Crime and Punishment I

Carolina Torreblanca

University of Pennsylvania

Global Development: Intermediate Topics in Politics, Policy, and Data

PSCI 3200 - Spring 2026

Agenda

  • Bateson, 2012

  • Weaver and Lerman, 2010

Why does the state exist?

The provision of security as the central reason for the state

  • Hobbes: Prevent chaos \(\rightarrow\) Protect our lives
  • Locke: Prevent abuse \(\rightarrow\) Protect our property (including our lives)

What do states do?

States provide public goods/services:

  • Health, schools, infrastructure
  • The provision of security (internal and external)

Security and Development

  • Security often thought as a prerequisite of development
    • Why?
  • Conversely, internal violence (civil war) characterized as “development in reverse” (Collier, 2004)
    • Why?

The inevitability of internal conflict

  • Yet, states often fail to provide internal security

  • Particularly developing states

  • Political violence, civil war, and most commonly… crime

Crime in the Developing World

Crime often identified as main concern of developing-world citizens

Crime in the Developing World


Consequences for Development?

  • In Latam, the most violent region in the world, the direct costs of crime in 2022 reached 3.44% of the region’s GDP (Source: IDB)

    • “The cost of crime is equivalent to 78% of the public budget for education, double the public budget for social assistance, and 12 times the budget for research and development.”

  • At least 115k people were murdered in Latin America and the Caribbean during 2023

Police

  • The state’s main tool to respond to insecurity is the police
  • Yet police in developing context are often repressive, abusive, over-target poor and marginalized communities.
  • Could policing backfire?

Crime and its Punishment

  • What are the consequences of criminal violence?
  • What are the consequences of poor policing?

Bateson, 2012

Research Question: Can crime motivate people to participate more in politics?

  • If the answer is “no” then crime could trigger a vicious circle of disengagement \(\rightarrow\) poor governance \(\rightarrow\) worse developmental outcomes


  • If the answer is “yes” then violence today might (ironically) mean better outcomes tomorrow: crime \(\rightarrow\) participation \(\rightarrow\) Better development outcomes

Theoretical Expectations


Why might we expect crime to depress political participation?

  • Fear
  • Cost of participating

Why might we expect crime to spur political participation?

  • Psychological reasons
  • Instrumental reasons

A quick aside


What were your priors?

Context / Data


Author leverages four regional surveys

  1. Americas Barometer
  2. Eurobarometer
  3. Asian Barometer
  4. Afrobarometer

You can download all of these surveys yourself!

Measurement

What is the dependent variable?

  • Non-electoral participation
    • community action
    • protest
    • political interest
    • town meetings

What is the treatment or independent variable?

  • Direct or indirect criminal victimization in the past year

Research Design


CIA or Selection on Observables

Treatment Group:

  • Direct or indirect victims

Control Group:

  • Non-victims

Controls:

  • Age, Education, Gender, Urban, Country FE

Threats to Inference

The author makes causal claims, what crucial assumption is she making?

  • No confounders!
    • After including controls, who gets victimized is independent of participation decisions
  • Assumption would not hold if:
    • People who participate more are more likely to be crimed
    • People who participate more are more likely to report they were crimed

Is Causality Plausible?

Results

Updating on the Results

Crime victims everywhere participate more in politics, regadless of the type of crime

  • Effect is roughly the same as 5-10 years of education (!!!)
  • Author also finds that victims are more skeptical of democracy, support authoritarianism, and vigilantism

Q1: Do we believe the results? Q2: What implications for democracy and development?

Weaver and Lerman, 2010

Context


For US citizens, contact with the criminal justice system is more common now than ever before

  • 1/100 US citizens are incarcerate
  • 1/3 black men will sever time in prison
  • Disadvantaged population routinely have unwanted/disciplinary interactions with police.

Authors call the spacial concentrated, punitive, surveillance-oriented system found in some (minority) communities “carceral state”

Research Question

How and in what ways encounters with the criminal justice system influence citizens’ political attitudes?

  • More specifically:

    • How does exposure to the criminal justice system socialize those exposed to it?
    • What do people learn from constant, unwanted, aggressive, interactions with the state’s criminal system?

Theory

Interactions with the criminal system shape participation through two channels:

  • Resources:
    • Fewer resources \(\rightarrow\) less political participation (i.e. spend $ in court, have a criminal record, etc.)
  • Learning:
    • What the authors call “interpretative effects”
    • Encounters with bureaucrats teach citizens’ about the goals and nature of the government
    • Bad experiences with police, courts, etc. may translate into bad impressions about the government

Hypotheses

Interactions with the criminal justice system will:

  1. Depress participation and civic engagement
  2. Reduce trust in the government

Data

Data comes from two panel surveys

  1. National Longitudinal Study of Adolescent Health (Add Health)

    • “20,000 adolescents who were in grades 7-12 during the 1994-95 school year, and have been followed for five waves to date”

  2. Fragile Families and Child Well-being Study

Both surveys are publicly available and VERY RICH (and linked in the slides)

Measurement

Treatment:

  • Criminal Justice Contact: no encounters, stopped by the police, charged, arrested, served time

Outcomes:

  • Electoral participation
  • Civic participation
  • Political participation
  • Trust in government
  • Civic obligations (jury, military)

Research Design

CIA or Selection on Observables

  • Multivariate OLS regression

  • Threats to inference: confounders!

  • What might predict participation and contact with criminal system?

  • Personality, criminality, income, education, etc.

  • Authors try to control for all these and run placebos to test whether they might be driving the results

Results

Implications

  • Criminal justice contact deters political and civic life
  • Especially for marginalized populations who are over-policed
  • Who should security policy be responsive to?
  • Implications for democracy?
  • For inequality?

Upside down world?

  • Both crime and its punishment have important implications for politics
  • Crime engenders participation and policing deters it (???)
  • A cause and a consequence of development
  • What is the appropriate response to crime? A very hard question to answer

For researchers

  • Both of these papers were published in the highest ranked journal of political science
  • Both use survey data which you yourself could have downloaded
  • You could have run all the analyses!
  • Why are they published where they are?
  • Why are they so influential?