Reading the Future of Oil: A Noncausal Approach to Supply News Shocks

This paper proposes a new strategy to identify oil supply news shocks by combining a Bayesian noncausal structural VAR with a Max-Share approach. The framework jointly resolves the problems of non-fundamentalness and recoverability that undermine standard (proxy) SVAR methods. Exploiting non-Gaussianity in a multivariate Student−t specification, we recover structural shocks from a two sided moving average representation and isolate the expectation driven component of oil supply innovations without external instruments. Applied to global oil market data, the model supports a non fundamental representation and detects anticipatory price and inventory movements consistent with rational expectations storage behavior. The identified shocks explain a substantial fraction of real oil price fluctuations, notably in the late 1970s–1980s and during the 2014–15 collapse, while the COVID-19 episode is predominantly demand driven. Decomposing global supply shows that these shocks are primarily OPEC-driven and generate stagflationary responses in output and inflation, underscoring the central role of expectations in oil market dynamics.