Valuation of maize storage facility

Abstract

Grain supply in Kenya depends strongly on weather, i.e. the seasons of the year. This implies that grain prices such as maize depend on the season of the year and overhead costs like storage and transportation. The analysis of historical spot prices of maize and weather shows a dependency between day-ahead prices and seasons of the year, especially around the harvesting period (October to December). Typically, the maize supply increases during this period. This study advocates for the adoption of a stochastic model that intertwines maize spot prices with the seasons, focusing on the assessment and risk mitigation of maize storage facilities. This is achieved through a spot-based valuation framework coupled with a financial hedging strategy, implemented via futures contracts. The key contributions of this study encompass proposing a comprehensive model that captures the dynamics of the futures curve and spot prices while accommodating essential characteristics of the commodities market, such as seasonality and sporadic spikes in the spot market. Additionally, the study addresses the evaluation of associated model risk. Employing both “intrinsic” and “extrinsic” valuation techniques, this study notably utilizes the “rolling intrinsic” valuation method. This approach considers both spot and future prices of maize in the valuation process, revolving around a trading strategy where a trader secures spot and futures positions by solving an optimization problem based on market information on the first day. The trader retains the flexibility to adjust positions over time in response to new market information, with the storage value calculated as the cumulative sum of the initial day’s value and subsequent added values. To solve the problem, this study formulates the Bellman equation and employs a recursive solution through Monte Carlo simulation with ordinary least square regression.

Keywords: Spot price; futures; valuation, hedging; “intrinsic” and “extrinsic” valuation; optimization; Bellman equation; Monte Carlo; ordinary least square regression.