Um modelo VAR para uma Avaliação Macroeconómica de Efeitos da Integração Europeia da Economia Portuguesa

Abstract

Procuramos com este estudo fazer uma identificação, seguida de quantificação, de possíveis choques de natureza macroeconómica a que a economia portuguesa esteve sujeita durante os finais dos anos 80. Propomos para isso um modelo resumido para representar o seu comportamento. As variáveis do modelo não excluem a hipótese de raiz unitária não sazonal e não apresentam qualquer relação de cointegração entre elas. O modelo VAR é estimado com variáveis I(1). Com base neste modelo propomonos conhecer as relações que se estabeleceram entre as variáveis macroeconómicas seleccionadas. O modelo exclui a hipótese nula de qualquer uma das suas variáveis e é estável. A análise dos choques permitenos avaliar as consequências de alterações imprevisíveis de cada uma das variáveis e de um mix de variáveis sobre todo o modelo. Num modelo deste tipo, os efeitos extinguemse ao fim de algum tempo e assim a análise apenas pode ser feita em termos de períodos durante os quais os choques se fazem sentir e dos valores transitórios desses choques. O modelo permitenos uma avaliação dos efeitos dos choques a que a economia portuguesa esteve sujeita no seu processo de integração. Verificamos ainda que os resultados desses choques já se encontram esgotados. Finalmente, uma palavra sobre este tipo de simulação: os resultados dos choques não são mais que as consequências dos valores dos choques que entretanto supusemos e que apenas devem relevar do bom senso.

The main objective of this paper is to simulate some possible positive effects of the Portuguese economic integration in the European community. An important by-product of this research is also addressed, the time profile of those effects upon the Portuguese economy. As far as I know, there is only a study of the main effects of the Portuguese participation in the EMU that is rather global and very detailed; and there are also some papers which purpose is to measure consequences of some specific community policies. I have built a VAR model, with 4 variables, and it was applied to quarterly data from 1977 to 2000. I would prefer a model estimated until the mid eighties, but with data only since 1977 it would not be very consistent. My first problem to address was the assumptions to make in order to simulate the positive shocks of our integration. To do that, I propose a very simple macroeconomic analysis of the end of the 70’s and the first half of the 80’s. The main issues are the openness of our economy with a similar behaviour of exports and imports; the financial innovation in parallel with financial liberalization meaning an extraordinary reduction of interest rates; and the inflationary process, not only in terms of the inflation values but also in terms of inflation in consumption and production. As a result, I will suppose that we can measure the quantitative consequences of the Portuguese economic integration only with shocks on production and on inflation. To build my model I have chosen the following four variables: real balances; production; annual inflation; and real interest rate. For the money balances I have used M1; for production the real GDP; for inflation the GDP deflator; and for the interest rate, a bank rate applied to credit from 6 months to one year. All these variables have a unit root. I have used also a dummy for the first quarter of 1993. To know the number of lags I have used a likelihood ratio (with Sims’ correction), and I have obtained 5 lags. For each equation, I can reject the presence of serial correlation and of ARCH from order 1 to 4, and I can accept the normality of the errors. The standard error of each estimation is very small. The errors are stationary. The values of the roots of the companion matrix show that the model is stationary. To test the stability of the parameters I have used three Chow tests, since 1992:2, that is, since the Portuguese presence at the European exchange rate mechanism. I conclude for the stability of the parameters. To test the possible exogeneity of each variable, I have used an ordinary Chi-squared rejection test. There are no reasons to exclude any variable from the model. The variance decomposition says that in general all the variables have a role in explaining the values of each one. It’s interesting to see the importance of inflation and of real interest rate in the explanation of the product. In the paper I analyse the results of shocks over each variable. These results are an illustration of the macroeconomic behaviour of the Portuguese economy. One interesting result is the fact that the very common behaviour of inflation and interest rate in inflationary process is the result of shocks in real balances or inflation and not shocks in the other variables. But as I have said at the beginning, my main interest is the simulation of the positive results of the Portuguese economic integration. So, I have supposed, as representatives’ shocks, an inflation decrease of 10 in percentage values and a production increase of 3%. The maximum effect over the product is obtained after 13 quarters with 19%, and after 33 quarters, the influence of shocks is still of 3,2%. Until 38 quarters, the effects are positives. In what concerns the rate of inflation its decrease is 12,4, after 7 quarters and the decrease goes on until 21 quarters. Suppose also that the shocks are of the type once for all at the first quarter of 1987. In this case, the maximum effect over the product will be at 1990:2, with 19,05%, and the maximum reduction of inflation will be at 1988:3 with 12,4. The positive effects over the production will vanish at 1996:2 and over the prices at 1992:2. The maximum effect over the real balances will be at 1990:1 and will vanish at 1995:4. And finally, over the real rate of interest, the maximum reduction will be at 1989:3, and about 1996:1 the effects contributing to its reduction will still exist. The results are in accordance with the evolution of Portuguese GDP growth rates in the last two cycles. The results also explain the recent performance problems of the Portuguese economy and that positive shocks don’t last forever, even if politicians use to forget such a simple principle.