Aplicación del modelo de Haavelmo para el análisis económico de la exportación de banano del Ecuador

Abstract

At present, there is no differential economic model to evaluate the Ecuadorian economy; therefore, the objective of this research was to apply Haavelmo’s economic growth model to Ecuadorian banana exports, in order to validate its adjustment and forecast levels. Therefore, as a methodology, real export data of this product were identified, in monthly and annual accumulated, by Free on Board (FOB) and Metric Tons (MT). The model variants were codified in MATLAB language, the simulation of these by the numerical methods of Euler and Runge - Kutta (RK4) and the local optimization to find the unknown constants, by means of three objective functions: adjusted R2 coefficient of determination (R2ajust), Pearson’s coefficient (Pxy) and the sum of Quadratic Residuals (SRC As main results, 48 experiments combining the above factors were analyzed. The combinations with acceptable fit and forecast were: (1) Euler method, linear variant, by R2adjust, with monthly FOB and MT and annual FOB, (2) RK4 method, linear variant, by SRC, with monthly FOB and MT, and (3) Euler method, non-linear variant, by SRC, with monthly FOB and MT. The unknowns found with the Haavelmo model were conceptually corroborated, where the rate of growth was always greater than the rate of decrease. Similar happened with the elasticity of production (a), where it was low to estimate FOB and MT monthly, linear variant, according to historical facts, such as COVID - 19 and Fusaruim Fungus race 4. It is concluded that the Haavelmo model, in its linear variant, by the Euler method, has preliminary potential to estimate the dynamic trajectory and forecast FOB and MT in banana exports. It is recommended to replace local optimization by global optimization, in order to find more accurately the unknown constants of the Haavelmo model.