This paper provides a description of some of the empirical regularities for the Mexican business cycle. The purpose is to have a benchmark for assessing dynamic stochastic general equilibrium models for the Mexican case. We follow the Kydland and Prescott methodology to describe the cyclical properties of the Mexican business cycle. We describe the volatility of several macroeconomic variables as well as their correlation with GDP. We use two filters to remove trends from the data: Hodrick-Prescott filter and Baxter-King filter. The idea is to check the robustness of the results. Qualitatively the findings are similar across both filtering methods. In order to analyze changes in the properties of the Mexican business cycle, the whole period of analysis is divided into two sub-periods, 1980-1995 and 1996-2006. The first sub-period is marked by high economic instability and the second and most recent sub-period is marked by a significant decrease in the volatility of all variables.