The power of investors’ optimism and pessimism in oil market forecasting
通过构建市场扩张、市场体制和市场流动性三个核心指标，研究投资者乐观与悲观情绪在石油市场预测中的作用，发现市场存在长期记忆和信息低效，且该预测方法优于传统模型。
By modelling dynamism in the global oil market by three essential market-centric observables (viz., Market Expansion, Market Regime, and Market Liquidity), we study forecasting potential of the future oil markets within a memory-driven interdependence setting. We combine spot prices with our derived market proxies to produce representative global market proxies. The latter are used to quantify the extent of static and dynamic persistence within the system. Our extensive empirical investigation exploits the rich features of fractionally cointegrated vector autoregression, where the rate of disequilibrium error correction within the system is modelled to be slow, approximating real life system dynamics. An advantage is that it explains why we often experience a slow response of a policy intervention. We present robust evidence of both system-wide long-memory and a long-memory in the market-centric observables. We introduce a memory of memory estimation to discern the magnitude of the relative rate of acceleration/deceleration of shocks within each observable, which reflects on the overall stability of the system. Our results show significant degree of non-linear error dissipation and high degree of informational inefficiency. Rigorous out-of-sample forecasting exercise produces robust predictions and demonstrate superiority of our approach.