texas. Second, the response of this industry to negative shocks to the supply of oil, for example, should be opposite that of aggregate employment. We characterize the univariate properties of the time series with a two-state Markov switching model, and we then estimate a vector autoregression with conditionally heteroskedastic errors, in which the conditional variance is permitted to interact with the conditional mean. We find that employment in oil and gas extraction cycles through episodes of boom and bust that are tightly linked to oil prices, in contrast to aggregate employment. We also find that uncertainty about oil prices has a negative and quantitatively large effect on employment in this sector. Our analysis is robust to different measures of uncertainty, including stochastic volatility and implied volatility.