Estimating the economic and social cost of crime


In 2017, more than 63,880 people were killed in homicides in Brazil. This equates to roughly 7 violent death per hour and sets an unprecedented record of violent crime in the country. In the 2017 Latinobarometer, 44% of Brazilians reported to have been victim of crime and fear of crime and public security rank high as concerns in Brazil and other countries in Latin America.


In this project we provide novel estimates on the economic and social cost of crime using cutting edge econometric techniques on large sets of administrative microdata from Brazil. We are particularly interested in documenting the cost of crime so far neglected in the literature and are absent from official cost estimates and often hidden for policy makers. Unique access to georeferenced crime data allows us to investigate the spatial dimension of exposure to crime and violence on a number of outcomes, including education, health, house prices and political outcomes.


We have access to a unique set of high quality administrative microdata on violence and crime from Brazilian government agencies. We combine the information on crime with administrative data on economic performance, educational outcomes, health realizations, outcomes of political processes among other.

We use information on the location of homicides in the public way to obtain the precise geocoordinates of incidents of crime over time. We display these on our Homicide Map, documenting the occurrence of all homicides in the public way in Brazil over time and across space at an unprecedented level of geographic precision.



Publications and Working Papers



This paper uses microdata from Brazilian vital statistics on births and deaths between 2000 and 2010 to estimate the impact of in-utero exposure to local violence – measured by homicide rates – on birth outcomes. The estimates show that exposure to violence during the first trimester of pregnancy leads to a small but precisely estimated increase in the risk of low birthweight and prematurity. Effects are found both in small municipalities, where homicides are rare, and in large municipalities, where violence is endemic, and are particularly pronounced among children of poorly educated mothers, implying that violence compounds the disadvantage that these children already suffer as a result of their households' lower socioeconomic status.

Paper: PDF

Data: ZIP

Code: ZIP


Martin Foureaux Koppensteiner & Livia Menezes

In this paper, we investigate the effect of exposure to homicides on the educational performance and human capital investments of students in Brazil. We combine extremely granular information on the location and timing of homicides with a number of very large administrative educational datasets, to estimate the effect of exposure to homicides around schools, students' residence, and on their way to school on these outcomes. We show that violence has a detrimental effect on school attendance, on standardised test scores in math and Portuguese language and increases dropout rates of students substantially. The effects are particularly pronounced for boys, indicating important heterogeneous effects of violence. We use exceptionally rich information from student- and parent-background questionnaires to investigate the effect of violence on the aspirations and attitudes towards education. In line with the effects on dropout and the longer-term human capital accumulation of students, we find that boys systematically report lower educational aspiration towards education. Making use of the very rich information from the homicides and education data, we explore a number of underlying transmission channels, including mechanisms related to school supply, bereavement and incentives for human capital investments.

Paper: PDF

Feature: IZA Newsroom

We study the effect of individual criminal victimisation in robbery and theft on birth outcomes using a unique dataset from Brazil combining information on birth records with information on criminal victimization . We find that victimisation in robbery during the first trimester reduces birthweight substantially, by about 60 grams (10 percent of a standard deviation in birthweight) and increases the likelihood for low birthweight by about 40 percent compared to the baseline. The results are robust to the inclusion of place of residence, hospital and time fixed effects and to the inclusion of a very large array of mother and pregnancy characteristics. We also show that victimisation leads to a substantial increase in fetal deaths and a positive selection of live births, hence likely providing a lower bound of the estimated effects on birthweight. Using the very rich information on victimization and health at birth we shed light on the mechanism underlying the estimated relationship.

Paper: PDF

Ongoing projects:

Crime and Productivity (Martin Foureaux Koppensteiner & Livia Menezes)

Crime and House Prices (Martin Foureaux Koppensteiner, Livia Menezes & Francisco Nobre)

Violence and Political Outcomes (Martin Foureaux Koppensteiner & Livia Menezes)

Violence and Health Outcomes (Martin Foureaux Koppensteiner & Livia Menezes)

Other related work:

The Impact of Household Shocks on Domestic Violence: Evidence from Tanzania  (Olukorede Abiona & Martin Foureaux Koppensteiner)

Project 360: An intervention to address victim-police engagement in repeat domestic violence cases (Martin Foureaux Koppensteiner, Jesse Matheson & Reka Plugor)


Martin Foureaux Koppensteiner & Livia Menezes, forthcoming at Journal of Labor Economics.



Associate professor in economics

Principal Investigator

Principal Investigator


Prof. Marco Manacorda, Queen Mary University of London & LSE

Dr Aline Menezes, UFRJ

Francisco Nobre, University of Surrey

Map development: Tomislav Bacinger


British Academy

Inter-American Development Bank

Global Challenges Research Fund 


Coming soon


School of Economics

University of Surrey



GU2 7XH 



Corresponding author:

Dr Martin Foureaux Koppensteiner


©2018 by Cost of Crime.