Próxima Página |
Página Anterior |
Índice2 WAS THERE A REAL SPILLOVER EFFECT?
Economies in Asia and Latin America are related through trade and direct investment. When markets open in one region, the fundamental information generated overnight during the business day in the other region are appraised by traders and analysts and incorporated in stock prices. These prices are affected by both macroeconomic data and industry news generated during the night. Global portfolios will also be rebalanced in accordance with the new information revealed. The general relevance of information flow increases in pace with the growing financial and commercial integration of the world markets.
Asset price moves in a country, however, may affect
those in another beyond what is justifiable by economic fundamentals. During
volatility upheavals, high frequency noise trading may overflow international
borders, particularly if the countries involved are perceived as being
similar. In a sense, this may be a partial explanation for what happened
in 1997, when similarities between emerging countries in Latin America
and Asia, both regions highly dependent on foreign capital, for some time
may have overruled their individual macroeconomic contexts in investors
minds. In such instances, when returns and volatilities from markets with
no overlapping trading hours are correlated, lagged spillovers may occur.
Foreign daytime returns or volatilities can determine, to a certain extent,
next day´s domestic returns and volatility.

In the following sections we try to detect this spillover effect, namely any statistical precedence of Asian returns and volatility over Latin American ones. The latter will be represented by the Brazilian market, which accounts for more than 40% in the MS EMF Latin America Index, and closely follows its moves (see Figure 5). Our analytical tool will be the Granger Causality Test, a standard statistical test for causality between two time series. Section 2.1 defines the series of Ibovespa (the São Paulo Stock Exchange index — Brazil’s most liquid market) overnight returns. The next section explains the Granger test and its premises in our context. Section 2.3 presents the results obtained from running the test with the series of returns. Finally, Section 2.4 presents the results obtained from running it with the series of squared returns, a proxy for the volatility.
Próxima Página |
Página Anterior |
Índice