← Back to KB Index
Kilian Framework
typeScenario — Analytical Framework

Lutz Kilian's 2009 structural VAR model distinguishing oil supply shocks, aggregate demand shocks, and oil-specific demand shocks — the foundational framework for decomposing oil price movements.

Overview

The Kilian framework, introduced in Lutz Kilian's seminal 2009 paper "Not All Oil Price Shocks Are Alike: Disentangling Demand and Supply Shocks in the Crude Oil Market" (American Economic Review, 99(3): 1053–1069), is the dominant structural vector autoregression (SVAR) model used in academic and policy analysis to decompose oil price movements into distinct causal drivers. Prior to Kilian's work, the literature treated oil price increases as homogeneous supply shocks. Kilian demonstrated that this conflation produces misleading policy conclusions — the same price increase can reflect fundamentally different market conditions depending on which shock drives it.

The framework identifies three structurally distinct shocks:

  1. Oil supply shocks — disruptions to crude oil production
  2. Aggregate demand shocks — shifts in global real economic activity
  3. Oil-specific demand shocks — precautionary demand for oil inventories driven by uncertainty about future supply

The Three-Shock Model

Oil Supply Shocks

Supply shocks represent unexpected changes in global crude oil production. These correspond to the traditional understanding of oil price spikes: wars, embargoes, OPEC production cuts, or infrastructure failures that physically remove barrels from the market.

Kilian measures supply shocks using monthly data on global crude oil production. The identifying assumption is that within a given month, global oil production does not respond to contemporaneous demand shocks — production is predetermined by existing capacity, contractual obligations, and geological constraints. This is the "short-run" exclusion restriction that makes the structural VAR identifiable.

In the Kilian decomposition, oil supply shocks account for a relatively modest share of oil price variation — typically explaining less than 15-20% of the forecast error variance in oil prices at horizons of 12-36 months. This finding surprised many economists who had attributed most oil price movements to supply disruptions.

Aggregate Demand Shocks

Aggregate demand shocks capture unexpected shifts in global real economic activity that drive oil demand. Kilian measures this using an index of global real economic activity constructed from dry cargo freight rates — the Baltic Dry Index and similar shipping rate measures. The logic is that shipping rates reflect real commodity demand and industrial production more accurately than GDP statistics, which are published with lags and revisions.

The identifying assumption is that real economic activity responds to oil supply shocks with a one-month delay but may respond contemporaneously to demand shocks. This allows the model to distinguish between supply-driven and demand-driven price increases.

Aggregate demand shocks explain the largest share of oil price variation — typically 40-60% of the forecast error variance at business cycle horizons. This finding reframes the historical narrative: many episodes previously attributed to supply disruptions (including parts of the 1970s shocks) had significant demand-side components driven by global industrial booms.

Oil-Specific Demand Shocks

The most innovative contribution is the identification of oil-specific demand shocks — precautionary demand driven by uncertainty about future oil supply. When geopolitical risk rises, market participants increase oil inventories as a hedge against potential supply disruptions. This precautionary demand bid is distinct from both physical supply reductions and broad economic demand.

Oil-specific demand shocks are identified as the residual after accounting for supply and aggregate demand shocks. They capture the "fear premium" — the component of oil prices driven by expectations of future supply disruption rather than current fundamentals.

These shocks explain 15-30% of oil price variation and are particularly relevant during geopolitical crises. The precautionary demand channel means that oil prices can spike before any actual supply disruption occurs, purely on the basis of increased risk perception.

Estimation Methodology

The Kilian model is estimated as a three-variable structural VAR:

The recursive identification structure imposes a lower-triangular contemporaneous impact matrix. This Cholesky-type identification means:

Kilian uses monthly data starting from 1973 (post-Bretton Woods) and estimates the model with 24 lags. The choice of lag length is important because oil market dynamics involve long propagation delays.

Extensions and Refinements

The original 2009 framework has been extended in several directions:

Kilian and Murphy (2014)

Introduced speculative demand as an additional structural shock, using data on net long positions of non-commercial traders in oil futures markets. This "fourth shock" captures financial speculation distinct from precautionary demand.

Kilian and Hicks (2013)

Extended the framework to incorporate revisions in global GDP data, showing that apparent demand shocks in real-time data are partly artifacts of measurement error.

Baumeister and Kilian (2016)

Developed conditional forecasts within the SVAR framework, allowing analysts to trace out the path of oil prices under specific supply disruption scenarios — directly applicable to policy analysis during crises.

Antolín-Díaz and Rubio-Ramírez (2018)

Used narrative sign restrictions as an alternative to the recursive identification, providing robustness checks on the three-shock decomposition.

Application to Historical Episodes

1973 Arab Embargo

The Kilian decomposition reveals that while the 1973 embargo was a genuine supply shock, a significant portion of the price increase reflected aggregate demand pressures from the preceding global economic boom and Nixon-era monetary expansion. The pure supply contribution was smaller than the headline production cut suggested.

1979 Iranian Revolution

The 1979 crisis is decomposed into a primary supply shock (Iranian production loss), a large precautionary demand shock (fear of broader Gulf disruption), and residual demand effects from global industrial activity. The precautionary component explains why prices remained elevated even after Saudi Arabia partially offset Iranian losses.

2003-2008 Price Surge

Kilian's framework shows that the pre-2008 oil price surge was predominantly driven by aggregate demand shocks — the global commodity supercycle driven by Chinese industrialization — rather than supply constraints. The financial speculation component (identified in the Kilian-Murphy extension) played a secondary but non-trivial role.

2014-2016 Price Collapse

The decomposition attributes the price collapse primarily to negative aggregate demand shocks (slowing Chinese growth) combined with positive supply shocks (US shale production growth). OPEC's decision not to cut production amplified the supply-side shock.

Relevance to the 2026 Hormuz Crisis

The 2026 Hormuz closure presents a uniquely complex challenge for the Kilian framework because it simultaneously involves all three shock types:

Supply Shock Component

The closure removed approximately 20 million barrels per day from global markets — by far the largest supply disruption ever recorded. The Kilian model would identify this as an extreme negative supply shock, 3-5× larger than the 1973 or 1979 episodes.

Aggregate Demand Shock Component

The simultaneous global recession triggered by the oil price spike generates a negative aggregate demand shock. The Kilian framework would need to disentangle the causal direction: did the supply shock cause the demand contraction, or are they independent? In a standard SVAR, the demand contraction would appear as a separate negative demand shock with a lag.

Oil-Specific Demand Shock Component

The precautionary demand surge was extreme. With the strait closed, market participants scrambled to secure physical inventories, driving a massive fear premium. The IEA's emergency release of 400 million barrels was a direct policy response to this precautionary demand shock — the releases were designed to signal that future supply would be available, reducing the incentive for precautionary hoarding.

Framework Limitations in 2026

The standard Kilian framework faces several challenges in the 2026 context:

The Dallas Fed's 2026 Hormuz closure scenario model (see Dallas Fed Scenarios) explicitly builds on the Kilian framework but incorporates non-linear dynamics to handle the unprecedented scale of the disruption.

Policy Implications

The Kilian framework's key policy insight is that the type of oil price shock matters more than the magnitude. A $30/bbl price increase driven by supply disruption calls for strategic reserve releases and supply-side responses. The same $30/bbl increase driven by aggregate demand reflects healthy economic growth and should not be policy-distorted. A precautionary demand spike calls for information interventions — credible signals about future supply availability — rather than immediate physical releases.

This decomposition directly informed the IEA's 2026 response strategy. The coordinated 400-million-barrel release was explicitly designed to address the precautionary demand shock component, with the IEA signaling that additional releases were available if needed. The policy calculus was: reduce the fear premium by credibly committing future supply, rather than attempting to fully offset the physical supply deficit.

Sources

Related

oil-shock-factory-kb · https://oil-shock-monitor.pages.dev
kilian-framework.md