Nonlinear Investing: Commodity

In this chapter, we employ a quantamental approach to build a nonlinear structural model for oil price forecast and commodity investing. We use West Texas Intermediate oil (WTI) as a commodity. We first present fundamental aspects of WTI pricing including a historical look at pricing regimes and evolutions over time. We then break down the total price of WTI into a long-term trend and a short-term deviation, and study the statistical characteristics of each component. It is well known that the causal factors and effects are not the same for long-term and short-term price changes of WTI, which means that a linear model for the total WTI price forecast will miss such nonlinear causality. To capture the nonlinear causality, we follow the quantamental approach and build a nonlinear structural model. This requires an understanding of long-term price trends and short-term price volatility, identifying respective causal factors, and building a long-term model to forecast returns of monthly trends and a short-term model to track daily price deviation. The combination of long- and short-term price forecasts will yield total daily price forecasts, which serve as nonlinear alpha for investing strategies in the oil commodity. We form both an ETF-based strategy without leverage and a futures-based strategy with leverage. Finally, we conduct an out-of-sample study to construct portfolios and analyze their performance. (This chapter is based on a research paper “Forecasting the Price of Oil: A Structural Equation Approach.”)