Abstract: The literature on housing stratification has ignored the effect of selfselection on housing stratification when residents face changes in the opportunity structure triggered by the housing reform. An analysis of the data from a questionnaire survey of over 1,000 households conducted in Guangzhou in2010 using the Endogeneity Switching Regression Model found the effect of selfselection in the housing stratification mechanisms during the process of marketoriented reform: Capable people within the state system before 1998 tended not to buy houses, and if they did, their housing class status was actually lowered; capable people outside the state system before 1998 and all capable people after 1998 tended to buy houses, and if they didn’t, their housing class status slipped. That is to say, after 1998, the housing class status of those with human or political capital would rise only if they did purchase houses; if not, their housing class status would not improve significantly. Thus, the study concluded that the power and market mechanisms were the structural factors which influenced housing stratification, and that people’s human and political capitals were necessary but not sufficient conditions for house purchasing. To some extent, the judgment that “the higher a person’s human capital and/or political capital is, the higher his/her housing class status is” ignored the role of the person’s selfselection. Of course, the decision making of whether to purchase a house or not might be a very complex selective process. In order to study the selective function in the process indepth, dynamic data including the buyer’s purchasing behaviors and the end results in different social contexts are needed. Because of the flaws in the real estate market and system, residents’freedom to make purchasing decisions is quite limited. These constraints for the current study were beyond the author’s means to overcome.