Search

We propose an ARCH model of the TGARCH type with an asymmetric Student's t distribution. It is built using the methodology of Fernandez and Steel (1998) and the traditional TGARCH model developed by Zakoian (1994). The model is used to describe series of stock market returns and to assess the validity of the rationality hypotheses in Latin America. The results suggest that: 1) The series can be described adequately with the proposed model; (2) the Samuelson´s rationality hypothesis is consistent with the evidence of the markets of Argentina, Brazil, Chile, Colombia and Mexico; 3) the traditional rationality hypothesis is consistent with the evidence of Peru; and (4) the volatility estimated with the proposed model are higher than those estimated with the traditional TGARCH model over the period 2008-2009.

We develop a GARCH model with autoregressive conditional asymmetry to describe time-series. This means that, in addition to the conditional mean and variance, we assume that the skewness describes the behavior of the time-series. Analytically, we use the methodology proposed by Fernández and Steel (1998) to define the behavior of the innovations of the model. We use the approach developed by Brooks, et. al., (2005), to build it. Moreover, we show its usefulness by modeling the daily returns of the Mexican Stock Market Index (IPC) during the period between January 3rd, 2008 and September 29th, 2009.

We develop a comparative study using the TARCH and EGARCH non-linear econometric models. We use them to describe Mexican stock market returns. We model daily series of returns for 30 stocks and the Stock Market Index (IPC) for the period between December 7, 2005 and August 1, 2011. Most of the series show leverage effects. The results also suggest that the AR(1)-EGARCH(1,1) model describes properly the aggregated returns of the stock market (measured by the IPC). They also show that the AR(1)-TGARCH(1,1) and AR(1)-EGARCH(1,1) models fit 19 and 11 stock return series, respectively. Finally, the results show that the return mean (variance) has decreased (increased) since August 2007.

We develop an investigation regarding the determinants of the stock prices listed in the Mexican Stock Exchange (BMV). We use the valuation Ohlson model and an extension of it. The Ohlson-Beta model includes the Beta elasticity as an additional explanatory variable. We use time-series and panel-data cointegration methodologies to assess the existence of meaningful long-run relationships among the variables postulated by both models. The main results suggest that the use of panel-data techniques may be more adequate than time-series ones. They also show that both Ohlson models are useful to describe stock prices. The traditional Ohlson variables are significant and have the postulated signs. The Beta variable is also significant. These results hold when the firms are considered as a whole and for the firms with long cycles. Thus the results may support the hypothesis that the information content of accounting and financial variables depends on the length of the cycles.

We develop an investigation regarding the determinants of the stock prices in six Latin American emerging markets (Argentina, Brazil, Chile, Colombia, Mexico and Peru). We test the traditional Ohlson model and an international version of it. The international model includes the Dow Jones index as an additional explanatory variable. We use time-series and panel-data cointegration methodologies to assess the long-run relationships among the variables postulated by both models. We use quarterly data for the period 2000:01-2010:03. The results suggest that panel-data techniques may be better than time-series ones for the assessments. They support the use of the Ohlson models and, specially, the international one. The variables are significant and have the postulated signs. These results hold when the firms are considered as a whole and for the commercial and construction firms. Furthermore, the results also suggest that the Latin American asset prices are complementary to the US ones in the long run.