Does Fear of Deportation Affect Victims’ Willingness to Report Crime?

Gonçalves, Jácome, and Weisburst (2026) examined whether immigration enforcement policies reduce public safety by discouraging crime victims from reporting crimes to police. They studied the Secure Communities program, which increased cooperation between local law enforcement and federal immigration authorities. The authors used data from the National Crime Victimization Survey, Immigration and Customs Enforcement records, FBI crime data, Census data, and police department records. They found that Hispanic victims became 30% less likely to report crimes to police after the policy was implemented. At the same time, victimization against Hispanics increased by 16%, suggesting that reduced reporting lowered offenders’ fear of being caught.

Why This Article Was Selected for The Policy Scientist

This article addresses a central question in public policy: whether aggressive enforcement strategies can unintentionally weaken the social cooperation required for effective governance. That issue extends well beyond immigration enforcement and applies broadly to policing, public health, education, and regulatory compliance. The article is especially timely because debates over immigration enforcement, police legitimacy, and institutional trust remain politically and administratively salient across the United States. Gonçalves, Jácome, and Weisburst have written extensively on immigration enforcement and community behavior, and this study advances that literature by directly linking reduced victim reporting to increased victimization. The data are unusually strong, combining NCVS survey records, ICE administrative data, FBI crime data, Census data, and local police records. The quasi-experimental difference-in-differences design provides substantially stronger causal leverage than standard multivariate regression approaches, and the findings are likely generalizable.

Full Citation and Link to Article

Gonçalves, F., Jácome, E., & Weisburst, E. (forthcoming). Community engagement and public safety: Evidence from crime enforcement targeting immigrants. American Economic Review.

Central Research Question

Gonçalves, Jácome, and Weisburst investigate whether immigration enforcement policies can unintentionally reduce public safety by discouraging victims from reporting crimes to police. Specifically, the article examines the Secure Communities (SC) program, a federal initiative launched in 2008 that expanded cooperation between local law enforcement and federal immigration authorities. The authors ask three connected questions. First, did Secure Communities reduce the willingness of Hispanic victims to report crimes? Second, did criminal victimization increase after the policy was implemented? Third, can changes in reporting behavior explain changes in offending? These questions are important because traditional crime statistics generally rely on reported crime data, making it difficult to separate actual criminal activity from changes in victims’ willingness to contact police. The article therefore addresses both a substantive policy issue and a major measurement problem in the crime literature.

Previous Literature

The article builds on several strands of literature in economics, criminology, and public policy. A large body of work following Becker’s economic model of crime has emphasized deterrence and the probability of punishment as central drivers of criminal behavior. Much of that literature focuses on police presence, incarceration, sentencing severity, or arrest rates. By contrast, relatively little research has examined the role of victims themselves in producing public safety outcomes. Existing studies have often assumed that victims report crimes consistently over time, even though scholars in criminology have long argued that trust in police shapes reporting behavior.

The article also contributes to research on immigration enforcement. Earlier studies examining Secure Communities generally concluded that the policy had little measurable effect on crime rates. However, those studies relied primarily on FBI reported crime statistics, which cannot distinguish between actual changes in victimization and changes in reporting behavior. Gonçalves, Jácome, and Weisburst argue that this limitation concealed important effects of the policy. Their findings therefore challenge the prevailing interpretation of Secure Communities in prior scholarship.

The study additionally connects to a growing literature examining how institutional trust affects cooperation with government authorities. Prior work has shown that policing practices, immigration enforcement, and high-profile incidents involving law enforcement can alter public willingness to cooperate with police. The authors extend this literature by providing direct evidence that lower reporting rates can produce measurable increases in criminal victimization. In doing so, they move beyond correlational claims and provide one of the strongest causal analyses to date linking community engagement with public safety outcomes.

Data

The study combines multiple large-scale administrative and survey datasets. Its primary source is the National Crime Victimization Survey (NCVS), a nationally representative survey administered by the U.S. Census Bureau and the Bureau of Justice Statistics. The NCVS is uniquely valuable because it asks respondents both whether they experienced a crime and whether they reported the incident to police. This distinction allows the authors to separate victimization from reporting behavior, something impossible with ordinary police crime statistics.

The restricted-access version of the NCVS also includes county identifiers, allowing the authors to match respondents to the staggered implementation of Secure Communities across counties. The authors supplement these data with Immigration and Customs Enforcement (ICE) records on detainers and removals, FBI Uniform Crime Reporting data, Census and American Community Survey demographic information, labor market indicators, and administrative records from police departments. They additionally assembled original datasets from police agencies, including 911 call records and arrest records from dozens of cities.

The data quality is exceptionally strong. The NCVS is widely regarded as the most reliable national source for measuring victimization because it captures both reported and unreported crimes. The combination of survey data, administrative enforcement records, and local policing data provides substantial triangulation across independent sources. The geographic breadth of the data also improves the external validity of the findings, particularly for large urban and suburban counties with sizable Hispanic populations.

Methods

The article employs a quasi-experimental difference-in-differences research design. Secure Communities was introduced gradually across counties because of technological and administrative constraints rather than local crime conditions. This staggered rollout created plausibly exogenous variation in exposure to immigration enforcement. The authors compare changes in outcomes before and after implementation across counties activated at different times.

The study uses modern difference-in-differences estimators designed to address problems associated with traditional two-way fixed effects models. The authors estimate dynamic event-study models to test for preexisting trends and demonstrate that treated and comparison counties followed similar trajectories prior to implementation. This strengthens the causal interpretation of the results.

The methodological approach is considerably stronger than standard multivariate regression because it leverages temporal and geographic variation in policy implementation. The authors also conduct extensive robustness checks, including alternative estimators, alternative control groups, triple-difference models, and analyses accounting for survey attrition and demographic composition changes. These exercises consistently reproduce the main findings.

The article further strengthens causal inference through several complementary analyses. The authors compare cohorts with differing reporting declines, examine police clearance rates and response times, analyze offender ethnicity patterns using arrest records, and conduct mediation analyses assessing whether economic distress or deportations can explain the increase in victimization. Together, these analyses substantially narrow the range of alternative explanations for the observed results.

Findings/Size Effects

The authors find that Secure Communities significantly reduced crime reporting among Hispanic victims. Reporting rates declined by approximately 9 percentage points, representing roughly a 30% reduction relative to pre-policy levels. The decline occurred primarily for property crimes, although the direction of the effect was similar for violent offenses. Non-Hispanic reporting rates remained largely unchanged.

At the same time, criminal victimization against Hispanics increased substantially. Hispanic victimization rose by approximately 0.15 percentage points per month, representing a 16% increase relative to baseline victimization levels. The authors estimate that the policy generated roughly 1.3 million additional crimes against Hispanics nationally during the first two years after implementation. Property crimes account for most of this increase, though violent victimization also rose.

Importantly, traditional reported crime measures showed almost no overall change. The increase in victimization was offset by the decline in reporting, causing standard crime statistics to mask the policy’s effects. This finding directly explains why previous studies relying on reported crime data concluded that Secure Communities had little impact on public safety.

The authors provide substantial evidence that reduced reporting was the principal mechanism driving increased offending. Counties with larger declines in reporting experienced larger increases in victimization. Police clearance rates and response times remained largely unchanged, suggesting that police effectiveness did not deteriorate. Arrest data also indicate that offenders became less Hispanic over time, weakening explanations based on worsening labor market conditions among Hispanic populations.

One of the article’s most important contributions is its estimate of the elasticity between reporting and offending. The authors estimate that a 10% decline in victim reporting produced approximately a 7.9% increase in offending. This is among the first causal estimates directly linking victim cooperation with crime levels.

Conclusion

The article makes a significant contribution to the economics of crime, immigration policy research, and the broader study of institutional trust. Its central finding is that enforcement policies may produce unintended public safety consequences when they reduce civilian cooperation with law enforcement. The study demonstrates that victim reporting is not merely a passive measurement issue but an active component of deterrence and crime control.

Methodologically, the article is notable for combining high-quality victimization data with rigorous quasi-experimental methods and multiple supplementary analyses. The findings are especially important because they reveal how standard administrative crime statistics can conceal meaningful policy effects. More broadly, the study suggests that public safety depends not only on formal enforcement capacity but also on the willingness of communities to engage with public institutions.