Geopolitical risk and forecastability of tail risk in the oil market: Evidence from over a century of monthly data

Highlights: • Variants of the conditional autoregressive value at risk (CAViaR) are used to measure oil-market tail risk. • A heterogenous-autoregressive-tail-risk model is used to link tail risk to geopolitical risk, threats, and acts. • Full-sample results show that threats increase tail risk and actual acts reduce tail risk at longer forecast horizons. . • Out-of-sample results show that threats are statistically significant predictors of oil-market tail risk. Using monthly data for the period from 1916 to 2020, we report that geopolitical risk, when decomposed into threats and actual risk, has predictive value for tail risk in the oil market. When we study the full sample of data, we find that threats increase tail risk in the oil market, while actual acts related risk reduces tail risk at longer forecast horizons. While the findings of the full-sample analysis show that the effect of threats and acts on tail risk in the oil market is quantitatively small, results of an out-of-sample analysis show that, for several model configurations, geopolitical risks associated with threats are statistically significant predictors of tail risk in the oil market, even after controlling for a factor capturing global equity-market tail-risk spillovers. Our results have important investment implications.