The cost of silence: unreported crime and homicides as correlates of capital formation in emerging economies

In the context of nearshoring, institutional strength and security have become critical considerations for the competitiveness of emerging economies. Little empirical work disaggregates how different crime types relate to formal investment or tests whether such relationships persist over time. This paper examines the association between large-firm growth and two crime metrics – lethal violence and institutional impunity (‘unreported crime’) – using a panel of the 32 Mexican states, 2016–2025 (2016–2024 with the per-capita GDP control, given data availability). Using fixed-effects models with clustered standard errors, we find that all three time-varying crime indicators – homicide rate, vehicle theft, and unreported crime – show statistically significant negative contemporaneous associations with large-firm growth, with unreported crime showing the largest and most consistent association (β ≈ −0.71 to −0.48, p < 0.05), robust to per-capita GDP, COVID-19, and linear-trend controls. However, none of these contemporaneous associations withstands temporal robustness testing: when crime measures are lagged by one to three years, the relationships reverse sign and lose coherence, and joint state-year fixed effects render all crime coefficients statistically indistinguishable from zero. These results show that, while the contemporaneous associations we document are robust to conventional controls, they lack the temporal persistence expected of a genuine causal or durable mechanism – more consistent with contemporaneous confounding or short-run sentiment effects than a structural deterrent to investment. We discuss the implications of this fragility for the wider crime-investment literature, which often relies on single, contemporaneous specifications without temporal checks, and caution against drawing firm policy conclusions from such evidence alone.