Forecasting coffee exports in Brazil using hierarchical models
The coffee export market in Brazil, the world's largest producer and exporter, faces significant challenges due to the volatility of international prices and variability in regional production. These factors make export forecasting a complex but crucial task for the economic sustainability of the sector. This study aimed to develop and apply a hierarchical model, integrating regional and temporal variables, to forecast coffee exports from Brazil. The research used data from six exporting states (Minas Gerais, Espírito Santo, São Paulo, Bahia, Paraná and Rio de Janeiro) from January 2012 to May 2022. For the analysis, two-level hierarchical models were used to capture fixed and random effects, comparing them with traditional multiple linear regression models. The methodology included transforming the data into natural logarithms, applying lags to dependent variables and using the regional economic activity index (IBCR) as a proxy for state GDP. The analysis was conducted using the R programming language, with specific packages for hierarchical modeling and descriptive statistics. The results indicated that the hierarchical model offered more accurate forecasts, excelling in capturing regional and temporal fluctuations. These forecasts improve planning and decision-making in the coffee sector, as well as reinforcing the need for public policies that consider regional particularities and help mitigate the effects of market volatility. The study confirmed that, contrary to initial expectations, the exchange rate was the main determinant of exports, while regional variables such as the Regional Economic Activity Index (IBCR) played a secondary, albeit significant, role. These findings suggest that economic and trade policies better adapted to regional realities could improve Brazil's competitiveness in the international coffee market.
Forecasting coffee exports in Brazil using hierarchical models
-
DOI: https://doi.org/10.22533/at.ed.973512507019
-
Palavras-chave: economic analysis; Brazilian coffee; hierarchical models; regional economic dynamics
-
Keywords: economic analysis; Brazilian coffee; hierarchical models; regional economic dynamics
-
Abstract:
The coffee export market in Brazil, the world's largest producer and exporter, faces significant challenges due to the volatility of international prices and variability in regional production. These factors make export forecasting a complex but crucial task for the economic sustainability of the sector. This study aimed to develop and apply a hierarchical model, integrating regional and temporal variables, to forecast coffee exports from Brazil. The research used data from six exporting states (Minas Gerais, Espírito Santo, São Paulo, Bahia, Paraná and Rio de Janeiro) from January 2012 to May 2022. For the analysis, two-level hierarchical models were used to capture fixed and random effects, comparing them with traditional multiple linear regression models. The methodology included transforming the data into natural logarithms, applying lags to dependent variables and using the regional economic activity index (IBCR) as a proxy for state GDP. The analysis was conducted using the R programming language, with specific packages for hierarchical modeling and descriptive statistics. The results indicated that the hierarchical model offered more accurate forecasts, excelling in capturing regional and temporal fluctuations. These forecasts improve planning and decision-making in the coffee sector, as well as reinforcing the need for public policies that consider regional particularities and help mitigate the effects of market volatility. The study confirmed that, contrary to initial expectations, the exchange rate was the main determinant of exports, while regional variables such as the Regional Economic Activity Index (IBCR) played a secondary, albeit significant, role. These findings suggest that economic and trade policies better adapted to regional realities could improve Brazil's competitiveness in the international coffee market.
- Gabriel Amil Bastos
- José Erasmo Silva