Abstract
Abstract Purpose Crime-and-place research increasingly distinguishes residential context from the populations that enter neighborhoods through daily activity. This study examines whether inbound routine activity mobility is associated with recorded theft differently across Shanghai communities with different hukou-based residential compositions. Methods This study links 2010 Population Census measures, 2016 China Telecom commuter and tourist inflows, and geocoded court-judgment-derived theft records for 4,067 communities. Logistic regression models recorded theft occurrence, and negative binomial regression models 2017 theft counts using 2010 census usual-resident population as an exposure. Standard errors are clustered by street, and conditional relationships are evaluated with average marginal effects and adjusted predictions. Results Non-hukou resident share and inbound mobility are each positively associated with theft occurrence in additive models, but the occurrence interaction is not estimated precisely. For theft intensity, inbound mobility is positively associated with theft counts, and its association becomes more positive as the non-hukou share increases. The count-model interaction remains stable across alternative count and spatial specifications, while exploratory tests provide no evidence of systematic differences across broad urban zones. Conclusions The findings suggest that dynamic activity-space exposure may be more informative when interpreted together with the residential context receiving those flows. In Shanghai, hukou-based residential composition provides one institutionally structured dimension of that context. The results concern community-level associations in recorded theft and do not establish causal mechanisms or imply criminogenic characteristics of non-hukou residents or mobile populations.