What does peer-reviewed research say about the impact of sanctuary policies on local crime rates?
Executive summary
Peer‑reviewed and working‑paper evidence consistently rejects the simple claim that sanctuary policies raise crime; a plurality of rigorous studies finds either no effect on violent crime or modest reductions in property and robbery rates, while a smaller set of analyses documents heterogeneous local effects and measurement caveats [1] [2] [3] [4]. The literature remains limited in scope, varies in how “sanctuary” is defined, and points to mechanisms — greater community trust in police and changes in reporting — that complicate interpretation of headline crime statistics [5] [6].
1. No overall uptick in crime — the modal finding across studies
Multiple broad analyses conclude sanctuary policies do not cause increases in crime: a PNAS paper compiling city‑level policy variation reports no effect on violent crime and an approximately 7% decline in property crime in treated jurisdictions [1], and several city‑ and county‑level econometric studies similarly find no evidence of higher crime following sanctuary adoption [2] [7] [8]. Working papers and peer‑reviewed articles by different researchers, using difference‑in‑differences and fixed‑effects models, reinforce the common result that sanctuary status is not associated with rising violent or property crime overall [3] [2].
2. Modest declines in property and robbery in some studies — not a uniform effect
A recurring pattern is modest crime reductions in property‑type offenses after sanctuary adoption: Gingeleskie and others report declines in property crime [1], Otsu and related work find decreases in robbery rates particularly in precincts or neighborhoods with high immigrant concentrations [4] [3], and some pooled county analyses show sanctuary practices strengthen negative links between foreign‑born Latino shares and property crime [9]. These findings suggest benefits are concentrated in certain offense types and geographic subunits rather than uniformly distributed across all crimes [4] [9].
3. Mechanisms matter — trust, reporting, and selection
Scholars advance two interlocking mechanisms to explain why sanctuary policies correlate with equal or lower crime: improved trust between immigrant communities and local police increases cooperation and informal social control, reducing crime or improving enforcement [2] [1], while sanctuary rules may also change reporting behavior — increasing recorded incidents for offenses previously hidden — which can make measured crime trends ambiguous [6] [4]. Researchers warn that heterogeneous selection into sanctuary status and changes in local enforcement practices complicate causal interpretation, so mechanisms must be tested directly rather than inferred solely from aggregate crime data [5] [9].
4. Heterogeneity and methodological caveats — why studies differ
Results diverge when researchers use different definitions (detainer refusal, municipal ordinances, or broader “immigrant‑friendly” measures), units of analysis (precincts, cities, counties), and empirical strategies (matching, conditional fixed effects, pooled cross‑sections), producing heterogeneity in effect sizes and statistical significance [3] [9] [5]. Review articles emphasize the limited number of high‑quality studies and warn that simplistic cross‑city comparisons can mislead because sanctuary jurisdictions differ systematically in demography and governance from non‑sanctuary places [5] [7].
5. What the evidence does not settle — open questions for policy and research
Peer‑reviewed work settles the narrow question that sanctuary policies do not reliably increase crime and often coincide with reductions in property and robbery, yet it does not fully resolve whether those crime changes are driven by causal deterrence, improved policing relations, reporting artifacts, or compositional shifts in populations [1] [6] [9]. The literature’s limits — inconsistent definitions, varying geographic scale, and relatively few longitudinal case studies — mean policymakers should treat aggregate crime findings as one piece of evidence and support further research on mechanisms, local heterogeneity, and long‑run fiscal and social effects [5] [2].