Real Options in Agricultural Commodities via Bayesian MCMC

Real Options in Agricultural Commodities via Bayesian MCMC

• Autor
• Rodrigo Hermont Ozon
• Co-autores
• José Donizetti de Lima
• Resumo

• Classical real options valuation relies on point estimates of stochastic process parameters, ignoring the uncertainty inherent in parameter estimation. This paper proposes a Bayesian framework that propagates full posterior uncertainty from Markov Chain Monte Carlo (MCMC) estimation into real options valuation for agricultural commodities. We estimate three stochastic processes—Geometric Brownian Motion, Mean-Reverting Jump-Diffusion, and Mean-Reverting Stochastic Volatility—for five agricultural commodity futures (corn, soybean, wheat, live cattle, and soybean meal) using Hamiltonian Monte Carlo via Stan. For each posterior draw, we simulate price paths and compute option payoffs, yielding complete Bayesian distributions of real option values rather than scalar estimates. Results show that parameter uncertainty widens the distribution of timing and abandonment option values relative to classical approaches, with important implications for agricultural investment decisions under uncertainty.

• Palavras-chave
• Real options; MCMC; Agricultural commodities; Stochastic processes; Bayesian inference.
• Modalidade
• Comunicação oral
• Área Temática
• Engenharia Financeira

Voltar Download