# Impact of fundamental variables on Mexican stock returns

This dissertation analyzes the cross-sectional variation in monthly average stock returns explained by the behavior of fundamental variables such as earnings to price ratio (E/P), book-to-market equity ratio (B/M), and size of market equity series (ME) in the Mexican stock market The Seemingly Unrelated Regression (SUR) model, using methodology similar to that of Chan, Hamao, and Lakonishok (1991), is used first to test the significance of the fundamental variables at the portfolio level, adjusting at the same time for portfolio risk where the betas are estimated simultaneously with the impact of fundamental variables. I find that E/P is the most statistically significant and economically important of the fundamental variables with a positive impact on expected stock returns for the 1989 to 2001 period. The size of the market equity series has a negative risk premium, but size is not significant after the adjustment in the level of prices in the fundamental variables for the 1989 to 2001 period. The B/M variable turns out to be highly dependent of the model and portfolio formation for the significance. The risk premium of B/M changes from positive to negative and several times is not significant for the 1989 to 2001 period The January effect versus the other months in relation to the fundamental variables and stock returns is also analyzed using SUR. I find that E/P is also the fundamental variable that has the largest average risk premium and significant, with a positive relation between expected returns and E/P Finally, a portfolio formation process similar to that of Fama-French (1992), using size-beta portfolios, is applied to study the relationship between fundamental variables, beta and the stock returns using SUR. Beta is previously computed with the monthly value-weighted index obtained from the sample. The sum beta is used for nonsynchronous trading. I find that average stock returns are positively related to market beta, but E/P is additional priced with a positive risk premium in that model with beta and the fundamental variables. E/P and beta are the only variables that are statistically significant for the period of 1989 to 2001