We use data from Bogotá and Medellín to describe key quality of life indicators of each city and illustrate their spatial segregation at the census sector level and present evidence that the main two Colombian cities are highly spatially segregated according to their education levels and access to education, coverage of public services, households headed by women and key demographic variables like their levels of adolescent pregnancy. Not surprisingly, our estimated quality of life indexes resembles the mentioned segregation patterns in each city. We present evidence that the spatial agglomeration is statistically significant for each of the variables enumerated. We estimate hedonic models of house values and life satisfaction for Bogotá and Medellín and find that the importance of the average level of education at the census sector level to determine house prices is striking. We also compare hedonic models for Bogotá and Medellín. Bogotá is better endowed than Medellín in the variables included in the analysis, in particular, it has higher education levels, and additionally, education is more equally distributed within census sectors. Bogotá has also better access to gas and has in general houses with better conditions. The models based on house values and life satisfaction approaches used in this article lead to similar conclusions in the aggregate when comparing their implied quality of life indexes. Although each approach allows us to rank the specific determinants of quality of life, and these determinants depend on the approach, their implied aggregated indexes suggest that they are just different faces of the same story. From a policy perspective, the evidence suggests that redesigning the current socioeconomic stratification system in a way that still allows reaching the poorest while preventing segregation to deepen, might be among the most important challenge to face in order to improve quality of life in main Colombian cities.
