Kilian (2008) Exogenous Oil Supply Shocks

Summary

Kilian builds a new event-based measure of exogenous OPEC oil-production shortfalls — counterfactual production paths for countries hit by wars and political disturbances since 1973, benchmarked against uninvolved OPEC producers — and uses it to ask three questions: how large are exogenous oil-supply fluctuations, how much of oil-price movements do they explain, and what are their dynamic effects on U.S. real GDP growth and CPI inflation. The answers overturn conventional wisdom: exogenous supply shocks explain only a small fraction of oil-price increases, and their average macroeconomic effect is modest (though non-trivial in specific episodes).

Key Claims

• A better shortfall measure. Versus Hamilton's (2003) quantitative dummies, the counterfactual construction (1) doesn't assume production would have been flat absent the event, (2)–(4) allows immediate/delayed, long-lasting, and time-varying responses, and (5) allows the response to be negative or positive — Middle East wars can raise output when combatants export oil to finance fighting.
• Weak-instrument problem. Using the shortfall series as an instrument for the oil price in macro projections (à la Hamilton 2003) suffers from weak instruments, undermining IV inference.
• A simpler OLS alternative. To measure the effect of exogenous supply shocks, directly project real GDP growth (or inflation) on current and lagged shock values plus lagged dependent values — OLS with standard inference, no IV needed.
• Modest but episodic macro effects. Statistically significant sharp drop in real GDP growth ~5 quarters after a shock and a CPI inflation spike ~3 quarters after; but averaged over 1973–, exogenous supply shocks mattered little for U.S. GDP/inflation overall — except in episodes like the 1990/91 Gulf War.
• Supply shocks explain little of oil prices. Exogenous production shortfalls account for only a small share of crisis-period oil-price increases (only 1980/81 is clearly supply-driven), pointing instead to oil demand shifts and precautionary/uncertainty demand as the main drivers — foreshadowing the structural demand/supply decomposition of Kilian (2009).
• Net oil price increase ≠ exogenous component. Popular statistical measures of the "net oil price increase relative to the recent past" do not isolate the exogenous part of oil prices.

Concepts Introduced or Extended

• Oil Price Shocks

Entities Mentioned

• Lutz Kilian

Quotes

"Only a small fraction of the observed oil price increases during crisis periods can be attributed to exogenous oil production disruptions."

"Exogenous oil supply shocks made remarkably little difference overall for the evolution of U.S. real GDP growth and CPI inflation since the 1970s, although they did matter for some historical episodes."

My Take

This is the demolition-work half of Kilian's oil-market project: before the celebrated structural decomposition in Kilian (2009), he first has to dislodge the entrenched "major oil price shocks are exogenous supply events" view, and he does it with a hand-built counterfactual shortfall series rather than a reduced-form VAR — which is what makes the exogeneity claim credible. The two durable methodological lessons for the wiki are (i) the weak-instrument warning against Hamilton-style IV oil-price regressions, and (ii) the pragmatic point that if you only want the effect of the exogenous shock, a direct OLS projection sidesteps IV entirely. Substantively it sets up the punchline of the whole literature — that oil prices are largely demand-driven — which the AER paper then formalizes; this paper is best read as its empirical and rhetorical prerequisite. The sign-can-flip insight (wars raising output) is a nice guard against mechanical dummy-variable event studies.