Comparação entre modelos de preços de commodities: dois e três fatores estocásticos

Abstract

Aiming to refine commodity price forecasting and assist market agents in decision-making, this study proposes a three-factor stochastic model, extending the work of Gibson and Schwartz (1990). In addition to spot prices and convenience returns, the mean of convenience returns is treated as a stochastic factor. To implement the model, weekly futures price data from the New York Stock Exchange between 1996 and 2018 were collected. Additionally, the Kalman filter methodology was employed to estimate the parameters and time series of the unobserved variables. Finally, we compared the two-factor and three-factor models, and the proposed model exhibited superior explanatory power compared to the traditional two-factor model, more accurately capturing the term structure of futures prices. This enhanced performance positions it as a valuable tool for traders, risk managers, and policymakers, aiding in strategic decision-making, risk mitigation, and outcome optimization.