State-Space Modeling of Mexican Agave Price Dynamics

Abstract The Mexican tequila industry depends heavily on Agave tequilana Weber var. azul, a crop with a long biological cycle and a highly volatile market. This study develops dynamic state-space models to analyze and forecast fluctuations in the mean price of blue agave at both national (Mexico) and regional (Jalisco) levels. Using annual data from 1999 to 2023, we apply subspace system identification (MOESP variant) and Kalman filtering to estimate and validate discrete-time models. Model selection is based on structural interpretability and a battery of statistical tests, retaining only models that passed all validation criteria at a 15% significance level. The final national model is second-order and highlights inflation and the MXN/USD exchange rate as dominant drivers—contrasting with previous literature focused on climatic or agricultural variables. In Jalisco, several valid first-order models were identified, often involving production value, suggesting regional-specific dynamics. The proposed framework offers a robust alternative to traditional static econometric models, enabling short-term forecasting and informing decision-making under uncertainty in agro-industrial systems affected by biological lags and economic shocks.