["Journal of Regional Science, EarlyView. ", "\nABSTRACT\nWhile the relationship between public infrastructure and crime location choices is well‐studied in criminology, the impact of crime as a spillover externality on public–private partnership (PPP) infrastructure financing remains unexplored. Leveraging a data set of 542 Private Finance Initiative (PFI)‐funded infrastructures in England and Wales from 2011 to 2018 period, we find evidence that local crime negatively affects government's reimbursement commitment to its PFI's partner, that is, a novel crime–PPP underpayment nexus. We also find evidence that this adverse fiscal resource re‐allocation effect is moderated by both contemporaneous local affluence and the predetermined urban proximity of the infrastructure. Specifically, the former (measured by local house price), while exhibiting a direct “crime attractor” effect, reflects the “nondependent on government to provide security” effect, as rich neighborhoods can afford their own private security and therefore less reliant on public police spending. This mitigates the negative effect of crime on PFI underpayment. The latter, which concerns the physical location of the infrastructure built, is hypothesized to exhibit the “guardian angel” effect, that is, its urban proximity means it is less costly for general public surveillance to take place, hence deterring crime. Finally, we provide evidence that the two moderation effects are rivalry, which contribute to the existing literature pertinent to the trade‐offs between the “crime attractor” and the “guardian angel” effects associated with public infrastructure in crime deterrence. Our findings reveal an additional fiscal resource re‐allocation dimension, even when these concern PFI‐financed projects (for which the governments ought to have a lesser stake).\n"]