A novel research strategy of measuring housing disadvantages of vulnerable populations for all income levels: the Propensity Score Matching approach

Housing is an important inclusion factor that reflects the degree of stability of vulnerable groups in terms of ethnic or racial origin, who face several difficulties to access good housing. Many studies focused on the marginalisation of this social groups identifying important problems and policy recommendations. Though, there is still plenty of opportunity to broaden the scope of the existing evidence. We propose an analytical framework that solve one of the main methodological problems of the existing studies: the selection bias. The proposed the Propensity Score Matching (PSM) technique in combination with the use of a housing quality indicator (HQI) with a continuous numeric scale. This combination offers several possibilities to tinge the results of former studies. The PSM method isolates that part of the difference in housing quality—defined here as the housing quality gap (HQG)—of the vulnerable people that is directly caused by their ethnic or racial origin from the fraction generated by other structural socio-economic differences between them and the rest of the Mexican population. In other words, it assesses whether the poorest vulnerable individuals suffer the same level of disadvantages as the non-vulnerable in the same socio-economic situation. To broaden the scope of existing studies we quantify the size of the HQ gap for each vulnerable person, to identify the profile of those who have a larger HQ gap than others considering not only the most common aspects (sex, age, educational level, household income) but also some characteristics barely used (housing tenure, participation in public social program, mental of physical disabilities, remittances, housing loan, etc.). This approach permits to study not only the poorest sectors of the population, though we can analyse whether members of the vulnerable population in terms of ethnic or racial origin with higher income levels also suffer worse housing conditions due to their vulnerability and create the corresponding profile of the most affected. Analysing the Mexican situation for the indigenous population and Afro-Mexican, we observed a negative gap for indigenous people, being wider for elderly, renters, single parents, those with higher educational level, those living in the so-called vecindades (poor multi-family dwellings) or house with no floor covering and those residing in the Central City or the largest localities. The model evaluates also some barely used aspects that intersect with ethnic vulnerability showing a wider gap for native speakers and those with physical or mental disabilities.