Does Sending Financial Crime Offenders to Prison Reduce Financial Misconduct?

This study asks whether sending people who commit financial crimes to prison reduces future offending and whether these sentences also discourage their coworkers from committing similar crimes. The authors use detailed administrative data from Finland, linking court records, workplace information, and criminal histories. Using random assignment of judges to identify causal effects, they find that a prison sentence reduces a defendant’s likelihood of reoffending by about 43 percentage points within three years. They also find spillover effects: coworkers of fraud offenders sentenced to prison are about 27 percentage points less likely to commit crimes in the next three years.

Why This Article Was Selected for The Policy Scientist

This article addresses a policy problem of broad relevance: whether incarceration meaningfully deters financial misconduct in an era of rising fraud, expanding digital vulnerabilities, and increasing concern about white-collar accountability. The question matters widely because financial crimes impose diffuse but substantial social costs, yet offenders often face comparatively lenient sanctions. The study's use of judge random assignment offers a strong causal design. Although Finland’s institutional setting differs from other jurisdictions, the mechanisms examined are general enough to inform wider debates. The article constitutes a significant contribution by advancing deterrence research and extending prior work on financial misconduct using methodologically rigorous tools.

Full Citation and Link to Article

Huttunen, K., Kaila, M., Macdonald, D. C., & Nix, E. (2025). Punishing Financial Crimes: The Impact of Prison Sentences on Defendants and Their Colleagues. American Economic Journal: Economic Policy (Forthcoming). https://doi.org/10.1257/pol.20240121 

Central Research Question

The article asks whether prison sentences causally reduce future offending among individuals who commit financial crimes, and whether these sentences also exert measurable deterrent effects on their workplace colleagues. The authors examine two related but distinct mechanisms: specific deterrence—the direct impact of imprisonment on the defendant’s own criminal behavior—and general deterrence—behavioral adjustments made by peers who observe the sanction. The paper also assesses whether financial crimes impose tangible economic costs on victims, in order to contextualize the broader policy relevance of sentencing decisions. Together, these inquiries aim to determine whether incarceration, in an environment where financial crimes are common and typically punished less severely than other non-violent crimes, meaningfully reduces the incidence of such offenses.

Previous Literature

This study engages with three major research areas. The first is the literature on the effects of incarceration on recidivism. Prior work using random assignment of judges—such as Kling (2006), Aizer and Doyle (2015), and Bhuller et al. (2020)—shows that prison can have heterogeneous effects across populations, with some papers documenting increases and others decreases in reoffending. These divergences likely reflect differences in offender characteristics, underlying labor-market attachment, and criminal histories. Financial crime offenders, who tend to have stronger labor-market ties, higher education, and significantly higher income than typical defendants, have rarely been studied in this context.

The second area concerns workplace spillovers and misconduct contagion. Prior studies document that unethical or illegal behavior can diffuse through professional networks, particularly in financial and corporate environments. Egan et al. (2019, 2021) show that adviser misconduct recurs and migrates across firms. Mohliver (2019) and Dimmock et al. (2018) provide evidence of contagion in corporate fraud and questionable accounting practices. Yet despite evidence that misconduct spreads, little is known about whether punitive actions—especially incarceration—generate downward spillovers that suppress misconduct among peers.

Third, the article contributes to research on the consequences of financial crimes for victims. While anecdotal evidence suggests substantial harm, credible causal estimates of labor-market impacts have been hindered by difficulties in linking victims to administrative data. Recent work on crime victimization, such as Bindler and Ketel (2022), has advanced these measures, but financial crime victimization specifically remains understudied. By integrating victim records with high-quality labor-market data, this study provides rare evidence of measurable costs to those harmed.

Data

The authors construct a uniquely comprehensive dataset using administrative records from Finland. The data include the full population of district-court financial crime cases from 2000 to 2018, linked to detailed labor-market information from the Finnish Longitudinal Employer–Employee Database (FLEED). The richness of the data allows the researchers to observe not only defendants, but also the establishments where they worked at the time of their offense and the colleagues employed alongside them. This structure makes it possible to track both direct and peer-level criminal outcomes.

Financial crimes are defined using Europol and FBI categories, with fraud comprising about 60 percent of cases, followed by business offenses, forgery, money laundering, and other forms of misconduct. The resulting dataset includes over 44,000 defendants and more than 100,000 colleagues in eligible workplaces (restricted to establishments with 50 or fewer employees).

The dataset also includes police records identifying victims of financial crimes, each linked to labor-market outcomes. The ability to merge victim identifiers with employment and earnings data enables a rigorous difference-in-differences design to estimate victimization impacts.

Finally, the authors collect information on judges from the National Court Registrar, enabling measurement of judge-specific sentencing tendencies. This is crucial for the instrumental-variables strategy, since Finnish law requires quasi-random assignment of criminal cases to judges.

Methods

The empirical strategy exploits random assignment of defendants to judges who vary systematically in their propensity to impose prison sentences. This variation yields an instrumental variable (IV) that isolates the effect of incarceration for marginal defendants—those whose sentencing outcomes are influenced by judicial heterogeneity rather than by underlying criminal severity.

The primary specification uses two-stage least squares. In the first stage, the probability of incarceration is regressed on judge stringency, controlling for court-year fixed effects and relevant case characteristics. In the second stage, the predicted probability of incarceration is used to estimate causal impacts on reoffending, employment, and income for defendants, as well as offending among colleagues in the three years following sentencing.

For victims, the authors use a matched difference-in-differences design. Victims are matched to nearly identical non-victims based on demographic and labor-market variables, followed by propensity-score refinement. The authors then estimate dynamic treatment effects in an event-study framework, tracking labor-market trajectories relative to the year before victimization.

Robustness checks include alternative establishment size thresholds, various case-restriction rules, placebo tests for pre-trends, and an examination of incapacitation versus deterrence channels.

Findings/Size Effects

The results indicate that incarceration has strong specific deterrent effects for financial crime defendants. Imprisonment reduces the likelihood of any new criminal charge within three years by 42.9 percentage points. This effect is large, statistically precise, and opposite in sign from naïve OLS estimates, which suggest an increase in recidivism. The timing of the treatment effects, which emerge more than two years after sentencing, rules out incapacitation as the primary mechanism. The authors also find little evidence that improved labor-market outcomes explain the reduction, making deterrence the most plausible channel.

The paper also documents significant spillover effects among coworkers. For colleagues of fraud defendants—the largest and most economically relevant category—exposure to a defendant who is sentenced to prison decreases the probability of being charged with a crime in the next three years by 27 percentage points. While standard errors are wider for this estimate, the point estimate is large and consistent with a meaningful general-deterrence effect. For other categories of financial crimes, spillover estimates are smaller and not statistically significant, suggesting heterogeneity in how misconduct is distributed or perceived within workplaces.

The authors additionally find that financial crimes impose tangible costs on victims. Victims experience a 5.3 percent decline in earnings in the year following the crime, relative to matched controls. Employment also falls modestly but significantly. These losses underscore that financial crimes—often described as “victimless”—carry meaningful economic impacts.

Descriptive evidence reinforces the importance of studying this population. Financial crime defendants exhibit relatively strong labor-market attachment yet recidivate frequently, with 45 percent reoffending within five years. They also receive substantially lighter sentences than other non-violent offenders. This combination of high incidence, high recidivism, and low punishment intensity increases the policy relevance of the findings.

Conclusion

The study provides compelling causal evidence that prison sentences reduce reoffending among financial crime defendants and produce meaningful deterrent spillovers among colleagues, at least in fraud-related cases. These findings enrich debates about appropriate sentencing for financial crimes, a category often treated more leniently than other forms of nonviolent misconduct despite its growing prevalence and substantial societal costs. The results also contribute to broader criminological and economic research by demonstrating that incarceration effects vary significantly across offender types and may be substantially more effective for white-collar defendants than for populations studied previously.

Although the Finnish institutional context, data richness, and random judicial assignment may limit straightforward replication, the underlying mechanisms—deterrence, workplace spillovers, and the structure of white-collar employment—are applicable to many advanced economies. The paper therefore offers both a rigorous empirical contribution and a foundation for future cross-national research on the criminal justice system’s role in mitigating financial misconduct.