Adverse oil supply shocks trigger asymmetric drop in US activity

Authors Eoghan O'Neill and Sofia Velasco apply the Factor Bayesian Additive Regression Tree (FABART) model to 62 monthly U.S. macro-financial and regional labor variables from 1974 to 2019.

Using high-frequency OPEC surprise instruments, the paper shows that oil supply news shocks display pronounced sign asymmetry.

Adverse shocks that raise oil prices cause deep contractions in U.S. industrial production and equity prices, whereas favorable shocks lowering prices yield muted expansions.

Across shock magnitudes, the jump from a 1.5 percent to a 5 percent oil price increase produces substantial real effects, with CPI inflation responses remaining significant for over two years, while 10 percent shocks yield diminishing proportional impact.

Regional data across 50 U.S. states demonstrates substantial divergence in labor-market adjustments.

Manufacturing-heavy states such as Michigan and Ohio experience prolonged job losses due to energy input costs and reallocation frictions.

In contrast, major oil-producing states like Wyoming and Alaska record stable or positive employment gains after oil price increases.

Incorporating forward-looking financial variables such as VXO and Treasury yields attenuates global inventory responses while capturing sharper drops in real activity.