Table 6.
Mechanism (3): State-level public services.
| All crime | Domestic violence | Lapsed alimony | Sexual crimes | Femicide | |
|---|---|---|---|---|---|
| (1) | (2) | (3) | (4) | (5) | |
| Panel A: 1(More than One Violence Shelter in the State) | |||||
| 1(Post-COVID-19) x 1(High # Violence Shelters) | 0.135 | 0.063 | 0.261 | −0.190 | 0.000 |
| (1.479) | (1.190) | (0.394) | (0.231) | (0.017) | |
| 1(Post-COVID-19) | −3.050*** | −1.801* | −0.775** | −0.459*** | −0.015 |
| (1.167) | (0.961) | (0.334) | (0.150) | (0.015) | |
| Panel B: 1(Higher than Median Per Capita Safety Expenditures) | |||||
| 1(Post-COVID-19) x 1(High Public Safety P.C. Expenditure) | −0.812 | −1.120 | 0.302 | 0.010 | −0.003 |
| (1.327) | (1.023) | (0.326) | (0.271) | (0.014) | |
| 1(Post-COVID-19) | −2.590** | −1.255* | −0.721*** | −0.601*** | −0.013 |
| (1.007) | (0.754) | (0.233) | (0.163) | (0.013) | |
| Panel C: 1(Higher than Median Public Safety Employees) | |||||
| 1(Post-COVID-19) x 1(High P.C. Safety Personnel) | −1.530 | −1.455 | 0.088 | −0.156 | −0.007 |
| (1.410) | (1.133) | (0.341) | (0.256) | (0.014) | |
| 1(Post-COVID-19) | −2.158*** | −0.999* | −0.632** | −0.516*** | −0.011 |
| (0.834) | (0.602) | (0.253) | (0.184) | (0.014) | |
| Panel D: 1(MORENA Political Affiliation) | |||||
| 1(Post-COVID-19) x 1(MORENA) | −2.127 | −2.035 | 0.444 | −0.519 | −0.017 |
| (1.719) | (1.262) | (0.318) | (0.365) | (0.013) | |
| 1(Post-COVID-19) | −2.327*** | −1.157* | −0.716*** | −0.444*** | −0.009 |
| (0.768) | (0.642) | (0.176) | (0.070) | (0.009) | |
| Observations | 640 | 640 | 640 | 640 | 640 |
| Pre-Lockdown Mean Dependent | 17.42 | 13.22 | 1.45 | 2.68 | 0.07 |
| Baseline FE | X | X | X | X | X |
| Time Trends | X | X | X | X | X |
Source: Mexico's National Public Security System (Secretariado Ejecutivo del Sistema Nacional de Seguridad Pública). See Appendix B for data sources used for the mechanisms. Notes: Difference-in-differences-differences estimates from Equation (3). Estimates show the grouped post-periods (month three 2020 onward) relative to the pre-pandemic (month three and before). The interacted estimate shows the post-period interacted with the mechanism of interest. Monthly linear time trends are included. State-level aggregate crime rates shown. Baseline fixed effects are included at the state, month, and year. The specification is weighted by the state-level population. Estimation includes January through October over 2019–2020. Robust standard errors are clustered at the state level. Significance levels: * , ** , *** Crimes are measured per 100,000 persons. Alternative specification with continuous measures shown in Table B.5.