Stock markets and their relationship
with the real economy in Latin America
Samuel Brugger *and Edgar Ortiz **
Impulse Response Analysis

Stimulus and response analysis is a useful tool for evaluating the congruence and dynamic sensibility of the variables specified in the model, while ensuring there are no covariances different to 0, that is correlated errors which would make it impossible to determine the response variables before specific variable impulses (Pindyck and Rubinfeld, 2001). The analysis indicates the dynamic response of the dependent variable in the VAR before error term shocks or innovations in all the endogenous variables, excluding the effects of the variables expressly assigned as exogenous (Eq. 11 and 12). It is important to highlight that the impulse response can only be calculated if the VAR is in long term equilibrium. The length of the shock should also be considered because if this is very short, neither the evolution of the shocks nor the dynamic stability of the VAR can be observed. In this way, if the VAR is stable, a disturbance would cause the system to deviate from its equilibrium path, although after a few periods it returns to equilibrium.

The analyses in the present study are based on three year intervals so that the shock process and adjustment can be observed as a complete process. Although the stimulus and response analysis generates four analyses all the variables are examined only the impulse generated by the stock market is examined, as this is the statistically significant relationship, as mentioned in the section on Granger Causality. Also, the question that specifically needs to be answered is if Latin American stock markets have an impact on economic development. The four scenarios can be observed in Figure 4.

In panel (a) in the case of Argentina, all 36 months are statistically significant at 5% and practically all at 1%. The impulse response of the MERVAL begins positively although it decreases during the first three months; there is a negative impact on GDP in the fourth month and a positive one again in the fifth month. The path then zigzags back and forth until reaching stability. The positive impact in the eighth month stands out and the considerably negative impact in the twelfth month.

Panel (b) shows the impulse response analysis of the BOVESPA for the Brazilian economy. As for Argentina, nearly all the periods are significant at 1% and all are significant at 5%. There is an almost equal positive response in the first two months, which becomes negative in the third and fourth months. The results then stabilize at 0 until the ninth month, when the stock market has an impact almost as positive as during the first months, turning negative again in the twelfth month. The response then zigzags before reaching a stable point.

Panel (c) summarizes the impulse response of the IGPA to monthly GDP for Chile. In this case, all the periods are significant at 1%. Impact is slightly positive during the first two months and then slightly negative until the twelfth month when there is a positive effect. The impacts then zigzag, but only slightly. It is possible therefore to confirm that the Chilean stock market has a minimum impact on the real economy, which to an extent contradicts the hypothesis that the stock market is essential for development. The Chilean stock market is the most developed in Latin America and the exemplary performance of the Chilean economy is practically aligned to the countrys stock market activity.

Finally, panel (d) shows the case for Mexico. All the periods are significant at 1%. Although the response is weak, as is the case for Chile, the response is interesting in that it is different to that of the other countries. The response is negative initially at the beginning of the period, positive in the second and negative in the third. The impulse then begins to disappear and zigzag back and forth until it stabilizes.