Inspired by watching the TV series Blacklist, things I've been reading, and my own brief experience with courts and imprisonment, I asked Claude whether it was a known phenomenon that the people who are identified by police, caught, and convicted of crimes are the ones who are "dumb" enough to be identified and caught, who aren't cunning and careful enough.
There isn't one single canonical named hypothesis for this exact framing, but it sits at the intersection of several well-developed criminological concepts, and the observation itself is quite respectable intellectually.
The closest formal framing is probably differential attrition or selection bias through the criminal justice funnel - the well-documented phenomenon that offenders are systematically filtered at every stage (offence → detection → arrest → prosecution → conviction → sentencing), and the filters are not random. They correlate with factors like impulsivity, social capital, organisational sophistication, and yes, intelligence.
Edwin Sutherland's white-collar crime thesis (1949) made essentially your argument in reverse: he pointed out that the most harmful criminals - those with social standing, organisational access, and the cunning to operate through legitimate institutions - are systematically absent from prison populations, which led criminologists to dramatically underestimate both the prevalence and harm of crime by studying only incarcerated samples. This is sometimes called the "respectable criminal" problem.
Gottfredson and Hirschi's self-control theory (1990) is also relevant - their argument that low self-control predicts both criminal behaviour and getting caught makes a theoretical prediction consistent with yours: the prison population is disproportionately low-self-control, which is a proxy for the kind of impulsivity and carelessness that leads to identification and betrayal.
The methodological version of your observation is sometimes called the "caught criminal fallacy" or more formally the "incarcerated sample problem" - the recognition that using prison populations to study criminal behaviour introduces severe selection bias, because you're studying precisely those who failed to evade the system.
Your specific framing - that the truly dangerous and sophisticated aren't in prison - doesn't have a single pithy name, but it's implicit in the literature on organised crime, white-collar crime, and elite deviance (a term associated with David Simon), all of which take as a founding premise that the most structurally damaging criminal actors are those least likely to be incarcerated.
So: no single named hypothesis, but a well-recognised problem with several handles depending on which angle you're approaching from.
I followed up with a question about the function of the criminal justice system in society, and if there was a known mismatch between society's perception and belief of the role and effectiveness of prisons and the real world performance of prisons. Is it mostly performative to give the impression of law and order when in reality the most heinous criminals are, like in Blacklist, evade detection and capture?
This is one of the most thoroughly critiqued misalignments in all of social science, and it has generated a substantial and quite damning literature. The short answer is: yes, the misalignment is real, well-documented, and arguably structural rather than incidental.
The performative function of punishment - Durkheim and after
Émile Durkheim made an early and important observation that punishment is primarily a social ritual - it reaffirms collective norms and produces solidarity among the law-abiding, regardless of whether it actually reduces crime. On this view, the trial, conviction, and imprisonment of a criminal is a moral theatre performed for society, not an instrumental act designed to make society safer. This separates the expressive function of punishment from its instrumental one, and the gap between those two has only widened in subsequent analysis.
Foucault's deliberately provocative inversion
Michel Foucault in Discipline and Punish made the argument that is probably closest to the spirit of your question. He noted that by any crime-reduction measure, prisons appear to fail - recidivism is high, serious crime continues, criminal networks are consolidated inside prisons rather than dismantled. His response was to ask: what if this "failure" is actually the system working as intended? His argument was that the prison system produces a known, managed, identifiable criminal class - the recidivist, the petty criminal, the street offender - and in doing so legitimates intensive state surveillance and control of the poor and marginalised, while leaving entirely untouched the structurally powerful. The system doesn't fail to catch criminals; it succeeds at catching the right criminals - meaning those whose incarceration serves the interests of social order without threatening entrenched power.
Garland's expressive turn
David Garland's The Culture of Control (2001) is probably the most sophisticated modern treatment of this. He argues that from roughly the 1970s onward, criminal justice policy in the US and UK underwent a fundamental shift - away from the rehabilitative ideal (which at least had an instrumental logic: change behaviour, reduce crime) toward what he calls expressive or punitive justice. Policies began to be designed not to reduce crime but to express moral condemnation, respond to public anxiety, and signal toughness to voters. Garland's term penal populism (developed further by John Pratt) describes how criminal justice becomes responsive to public appetite for punishment rather than to evidence about what works. The result is a system oriented around performance and legitimacy rather than outcomes.
Jeffrey Reiman's explicit structural argument
Jeffrey Reiman's The Rich Get Richer and the Poor Get Prison is perhaps the most direct articulation of your core intuition. His central thesis is that the American criminal justice system is not a failed system - it is a successful one, but its success is measured by its function of criminalising poverty and managing the lower classes, not by reducing harm. He makes the observation that corporate crime, white-collar fraud, industrial negligence, and state violence cause vastly more harm, death, and suffering than street crime, yet receive a tiny fraction of the enforcement attention. This is not accidental incompetence; it is structural. The system is calibrated to the street, not the suite.
The dark figure and the funnel's true shape
Criminologists use the term the dark figure of crime to refer to offences that never enter the funnel at all - never reported, never recorded. Estimates consistently suggest that for many serious crime categories, the dark figure dwarfs recorded crime. When you combine this with the selection bias in who gets arrested, prosecuted, and convicted, the prison population becomes an extremely distorted sample of actual criminal behaviour - skewed heavily toward the poor, the impulsive, the socially isolated, and minorities subject to disproportionate policing. The funnel doesn't just filter; it actively distorts the picture of who "criminals" are.
The legitimation problem
The deepest version of your question is about legitimation - the function the system plays in maintaining public belief in the social order. There's a coherent argument that the criminal justice system's primary social function is to sustain the narrative that wrongdoing is identified and punished, that society is protected, and that the state has the situation under control. This narrative is necessary for social cohesion regardless of whether it's true. The visible prosecution of street crime, the prison population as a kind of displayed trophy of justice, the ritual of trial and sentencing - all of this performs order for public consumption. The fact that the most destructive actors - financial fraudsters, corrupt officials, sophisticated organised crime leadership, corporate executives making lethal decisions - rarely enter the funnel is, from this perspective, a feature of the legitimation function rather than a failure of it. Their prosecution would be destabilising in ways that prosecuting a drug dealer is not.
Where this leaves rehabilitation
Rehabilitation programs are caught in a particularly uncomfortable position within this framework. If the people in prison are disproportionately there because of impulsivity, poor social capital, mental illness, addiction, and poverty - rather than because they are the most harmful - then rehabilitation programs are, at best, addressing the symptoms of structural inequality rather than crime itself. At worst, they provide a humane veneer on a system whose actual function has little to do with reducing harm.
The honest summary of the literature is this: there is broad scholarly consensus that the criminal justice system as it currently operates in most Western countries has a weak relationship with crime reduction, a strong relationship with the management of the marginalised, and serves a powerful symbolic/legitimating function that is largely independent of its instrumental performance. Whether this is the result of institutional drift, class interest, political incentive, or some combination is where the debate sits - but the misalignment you're identifying is not a fringe critique. It's close to the mainstream of criminological and sociological scholarship.