Hedging Effectiveness of Live Cattle in the State of São Paulo: An Approach Using LSTM (Long Short-Term Memory)

• Grant number: 24/22064-5
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: April 01, 2025
• End date: March 31, 2026
• Field of knowledge: Applied Social Sciences - Economics - Agrarian and Natural Resource Economics
• Principal Investigator: Nicole Rennó Castro
• Grantee: João Guilherme Perotto Fernandes
• Host Institution: Escola Superior de Agricultura Luiz de Queiroz (ESALQ). Universidade de São Paulo (USP). Piracicaba , SP, Brazil

Abstract

Hedging is a financial mechanism that provides security to beef cattle producers by pre-establishing the price of an asset with the aim of selling it at the same price on a future date. However, basis risk, which is the difference between the covered asset's price and the instrument used for hedging , introduces uncertainties regarding returns, as the models employed for price forecasting must be sufficiently robust to ensure accuracy in market positions. Thus, to improve the predictability of returns in cattle trading, this project adopts a neural network-based approach aimed at identifying price trends and making strategic adjustments in response to price fluctuations, thereby optimizing the outcomes of hedging operations for producers. The project proposes simulating the purchase of feeder cattle in May, under the assumption of confinement during the off-season period, to sell them in November, ensuring a liquidation price for the confined cattle, as they would reach 16 arrobas-a weight typically considered ideal for slaughter. The study will incorporate various macroeconomic parameters to strengthen the hedged position, providing greater stability to the producer's assets under different scenarios. Consequently, the study seeks to: (i) forecast the price of finished cattle using LSTM (Long Short-Term Memory); (ii) regularly monitor and adjust the hedge's performance to adapt to price volatility; (iii) analyze the sensitivity and robustness of the model through LSTM adjustments; and (iv) assess the effectiveness of minimum variance hedging in the context of using LSTM in the finished cattle market.