Dívida pública e os efeitos sobre o PIB e o Gap do Produto (1994-2018): Modelos ARDL de dados em Painel

Abstract

The present work main goal is to investigate the role of public debt (% of GDP) for the GDP and output GAP for a sample of 56 countries from 1994 to 2018, using PMG ARDL panel data models. The empirical results indicate that the long run coefficients for the GDP models are all negative for the public debt, in other words, higher (lower) levels of public debt tend to reduce (foster) the GDP. For the output GAP models, the long run analysis indicates that all estimation for the public debt are statistically significant with negative coefficients, meaning that higher (lower) levels of public debt tend to affect negatively (positively) the output GAP. For the estimation of the error correction models, it is worth to highlight that the adjustment towards the equilibrium is faster (slower) for the output GAP (GDP), and the difference is of 10% and 23% respectively. The short run estimated coefficients for the GDP models are all negative and statistically significant for the change in public debt, while the coefficients for the output GAP models the public debt coefficients are all negative and statistically significant for most of them. The results for the conventional Granger causality tests and the Dumitrescu-Hurlin (DH) version indicate that there is Granger causality from the public debt to output GAP with two and three lags, and in the DH version, the causality runs from the public debt to GDP for all three lags.