Skip to main content

Table 8 Robustness checks: lagged crime and migration patterns

From: Violence and migration: evidence from Mexico’s drug war

Panel A: domestic migration

 2nd stage IV results

State level panel (annual)

 Migration type

Another state

Anywhere in Mexico

 

(1)

(2)

(3)

(4)

 Homicides, 1-year lag

−0.00758*

 

−0.01709

 
 

(0.00401)

 

(0.01070)

 

 Homicides, 2-year lag

 

−0.01132**

 

−0.02268

  

(0.00524)

 

(0.01423)

 Observations

192

160

192

160

 Angrist-Pischke F test

7.975

10.21

7.975

10.21

 Anderson-Rubin LM ChiSquared

8.145

10.26

8.145

10.26

Panel B: international migration

 Geographic and time variation

Municipal panel (annual)

State panel (annual)

 

(1)

(2)

(3)

(4)

 Homicides, 1-year lag

−0.00703**

 

0.00495*

 
 

(0.00284)

 

(0.00267)

 

 Homicides, 2-year lag

 

−0.01096**

 

0.00573*

  

(0.00451)

 

(0.00310)

 Observations

11,890

9512

192

160

 Angrist-Pischke F test

12.47

10.49

7.975

10.21

 Anderson-Rubin LM ChiSquared

17.20

12.96

8.145

10.26

  1. Robust standard errors in brackets
  2. ***p < 0.01, **p < 0.05, *p < 0.01
  3. Lagged homicides instrumented with kilometers of federal highways multiplied by a one period lag in cocaine seizures
  4. All regressions include state or municipal controls and year fixed effects
  5. Source: Mexican Census, as accessed through IPUMS, ENOE, INEGI, Colombian Ministry of Defense