Determinantes dos preços da cesta básica nas capitais do Brasil com base no modelo threshold autorregressivo de painel com efeitos fixos

Abstract

This study analyzes the determinants of the price of the basic food basket in 10 Brazilian capitals from 2016 to 2022, based on a threshold autoregressive model for panel data with fixed effects. Two models are analyzed, with the Broad National Consumer Price Index (IPCA) as the threshold variable and rainfall precipitation and diesel prices as explanatory variables. In the first analysis, the model uses the value of the basic food basket as the dependent variable, and following Hansen's (1999) procedure, the existence of two thresholds is found. The first threshold occurs at 0.53% of the IPCA, corresponding to June 2021, and the second threshold occurs at 1.01% of the IPCA, corresponding to February 2022. The second model uses the number of working hours required to purchase a basic food basket as the dependent variable. A threshold at 0.54% of the IPCA is found, occurring in January 2022. Comparing the estimates of the two models, some consistency is observed in the change in behavior of the dependent variables, which is further supported by the threshold periods. In both models, the inclusion of the lagged dependent variable by one period had a positive impact on the current dependent variable, regardless of the inflation regime. The explanatory variables did not show consistent results across the two models analyzed, but in the basic food basket value model, rainfall increases the price of the basic food basket, and in the working hours model, an increase in fuel prices results in more working hours being required to purchase a basic food basket.