Who Are the Invisible Dead?
The Demographics of Incarceration and the People Behind the Uncounted Deaths
A Companion Investigation to “The Invisible Dead”
Bill Friend • February 13, 2026
Introduction
In a companion investigation, “The Invisible Dead: Why the World Cannot—or Will Not—Count Its Prison Deaths,” we documented the global failure to track mortality behind bars. That investigation established the “what”: governments across the world, from authoritarian regimes to established democracies, systematically fail to count, classify, or publish data on deaths in their custody. This investigation turns to a different and arguably more disturbing question: who are the people most likely to be among those uncounted dead?
The answer demolishes a set of assumptions that most of the public has never examined. When people imagine a prison, they imagine violent criminals—murderers, rapists, armed robbers—people whose incarceration seems self-evidently justified and whose deaths behind bars, while regrettable, carry an implicit asterisk of diminished concern. The demographic reality of incarceration tells a radically different story.
Across every First World country for which reliable data exists, the incarcerated population is dominated by three groups: people who have not been convicted of anything, people convicted of non-violent offenses, and people drawn disproportionately from marginalized communities. The genuinely dangerous—those convicted of the most serious violent offenses—constitute a minority, often a small one, of the total. This means that the uncounted dead documented in our first investigation are overwhelmingly people the system had the least justification for holding in the conditions most likely to kill them.
1. The Global Scale
More than 10.7 million people are held in penal institutions worldwide, according to the World Prison Population List. When accounting for unreported populations in countries like China and North Korea, the true figure likely exceeds 11 million. Of these, more than 3 million are held in pre-trial detention—people who have been accused but not convicted, who are legally presumed innocent, and who in many cases are held simply because they cannot afford bail or because the court system has not processed their cases.
The United States alone accounts for nearly 2 million of the global total, incarcerating its residents at a rate of 580 per 100,000—the highest of any independent democracy on earth. But the United States, while extreme in scale, is not unique in the composition of its incarcerated population. The patterns documented here recur with striking consistency across the developed world.
2. The United States: A Detailed Portrait
2.1. Custody Status: The Unconvicted Majority in Jails
The U.S. incarceration system operates on two distinct tracks that most people conflate. State and federal prisons hold people who have been convicted and sentenced, typically for felonies. Local jails—of which there are more than 3,100—hold a fundamentally different population: predominantly people who have not been convicted of any crime.
On any given day, more than 450,000 people sit in local jails awaiting trial. Most are there not because a judge has determined they pose a flight risk or a danger to the community, but because they cannot afford to post bail. The median bail amount is approximately $10,000—a sum that might seem modest but is frequently beyond the reach of the population that cycles through jails, whose pre-incarceration incomes sit far below the national median. An additional 14,000 people are held in the federal pretrial system. Taken together, the unconvicted constitute roughly a quarter of the total incarcerated population.
This is not a marginal detail. These are people for whom the presumption of innocence is a constitutional guarantee that has been functionally suspended by their inability to pay. When they die in custody—of untreated illness, of suicide, of violence in overcrowded facilities—they die as people the state had not proven guilty of anything.
2.2. Offense Breakdown: What People Are Actually Serving Time For
Among the sentenced population, the composition has shifted significantly over the past three decades in ways that are widely misunderstood. In state prisons, 63 percent of inmates in 2021 were classified as serving time for a violent offense, compared to 46 percent in 1990. This figure is frequently cited as evidence that prisons are reserved for dangerous people. It is, in fact, evidence of the opposite: the violent share grew not because more violent offenders entered the system, but because non-violent offenders left it faster. Between 2011 and 2021, the prison population serving sentences for non-violent offenses decreased by 36 percent, while the population serving sentences for violent offenses decreased by only 12 percent. The denominator shrank around a relatively stable numerator.
The federal prison system tells a different story entirely. Drug offenses account for approximately 45 percent of the federal prison population—making it the single defining characteristic of federal incarceration. Weapons offenses, immigration offenses, and fraud constitute much of the remainder. The federal system is, in essence, a drug-war apparatus with a prison attached.
Across all systems combined—state, federal, and local—drug offenses alone account for more than 360,000 incarcerated people. When combined with property crimes, public order offenses, and technical violations of probation or parole, non-violent categories account for approximately 40 percent of the total incarcerated population even after three decades of decarceration concentrated among non-violent offenders.
Even the “violent” classification demands scrutiny. The reported offense data oversimplifies how people interact with the criminal legal system. It reports only one offense category per person and reflects the outcome of the legal process rather than the actual events. Almost all convictions result from plea bargains, where defendants plead guilty to a lesser offense, sometimes in a different category entirely. The felony murder rule further inflates the numbers: a person who served as a lookout during a break-in where someone was accidentally killed can be—and routinely is—convicted of murder. The label “violent offender” encompasses serial killers and people who got into a bar fight with equal statistical weight.
3. The Pattern Across First World Peers
If the composition of the U.S. incarcerated population were an American anomaly—a product of uniquely punitive policies and the War on Drugs—the implications would be significant but contained. The data shows something more unsettling: the basic architecture of who gets locked up is remarkably consistent across the developed world. The ratios shift, but the pattern holds.
3.1. England and Wales
England and Wales held approximately 87,300 prisoners in 2024, the highest level recorded since 1900, at a rate of roughly 170 per 100,000 adults. The remand population—those held before trial or before sentencing—reached 17,700 by September 2025, representing 20 percent of the total prison population and the highest figure in at least 50 years. Within that remand population, the untried segment—people who have not yet been convicted—grew by 3 percent in a single year.
Among the sentenced population, violence against the person and drug offenses are the most prominent categories, with violent offenses constituting roughly 30 to 35 percent of sentenced prisoners. As of 2018, most women in the system were serving time for non-violent offenses. The Ministry of Justice’s offense-group data confirms that drug, theft, fraud, and public order offenses collectively account for the majority of the non-violent remainder.
3.2. Europe Broadly
The Council of Europe’s SPACE I survey—which in 2024 achieved 100 percent participation across all 51 prison administrations in 46 member states for the first time—provides the most systematic comparative data available. As of January 2024, more than one million individuals were held across European prison systems. The average European prison population rate stood at 122 per 100,000 inhabitants, with a median of 105.
On pre-trial detention: European prison administrations report, on average, that 30 percent of the inmate population consists of individuals not serving a final sentence. On offense composition: drug-related crimes are the single most frequent principal offense category in Europe at 16.7 percent, while approximately one-third of prisoners are incarcerated for violent offenses. This means that roughly half of all European prisoners are serving time for non-violent, non-drug offenses—property crimes, theft, fraud, public order violations, and other categories.
The variation within Europe is instructive. Countries like Norway, Finland, and the Netherlands maintain incarceration rates of just 54 per 100,000—less than one-tenth the American rate—while Germany sits at 71 and Switzerland at 77. These countries face comparable crime challenges but have made fundamentally different policy choices about who warrants incarceration. Their existence proves that the composition of prison populations is a policy outcome, not an inevitable response to crime.
3.3. France
France held approximately 78,000 prisoners in 2024, with a pre-trial detention rate of roughly 28 to 30 percent. The offense breakdown is telling: only 22.6 percent of prisoners were incarcerated for violent act offenses, and 18 percent for theft offenses. That leaves approximately 60 percent of the French prison population incarcerated for drug offenses, public order violations, fraud, immigration offenses, and other non-violent categories. France’s violent-offense share is notably lower than that of the U.S. or Australia, yet its pre-trial proportion is comparable, reinforcing the pattern: a large share of unconvicted people, a non-violent majority among the sentenced.
3.4. Australia
Australia presents perhaps the most striking data point on pre-trial detention among First World nations. As of 2024, 41 percent of all prisoners in Australia were on remand—up from 38 percent just one year earlier. This means that nearly half of Australia’s incarcerated population had not been convicted of any crime.
Among those who had been sentenced, the most common offenses were acts intended to cause injury (28 percent), sexual assault and related offenses (17 percent), and illicit drug offenses (11.7 percent). The combined violent category (assault plus sexual assault) constitutes roughly 45 percent of sentenced prisoners, with the remaining 55 percent distributed across drug, property, weapons, justice system, and other offense categories.
Australia also offers a stark illustration of the demographic dimension. Aboriginal and Torres Strait Islander peoples constitute approximately 3.8 percent of Australia’s general population but account for 37 percent of all prisoners. Their imprisonment rate—2,500 per 100,000 Aboriginal and Torres Strait Islander adults—exceeds the American rate by a factor of four. Any accounting of who is dying uncounted in Australian prisons must reckon with this disproportion.
3.5. Germany and the Nordic Countries
Germany, with an incarceration rate of 71 per 100,000, and the Nordic countries, with rates ranging from 54 to 74 per 100,000, demonstrate what is possible when policy prioritizes alternatives to incarceration. Their pre-trial detention rates run between 20 and 30 percent, and their offense compositions follow the European pattern: drug offenses prominent, violent offenses around one-third, and non-violent categories comprising the majority.
These countries are important not as utopias but as proof of concept. They face drug trafficking, property crime, domestic violence, and organized crime. They choose, through policy, to incarcerate fewer people for these offenses and to hold pre-trial detainees for shorter periods. The result is not higher crime rates. It is fewer people in conditions where preventable deaths occur.
4. The Numbers Side by Side
| Country | Total Incarcerated | Rate per 100K | Pre-trial % | Violent % | Drug % | Non-violent / Other % |
| United States | ~1,900,000 | 580 | ~25% | ~40%* | ~19% | ~41% |
| England & Wales | ~87,300 | 170 | 20% | ~30–35% | ~15% | ~50–55% |
| France | ~78,000 | ~105 | ~29% | ~23% | ~15% | ~62% |
| Australia | ~47,000 | 216 | 41% | ~45% | ~12% | ~43% |
| Germany | ~44,000 | 71 | ~21% | ~33% | ~17% | ~50% |
| Europe avg. | ~1,021,000 | 122 | 30% | ~33% | ~17% | ~50% |
| Nordics (avg.) | varies | 54–74 | ~25% | ~30% | ~15% | ~55% |
* U.S. violent percentage reflects all systems combined (state prisons skew higher at ~63% due to denominator effects; federal skews much lower). Offense percentages for sentenced populations only; pre-trial populations are excluded from offense calculations in most jurisdictions since they have not been convicted.
Table compiled from sources listed in the Sources section of this document, including: Prison Policy Initiative (U.S.); Bureau of Justice Statistics (U.S.); UK Ministry of Justice (England & Wales); Council of Europe SPACE I survey (Europe, Germany, Nordics); Ministère de la Justice (France); and Australian Bureau of Statistics (Australia). All figures are approximate and reflect the most recent available data as of early 2026. Offense classification systems vary by jurisdiction; cross-national comparisons should be interpreted with caution.
5. What the Numbers Reveal
5.1. The Pre-Trial Crisis
Across every country examined, between one-fifth and two-fifths of the incarcerated population has not been convicted of any crime. Globally, the figure exceeds 3 million people. These individuals are held because of court backlogs, because they could not afford bail, because the system moves slowly and nobody with power is in a hurry to move it faster. They are, by law, presumed innocent.
Pre-trial detainees are typically held in the worst conditions. Jails and remand facilities are designed for short stays but often house people for months or years. They are frequently the most overcrowded, the most understaffed, and the most medically underserved facilities in any correctional system. Suicide rates in pre-trial detention consistently exceed those among sentenced prisoners. When deaths occur in these facilities—and they do, routinely—they are deaths of people the state took custody of, failed to process, and failed to protect. Every such death is a death of someone who has not been proven guilty of anything.
The pre-trial population is also where the “definition problem” documented in our companion investigation does its most corrosive work. In the United States, jails and prisons are administered by entirely different levels of government—county versus state—with separate reporting systems, separate oversight bodies, and separate definitions of what constitutes a reportable death. A person who dies in a county jail awaiting trial on a misdemeanor charge may never appear in the state’s prison mortality statistics. A person who dies in police lockup before being transferred to jail may not appear in anyone’s statistics. The jurisdictional seams between arrest, booking, jail, and prison create gaps through which pre-trial deaths vanish from the record entirely. It is precisely the population most likely to die—the unconvicted, held in the worst conditions, with the least oversight—whose deaths are most likely to go uncounted.
5.2. The Non-Violent Majority
Among the sentenced population, non-violent offenses dominate in every country surveyed. Drug offenses, property crimes, theft, fraud, public order violations, and technical violations of probation or parole collectively account for 40 to 60 percent or more of sentenced prisoners across the developed world. The technical violations category deserves particular attention: these are not new crimes. A technical violation means a person on parole or probation missed a meeting with their officer, failed a drug test, traveled outside a permitted area, or violated a curfew. No victim, no criminal act—merely a breach of the administrative conditions attached to their release. Yet in the United States alone, roughly 280,000 people are incarcerated on any given day for technical violations, returned to the same facilities, the same conditions, and the same mortality risks as people convicted of violent felonies. When combined with the pre-trial population, the picture becomes even starker: in most systems, a clear majority of all incarcerated people are either unconvicted or convicted of offenses that most members of the public would not consider dangerous.
The retired prison warden’s framework—that violent offenders should be incarcerated, while non-violent offenders should be placed on work-release programs where they repay their debt to society—finds empirical support in these numbers. The current system does not sort people by the danger they pose. It sorts them by the resources they lack. The person who cannot afford bail sits in a cell next to the person convicted of assault. The person convicted of drug possession serves time in the same facility as the person convicted of armed robbery. The conditions that kill them do not discriminate by offense category.
5.3. The Consistency Across Nations
Perhaps the most important finding is how consistent this pattern is. The United States is an outlier in scale—it incarcerates at roughly five times the rate of most Western European countries—but it is not an outlier in composition. The basic architecture of who gets locked up is the same everywhere: a large pre-trial segment, a non-violent majority, and a genuinely dangerous minority.
This consistency matters because it eliminates the comfortable explanation that the problem is uniquely American. It is not a product of the War on Drugs, of private prisons, of mandatory minimums, or of any single policy failure. These factors amplify the problem in the United States, but the underlying structure—the systematic incarceration of people who do not need to be incarcerated—is a feature of how developed nations use imprisonment, full stop.
The Nordic countries and Germany demonstrate that this feature is not inevitable. They face the same categories of crime. They make different choices. Their incarceration rates are a fraction of the American rate, and their crime rates are comparable or lower. The people who fill the prisons of the United States, England, France, and Australia are, in large part, people who would not be incarcerated at all under the policies of Norway or the Netherlands.
6. Connecting Demographics to Mortality
In “The Invisible Dead,” we documented a global system that cannot or will not count its prison deaths. The demographic evidence presented here answers the question that investigation left open: who are the dead likely to be?
They are likely to be the pre-trial detainee held for months on a minor charge who develops an untreated infection. The drug offender in an overcrowded facility where tuberculosis spreads unchecked. The property crime defendant who cannot afford bail and hangs himself in a holding cell. The Indigenous Australian held on remand at a rate four times the American national average. The woman serving time for a non-violent offense in a facility with no mental health screening.
Deaths in custody are not disproportionately concentrated among the most dangerous offenders. They are distributed across the entire incarcerated population—and that population is composed overwhelmingly of people whose presence behind bars was, at best, discretionary and, at worst, a consequence of poverty, marginalization, and systemic neglect. Preventable causes—medical neglect, suicide, untreated illness, violence in overcrowded facilities—hit hardest where conditions are worst, and conditions are worst in precisely the facilities that house the highest concentrations of pre-trial and non-violent detainees.
There is a reason these particular people produce the most invisible deaths. The demographics of incarceration are also the demographics of invisibility. The poor cannot afford attorneys who will demand an inquest. The unconvicted have no advocacy organizations monitoring their cases, because they have no cases—they are in legal limbo, not yet part of the system’s formal record-keeping. The person convicted of a drug offense or a property crime has no constituency demanding accountability for their treatment. Indigenous Australians, Black Americans, Roma in Europe, and other marginalized communities overrepresented behind bars are the same communities whose political power to compel government transparency is weakest. The people most likely to die in custody are, by definition, the people least equipped to make their deaths matter—the people with the fewest legal resources to demand an accounting, the fewest social connections to sustain public attention, and the least political leverage to force institutional change. This is not coincidental. It is structural. The same forces that put them in custody are the forces that ensure no one counts them when they stop breathing.
This is the cascading failure of proportionality. The punishment does not fit the offense. The conditions do not fit the punishment. The oversight does not fit the conditions. And the reporting does not fit the oversight. At each level, the gap between what should happen and what does happen widens—and the people caught in that gap are disproportionately the ones the system had the least justification for holding in the first place.
Conclusion
The demographic composition of the world’s incarcerated population is not a secret. The data presented here is drawn from government statistical agencies, international monitoring bodies, and established research organizations. It is publicly available to anyone who looks. And what it shows is unambiguous: the majority of people behind bars in the developed world are either legally innocent or convicted of offenses that do not require incarceration to protect public safety.
When these people die—and they do, in numbers that no government reliably counts—their deaths are rendered invisible by the same system that rendered their humanity invisible the moment they entered custody. The failure to count the dead, documented in our companion investigation, is not a technical oversight. It is the final act in a process that begins with the decision to cage people who do not need to be caged, continues with the failure to provide them adequate care, and ends with the refusal to acknowledge what that failure costs in human lives.
The information is not unknowable. The people are not invisible because they cannot be seen. They are invisible because the systems that hold them have decided, through action and inaction, that they do not matter enough to count.
Sources
Prison Policy Initiative, “Mass Incarceration: The Whole Pie 2025” (March 2025).
Bureau of Justice Statistics, “Prisoners in 2023 – Statistical Tables” (September 2025).
Bureau of Justice Statistics, “Preliminary Data Release – Jails 2024” (December 2025).
UK Ministry of Justice, “Offender Management Statistics Quarterly” (2024–2025).
House of Commons Library, “UK Prison Population Statistics” (February 2026).
Council of Europe, “SPACE I – Prisons and Prisoners in Europe 2024: Key Findings” (July 2025).
Ministère de la Justice (France), “Distribution of Inmates by Offense” (2024).
Australian Bureau of Statistics, “Prisoners in Australia, 2024” and “Corrective Services, Australia” (2024–2025).
Institute for Crime & Justice Policy Research, “World Prison Population List” (14th ed., May 2024).
Institute for Crime & Justice Policy Research, “World Pre-trial/Remand Imprisonment List” (5th ed., November 2025).
Council on Criminal Justice, “Justice Data Snapshots” (2025).