Correlação poupança-investimento: o que é possível dizer sobre o modelo de correção de erros a partir dos dados brasileiros

Resumo

The purpose of this paper is to determine if the error correction model is superior to the other time series tests (cointegration between saving and investment and stationarity of the current account) which focused the long run relationship in evaluating the degree of capital mobility, as suggested by Jansen (1996). Besides it seeks to determine which kind of bias results from the estimates of the short run correlation, with variables in levels or first differences. Using Brazilian data for the period 1960-1996 as example all the tests point in the same direction, that is, capital mobility. The same result was previously found for the Brazilian economy using different methodologies. When the short run relationship is estimated, the regression in levels implies a high bias while the regression in differences implies no bias.