Relações entre variáveis econômicas e produtivas de soja no Brasil via correlação canônica

Abstract

This work aims to identify and understand the relationships between productive and economic variables that enabled and influenced the growth of soybean production in the country, which recorded an increase of 115.01%, according to data from the National Supply Company - Conab, from 2006 to 2021. To achieve this objective, multivariate statistical techniques were used, specifically Principal Component Analysis (PCA) and Canonical Correlation Analysis (CCA). These analyses were conducted based on a dataset composed of 166 observations, covering the ten largest Brazilian states involved in soybean production over the mentioned period. Therefore, there are two sets of variables: economic variables (X) and soybean production and trade variables (Y). The identification of data characteristics was performed through PCA, followed by the use of CCA to identify the economic variables related to soybean production variables. Thus, one of the canonical functions showed a significant canonical correlation of 98.57% and explained 75.28% of the joint variability of the data. PCA provided an in-depth view of the internal structure of the data, highlighting patterns of variation and identifying less relevant variables. CCA, in turn, provided a multivariate analysis of the relationships between variable sets, revealing substantial correlations and explaining a large part of the joint variability of the data.