Residential burglary and concentrated disadvantage: A spatial heterogeneity analysis in Mexico City

Previous empirical studies on the correlation between residential burglary and concentrated disadvantage (CD) in Latin America commonly omit the spatial elements of the relationship. Using Mexico City (CDMX) residential burglary data for the period 2016 to 2018, we examine the predictive capacity of concentrated disadvantage in relation to residential burglary patterns, using a Geographically Weighted Regression approach to check whether their correlation varies across CDMX police quadrants. Controlling for relevant structural variables associated with residential burglary in previous studies, we find that the relationship between CD and residential burglary is positive in 844 out of 846 police quadrants (99.7%) and significantly much steeper in some quadrants than others –up to four times the median local slope. Thus, one key implication is that as this relationship is affected by spatial heterogeneity, traditional regression-to-the-mean analyses may misinform evidence-based crime prevention policies.