← Back to Library

Record low crime rates are real, not just reporting bias or improved medical care

In a landscape dominated by alarmist headlines, Scott Alexander makes a startling, data-driven claim: violent crime in the United States has plummeted to its lowest point in 250 years, and this decline is real, not a statistical illusion. While public intuition screams that the world is becoming more dangerous, Alexander marshals a century of conflicting datasets to argue that we are living in an era of unprecedented safety. For the busy professional trying to cut through the noise of daily news cycles, this piece offers a rare, rigorous sanity check on the state of public order.

The Data vs. The Narrative

Alexander begins by dismantling the most convenient skepticism: that crime rates are only dropping because people have stopped calling the police. He acknowledges the validity of this concern in specific pockets, noting that "shoplifting in San Francisco" has seen reporting collapse. However, he argues this cannot explain a national trend. To prove the decline is genuine, he leans heavily on the National Crime Victimization Survey, a government study that bypasses police entirely by asking 240,000 randomly selected citizens directly if they were victimized. "The surveyors ask the victims directly," Alexander writes, and the results mirror police reports. This is a crucial distinction; if the drop were merely a failure of reporting, the survey would show a rise in unreported crimes, but it shows a parallel decline.

Record low crime rates are real, not just reporting bias or improved medical care

The argument gains further traction when examining crimes where reporting is mandatory. Alexander points out that "car theft is consistently reported to the police, because insurances require a police report before they will compensate the lost car." Yet, car theft has fallen at the same rate as other crimes. This convergence of data sources—from mandatory insurance claims to voluntary victim surveys—suggests the drop is structural, not bureaucratic. As Alexander puts it, "It would be an extraordinary coincidence if they exactly matched the proposed reporting bias to police."

It would be an extraordinary coincidence if they exactly matched the proposed reporting bias to police.

Critics might note that the data for 2020-2025 shows some friction between police reports and victim surveys, with the latter suggesting a less dramatic drop than the former. Alexander admits this discrepancy but maintains that the long-term trajectory remains unambiguous. The core of his analysis is that while short-term fluctuations exist, the century-long arc points decisively downward.

The Medical Care Counter-Argument

Perhaps the most sophisticated part of Alexander's piece is his engagement with the theory that better trauma care is masking a rise in lethal violence. The logic here is seductive: if doctors are better at saving lives, a gunshot that would have been a murder in 1960 is now an aggravated assault, artificially lowering the murder rate. Alexander traces this theory to research by Harris et al., which suggested that between 1960 and 1999, medical technology was the "principal explanation of the downward trend in lethality."

However, Alexander systematically dismantles this hypothesis using more recent research, particularly by Eckberg (2014). When looking at the National Crime Victimization Survey data, which is immune to police classification errors, the gap between assaults and murders disappears. "Their lethality trend is not compatible with the previous finding," Alexander notes, citing Eckberg's conclusion that lethality has remained stable or even increased. The reason? The nature of the violence has changed. Modern shooters are inflicting more severe injuries to compensate for better medical care. Alexander cites studies showing that the proportion of patients with three or more gunshot wounds nearly doubled between 2000 and 2011.

This creates a grim but clarifying picture: medical advances are indeed saving lives, but they are being met with more aggressive weaponry and tactics. "Gun injuries are getting worse over time," Alexander writes, pointing to the shift toward semi-automatic weapons and a criminal culture that now expects to fire multiple shots to ensure a kill. The cancellation of these two forces—better medicine and worse violence—results in a stable lethality rate, meaning the drop in murder statistics is not a medical artifact. It is a genuine reduction in the number of violent acts.

Gun injuries are getting worse over time, yet the murder rate is at historic lows. This suggests fewer people are being shot at all.

Why We Feel Unsafe

If the data is so clear, why does the public feel so threatened? Alexander suggests that our cultural intuition is simply out of sync with reality. He draws a parallel to other areas of safety, observing that "we're a safetyist culture" where car fatalities and childhood injuries have also hit historic lows due to regulation and technology. "Why should crime be the exception?" he asks. The disconnect, he implies, stems from a media ecosystem that amplifies rare, horrific events while ignoring the quiet, statistical reality of safety.

He offers a buffet of potential explanations for the decline, ranging from the reduction of lead poisoning to the impact of mass incarceration, though he admits no single theory fully explains the drop. One of his more provocative suggestions involves the anti-police backlash following Black Lives Matter, which he argues may have triggered a "backlash-to-the-backlash" that ultimately bolstered community support and resources for police in 2023, 2024, and 2025. This is a contentious point; critics might argue that attributing a crime drop to a political backlash oversimplifies complex social dynamics and ignores the role of economic factors or demographic shifts. Yet, Alexander's willingness to entertain unconventional hypotheses keeps the analysis fresh and avoids the trap of repeating standard criminological talking points.

Bottom Line

Scott Alexander's strongest contribution is his rigorous defense of the data against the twin skeptics of reporting bias and medical artifacts, proving that the drop in crime is real and not a statistical mirage. His argument's greatest vulnerability lies in the speculative nature of why this is happening, where he offers a range of plausible but unproven theories rather than a definitive cause. Readers should watch for how this data holds up against the next year's statistics, as the gap between public perception and statistical reality remains the most dangerous variable in the equation.

Sources

Record low crime rates are real, not just reporting bias or improved medical care

by Scott Alexander · Astral Codex Ten · Read full article

Last year, the US may have recorded the lowest murder rate in its 250 year history. Other crimes have poorer historical data, but are at least at ~50 year lows.

This post will do two things:

Establish that our best data show crime rates are historically low

Argue that this is a real effect, not just reporting bias (people report fewer crimes to police) or an artifact of better medical care (victims are more likely to survive, so murders get downgraded to assaults)

Here’s US murder rate, 1776 - present:

The pre-1900 estimates come from Tcherni-Buzzeo (2018); their ultimate source seems to be work by sociologist Claude Fisher which I can’t access. The 1900 - present data come from historian Randolph Roth’s American Homicide and the FBI’s Uniform Crime Reporting, both by way of the Council on Criminal Justice.

There’s less historical data for property crimes, and the nature of property has changed throughout history in ways that make numbers incommensurable (is it bad if we have a higher grand theft auto rate today than in 1840?) I was only able to get good data since 1960, but here it is:

The 1960-2023 data come from FBI Data Explorer via Vital City; the 2024 and 2025 data come directly from the FBI website, with 2025 annualized via incomplete Jan - Oct data. This one may or may not be an all-time low, but it’s pretty good.

These data are counterintuitive. Are they wrong?

Could This Be An Artifact Of Reporting Bias?.

People could be so inured to crime that they stop reporting it to the police. Or the police could be so overwhelmed that they stop accepting the reports. Since most crime statistics are based on police reports, this would look like crime going down. There’s some evidence of this happening in specific situations, like shoplifting in San Francisco. Could it be the whole effect?

No, for three reasons.

The National Crime Victimization Survey is a government-run survey of a 240,000 person nationally representative sample. They find random people and ask whether they were the victims of crimes in the past year. This obviously doesn’t work for murder, but they keep statistics on rape, assault, larceny, and burglary. Their numbers mostly mirror those reported by police and used in the usual statistics about crime rates. But here there’s no extra step of needing to trust the police enough to make ...