6. How to Forecast Commodity Price Movements: Time Series Models

In this chapter, we focus on commodity pricing and investment with time series models. For stock selection strategies, the ability to forecast individual stock returns is critical and usually relies on company-level factors, such as profitability, management quality, etc. In commodity investing, a deep understanding of the geopolitical dynamics and identification of macro-level factors are very important for a successful strategy. A stock selection strategy requires cross-sectional analysis, while commodity investing requires time series analysis. In this chapter, we get into details about the special features of time series models, introduce the concepts of unit root, spurious relationship, and cointegration, and show how they can be employed for quantitative investing in crude oil and pair trading.