Stock markets and their relationship
with the real economy in Latin America
Samuel Brugger *and Edgar Ortiz **
THEORY AND EVIDENCE ( ...continuation )

Evidence related to stock markets and development is abundant and controversial. In a pioneering study, Fama (1990) used future production to explain, with different frequencies, stock market yields in the United States between 1953 and 1987. His study demonstrates that the real economy is the cause of stock market yields, which differs from the majority of other works that suggest the reverse. More directly, the study by Hassapis and Kalyvitis (2002) identifies positive relationships between the changes in share prices and the growth rate of countries in the OECD for the period between 1950 and 1990 in the majority of countries in the OECD except Italy and France. Nasseh and Strauss (2000) also present evidence of a long term relationship between share prices and industrial production in the major European economies.

For countries with underdeveloped stock markets, there are few studies and the evidence is even more contradictory. In a cross sectional study of more than forty countries, Atje and Jovanovic (1993), present evidence of a correlation between average economic growth and the capitalization of stock markets. Levine and Zervos (1998) take the study further including more variables and lengthening the study period, and find that the liquidity of the stock market is related significantly to economic growth, capital accumulation and productivity, but not with the size of the capital market. Mauro (2003) looked for a correlation between production growth and stock market profits with lags in emerging and developed economies. His study showed that the proportion of developed and underdeveloped countries in which the correlation is significant, is the same if annual data is used, but is not significant if quarterly data is used. Caporale, Howells and Soliman (2003 and 2005) used a VAR model to show causality and the results showed that the development of the stock market improved economic development in Chile, Korea, Malaysia and the Philippines long term, particularly through the impact of investment. Choong, Yusop, Law and Sen (2003) demonstrated through an Autoregressive Distributive Lag (ARDL) and Granger causality test that there was co-integration between the development of the stock markets and economic growth for Malaysia from 1978 to 2000. Nyong (1997) used a development capital market index in Nigeria and found that development is negatively correlated with long-term growth in this country.

In summary, traditional financial and development theory postulates a direct correlation between the financial sector and economic development. This theory has been recently extended to suggest a direct link between stock market performance and economic growth. In both cases, alternative views have been put forward and empirical evidence has produced results that endorse both views or demonstrate two-way impact between finance and development. Nevertheless, empirical evidence reveals the need to strengthen research on the correlations between stock markets and long term development, particularly for emerging economies, including Latin American countries, for which financial literature is scarce.


Time Series

To examine the long-term relationship between stock markets and the real economy in Argentina, Brazil, Chile and Mexico, this study utilizes monthly time series of GDP and stock market indices for the period 1993.1 to 2005.12. The set of financial time series was obtained from the database Economtica and the set of economic time series and the GDP for each country, from the database CEPALStat. The financial series consists of stock market indices for Argentina (MERVAL), Brazil (BOVESPA), Chile (IGPA), and Mexico (IPC). Both the financial series and those of the real economy are standardized to a base of 100=1993. The series consists of a total of 196 observations.

Econometric Modeling

The methodological framework applicable consists of six econometric models, which aim to establish a comprehensive picture of the relationship between the stock markets and economic development in LA. These tests are: (a) unit root analysis; (b) co-integration analysis; (c) error-correction mechanism tests; (d) Granger causality modeling; (e) impulse response functions; and (f) autoregressive vector analysis.

Prior to any econometric test, it is useful to determine if the series are stationary so as not to make spurious correlations between the variables under analysis, a problem that occurs in financial time series with rising tends around the mean and variance. In order to determine if there is a balanced long term relationship between stock markets and the real economy it is necessary to remove the unit roots from the series and perform co-integration tests. The augmented Dickey-Fuller unit root test for a time series consists of applying a lineal regression model where the first differences of the series are the dependent variable and applying k lags of the time series as independent variables.

( 1 )