When “Criminal” Becomes a Slur: The Record Doesn’t Tell You What You Think It Does

I have met people with criminal records who were among the most thoughtful, careful, and decent people I have encountered in this work. And I have met people who embodied every behavior I would describe as criminal, who carried no record at all, who moved through institutions with full standing and the benefit of every doubt. That observation is not a sentimental one. It is, as it turns out, empirically supported.

The word “criminal” functions in American public life the way slurs function: as a term that purports to describe a category of person, when what it actually does is assign social status. It collapses the distinction between an act and an identity. It converts a legal classification into a character judgment. And once applied, it operates as a sorting mechanism that shapes what a person can access, where they can live, whether they can vote, and how institutions calculate their worth, often for the rest of their life.

That is worth examining carefully, because the systems that deploy the label are not neutral. They were built to sort. What they sort for, and what they miss, tells us something important about what “criminal” actually measures.

What the Label Measures

A criminal record documents system contact. It documents that a person was arrested, prosecuted, and convicted. It does not document character. It does not document current risk. It does not document what a person has done since the conviction, or who they are in the rooms where no one is watching.

The sociological literature on this point is now more than sixty years deep. Howard Becker’s foundational 1963 work argued that deviance is not a quality of the act a person commits but a consequence of how institutions apply rules and then apply labels to the people they find in violation of them. Social groups, in Becker’s framing, create deviance by making the rules whose infraction constitutes deviance, and by applying those rules to particular people and labeling them as outsiders. Edwin Lemert built on this by distinguishing between primary deviance, the initial act, and secondary deviance: the identity and behavior that develop in response to how others respond to the label once applied. According to Lemert, by being stigmatized as criminals, offenders were excluded from mainstream society and left with limited alternatives, a structural outcome, not a moral one.

What this framework surfaces is something the criminal legal system tends to obscure: the label is not a neutral measurement. It is an institutional act with downstream consequences that the labeled person neither controls nor consented to. The conviction is one event. The label is a permanent assignment.

What the Label Misses

The more significant analytical problem is not what the label does to the people it touches. It is what the label systematically fails to capture.

An estimated 87 percent of fraud perpetrators have no prior criminal record. Read that again. The background check, the screening algorithm, the landlord’s denials, the employer’s pass: none of those filters would have caught most of the people causing the most documented financial harm in this country. They would have come back clean. They would have gotten the job, the apartment, the professional license.

White-collar crime is estimated to produce annual losses anywhere from hundreds of billions to well over a trillion dollars, and up to 90 percent of such crimes go unreported. Federal prosecution of these cases has been declining for years. In FY 2024, only 44 out of more than 4,300 white-collar prosecutions involved charges filed against a corporation rather than an individual. The institutions committing the most economically destructive conduct in the country are, statistically, the least likely to carry the label.

Edwin Sutherland, who coined the term “white-collar crime” in 1939, defined it as crime committed by a person of respectability and high social status in the course of their occupation. That category of harm does not tend to produce criminal records. It tends to produce settlements, regulatory fines, and deferred prosecution agreements, when it produces any legal response at all.

This is the other direction the label fails. It marks heavily in one direction and barely at all in another. What distinguishes the two is not severity of harm. It is proximity to institutional power and the resources to avoid formal processing.

The Machinery of Permanent Classification

Once the label is applied, systems take over. The National Inventory of Collateral Consequences of Conviction catalogs more than 44,000 distinct legal sanctions that attach to a criminal conviction nationally. One in three people in the United States has a criminal record, meaning they are subjected to that sprawling architecture of restrictions — covering employment, housing, voting and civic engagement, education, and overall financial security.

The screening infrastructure is comprehensive. Nine in ten employers, four in five landlords, and three in five colleges and universities now use background checks to screen out applicants with criminal records. These are not marginal actors. They represent the dominant practice across every major pathway to economic stability.

The employment data is precise and worth sitting with. Sociologist Devah Pager’s 2003 audit study sent matched pairs of job applicants, identical on qualifications, varying only in criminal history and race, to apply for actual entry-level positions. Across teams, a criminal record reduced the likelihood of a callback or job offer by nearly 50 percent. The negative effect of a criminal conviction was substantially larger for Black applicants, with the magnitude of the penalty roughly double that for white applicants with a record.

Pager’s finding that white job seekers with criminal records stood a better chance of getting a callback than Black job seekers without criminal records is one of the more precise documents of how the label interacts with race. The label, in practice, does not operate identically across populations. It amplifies existing exclusion rather than functioning as a clean factual disclosure.

The housing consequences follow the same pattern. Collateral consequences contribute to a large percentage of returning residents remaining unhoused, unemployed, and disconnected from social institutions upon reentry, increasing recidivism rates — and these consequences disproportionately harm poor communities and communities of color, leading to a cycle of system involvement and a stark racial wealth gap.

The Label as Identity Assignment

The framing I keep returning to is this: when we call someone a criminal, we are not describing what they did. We are assigning what they are. That is a different kind of claim, and it carries different moral weight.

Slurs work by collapsing the distance between an attribute and a person’s full identity, treating one feature as the totality of what someone is, and using that totality to justify exclusion. The word “criminal” does this with a legal classification as its source material. It takes a record, which is a document of what happened inside a system, and converts it into a description of what a person is, permanently, in all contexts, regardless of what has happened since.

The system itself acknowledges this, imperfectly, through mechanisms like expungement, record sealing, and ban-the-box legislation. These policy instruments exist because legislatures have recognized that the label, applied without limit, produces outcomes disproportionate to any legitimate public safety interest. Research suggests that if a person does not commit a new crime within four to seven years of release from prison, their likelihood of doing so drops to a rate comparable with people without a criminal record. The risk the label is supposed to signal dissipates. The label itself does not.

What This Costs

The costs are not evenly distributed. Formerly incarcerated people are unemployed at a rate of more than 27 percent, higher than the total U.S. unemployment rate during any historical period, including the depths of the 2008 financial crisis. That figure is not a reflection of capacity or willingness. It is a structural outcome produced by screening systems that use the label as a categorical disqualifier.

And the costs are not only individual. When systems exclude one-third of the adult population from stable employment and housing on the basis of a record that tells them nothing about who that person is today, they are making a societal bet that is expensive and, by the available evidence, frequently wrong. The harm flows in both directions: to the people carrying records who are excluded from full participation, and to institutions that screen out capable people in favor of a proxy measure that is neither accurate nor honest about what it is doing.

Honest accounting requires naming that. The word “criminal” does not describe a type of person. It describes a person the system has processed. The distinction matters for policy, for how institutions design their screening criteria, and for how we talk about people whose records follow them decades past whatever event produced them.