Seasonal Effects in the Mexican Stock Market

Resumen

Starting in the mid-1960s, and for most of the next two decades, financial theory was based on the trinity of a single factor pricing model, the non-predictability of the returns of securities and the three levels of market efficiency proposed by the Efficient Markets Hypothesis (EMH). However, since the 1980s the challenges to the EMH have increased in view of several findings that contradict its propositions. Among these findings are the so-called market anomalies, including several calendar effects, related to the predictability of stock returns.

The purpose of this research is to empirically test for the presence of six calendar anomalies in the Mexican Stock Exchange (BMV), more specifically, the day-of-the-week, holiday, turn-of-the-month, month-of-the-year, January barometer, and Halloween effects. The analysis will include the returns of the Price and Quotation Index (IPC), and the large capitalization (IPC LargeCap), medium capitalization (IPC MidCap) and low capitalization (IPC SmallCap) indices of the BMV. The period under study for the IPC will be from January 2000 to September 2018, and for the size-related indices the period will be from November 2006 to September 2018.

This research will apply ordinary least squares models with dummy variables when the ARCH–LM test fails to reject the hypothesis of homoskedasticity of the error term. Otherwise, GARCH (p, q) models will be applied to determine the conditional variance or the disturbance. Also, TGARCH and EGARCH models will be used in order to examine for possible asymmetries in the volatility clustering of the return series.