Intra-Horizon Risk Measures and Continuation Liquidity Requirements in Commodity Futures Markets

Traditional Value at Risk (VaR) and expected shortfall (ES) measures that are based on terminal loss may underestimate the risk induced by interim price movements. In this article, we evaluate the intra-horizon risk measures MaxVaR and MaxES as measures of the continuation liquidity buffer required to sustain futures positions. We propose models with asymmetric jumps and/or seasonal volatility and generate out-of-sample MaxVaR and MaxES forecasts over 10-, 22-, and 66-day horizons. Using data for heating oil, natural gas, corn, and soybeans from 2000 to 2021, we find that considering both asymmetric jumps and seasonal volatility generates the most accurate MaxVaR and MaxES forecasts. Liquidity buffers that are based on these measures also exhibit substantial seasonal variation and significant asymmetry between long and short positions. Furthermore, comparisons with initial margins indicate that continuation liquidity buffers represent a substantial component of the total liquidity capacity required to initiate and maintain futures positions.