Análise dinâmica da relação entre as variáveis macroeconômicas e o mercado acionário brasileiro: uma abordagem do modelo de correção de erros

Abstract

The present study aims to analyze the short and long term relationship between the macroeconomic variables, represented by the interest rate, exchange rate, inflation and country risk, as compared to the brazilian stock market, represented by the Stock Exchange Index São Paulo (Ibovespa), during the period from January 1995 to June 2016. For that, the multivariate approach VAR was used. The implementation of the proposed model was done through the unit root tests of Dickey and Fuller Increased (ADF) and Phillips and Perron (PP), and the Johansen cointegration test. The unit root tests indicated that the series are integrated of order I(1). The cointegration test, both by the trace statistic and by the maximum eigenvalue, indicated the existence of at least one cointegration vector, thus enabling the use of the Vector Autoregressive method with error correction term (VEC). The analysis of the long-term coefficients, described by the normalized cointegration vector, showed positive exchange rate and inflation, and negative interest rate and country risk behavior, with Ibovespa. Analyzing the VEC coefficients, the adjustment parameter (α) was statistically significant, indicating that the Ibovespa reacts in the long-term equilibrium trajectory to changes in the short-term macroeconomic indicators. By the Granger causality test based on the VEC, we observed the existence of short-term unidirectional causality of Ibovespa to country risk. As verified in the Impulse Response Function (IRF), the Brazil risk is the macroeconomic variable that most impacts Ibovespa. The results concerning the Decomposition of Variance of Forecast Errors (VD) also showed the great explanatory power of the Brazil risk on the São Paulo Stock Exchange index, explaining approximately 28% of the index variations. The results obtained from the structural analysis of the VEC make clear the high sensitivity of Ibovespa to the country risk, represented by EMBI + Brazil.