Stock markets and their relationship
with the real economy in Latin America
Samuel Brugger *and Edgar Ortiz **
Basic Statistics

The main characteristics of GDP growth and yields from stock market indices in Argentina, Brazil, Chile and Mexico, are summarized in the following paragraph. The resulting trends are standardized in US dollars for the purpose of comparative analysis and for the purpose of investors and international economic agents.1

Starting from a common base of 100=January 2, 1993 in the panel of Table 1 it is evident that the Chilean economy experienced the highest economic growth of the four countries; its monthly average for the period under analysis (1993-2005) was 0.59%. On a descending scale, Brazil follows with a GDP of 0.33%, followed by Argentina (0.32%) and with the least growth in GDP, Mexico (0.31%). The dispersal of growth in aggregate production of the major economies in LA do not follow the global GDP growth rate pari passu. The first two places are reversed; the standard deviation of Brazils GDP growth is the greatest (8.129 points), followed by Chile (7.697 points) and the last two positions are also reversed; Argentina takes third place, her standard deviation in GDP growth being 4.08 points, whereas GDP growth volatility in Mexico registers a standard deviation of 6.413 points.

The asymmetry statistic in Table 1, panel a, reveals that the GDP growth series for Brazil, Chile and Mexico show a positive bias, whereas the series for Argentina shows a negative bias. The most positive bias is for Chile (0.1.286) and Argentinas negative growth bias is only -0.285 points. What stands out in relation to the kurtosis test is the fact that the distributions of the GDP growth series for Argentina and Chile are leptokutic (their value is greater than 3), i.e. the distributions are pointed. The distribution of Mexicos GDP growth is clearly mesokurtic, i.e. it flattens out in relation to the normal curve; Brazil is found on the inferior limit of the leptokurtic distribution. Finally, measuring the normality of the series according to the Jarque-Bera Test in Table 1, panel a, shows that in the case of Argentina and Chiles GDP growth, their statistics are greater than the critical value of 5.9 at a confidence rate of 5.0%, indicating the absence of normality; the growth of Chile is the furthest from normality. The Jacque-Bera test suggests normality in the cases of Brazil and Mexico.

Examining the returns, San Paulos stock market was the most profitable owing to its inflationary growth, registering a monthly profit of 5.03%. Next on a descending scale, is Mexico with profits of 1.37%, followed by Chile (0.75%) and finally registering the least growth the Argentinas MERVAL (0.11%). The behavior of volatility is a little different. In first place is the Brazilian market with a standard deviation of 51.48 points, followed by the MERVAL of Buenos Aires (11.92 points), then the IPC of Mexico (11.52 points) and finally the least dispersed, Chiles stock market (7.57). These measurements indicate high volatility in LAs capital markets with a big difference in the profit-risk relationship between these regional stock exchanges. Effectively, in terms of the standard deviation, the profit per point of risk assumed, measured by the profit/deviation coefficient standard is better in the case of the Mexican stock market (0.119%) followed by Chile (0.0989%) and Brazil (0.0977%) and finally Argentina (0.00977%).

Regarding asymmetry, what particularly stands out is the fact that three of the four markets analyzed have a negative bias (Argentina, Brazil and Mexico) and only Santiago de Chiles stock market has a positive bias, suggesting far from normal distributions in all cases. This fact is confirmed by the kurtosis statistic, which for all the markets under analysis is under 3, indicating the existence of sharp distributions. However, in the case of Chile, with a kurtosis of 3.5, the Jarque-Bera test does not reject the hypothesis of the absence of normality. In the case of all the other stock markets, this statistic is greater than the critical value of 5.9 confirming the absence of normality. In the case of Chile, the apparent contradiction can be explained by the presence of the positive bias, close to zero. Equally the kurtosis previously identified is low.

1 A table of basic statistics is available from the authors.