Disorder is Declining Along with Violence

All crime is not local

A note: I read with interest the Guardian’s review of Mister Beast, the world’s largest producer of content. They report that Mister Beast maximizes engagement by overwhelming the viewer with hype throughout the first minute of engagement so they hang-in until that magical second minute when monetization really begins. In the spirit of finding that to be horrifying, I remain steadfast in the Crock Pot business model where together we stew and simmer for a while. I will get to the point here, which is that disorder is also falling quite swifty along with crime, but it won’t be in the first minute.

Another note: I have published this newsletter haphazardly of late, occasionally at best, and thought perhaps an explanation was in order. This is usually where your favorite Substacker announces their pending book, but alas, not yet. It is a difficult time for social science research in America and my corner is not immune. To carry on in tumultuous seas, one trims the main and sails close-hauled and pinched. If you make a mistake in one direction, you speed up a little, which is easily corrected. If you make a mistake in the other, you luff, which is to say, you stop forward progress. That’s ok too. It’s capsizing that is really bad.

So, the question comes naturally, what to comment on and what to avoid? And it follows, what is this newsletter about, and what is it not?

External Processing is about thoughts on governing, of course, and there are three essential mechanisms of governing that one could plausibly comment on: policy, politics, and power.

Long ago I vigorously studied political philosophy, including sitting for a comprehensive exam in moral philosophy. But those ideas are mainly ghosts to me now, and children and work prevent me from returning to them because philosophy is a greedy partner and when studying it you are not much fun to be around (apologies to gregarious philosophers everywhere). Thus, I have little to say about power that is new and engaging.

I worked in politics for a time, but that too is ancient history. What I know about the mechanics of messaging and strategic bargaining is anachronistic. Also, politics is a trade, and trade discussions quickly become more about mechanics than science.

But I am quite interested in the policy-politics overlap. I was delighted to see Abundance enter that space with thoughtful ideas. Whether you find it compelling or not, perhaps you agree that the main discourse in these areas has been on the power-politics overlap, not the politics-policy overlap. Any attempt to thoughtfully fill the politics-policy void is warmly received, at least by me.

I am mainly concerned here with policy. And specifically, policy around crime and justice.

How society interacts with those most likely to victimize or be victimized is perhaps the most reliable indicator of the sustainability of the social contract.

So when time and space allow, I will offer some ideas to chew on in the spirit of progress. I was heartened by a recent article that argued that policy ideas tend to bubble from ideation to political action over a decade or a generation—not days or months—so all is not lost if there is little agency for them at gestation. So, my friends, onward!

Crime Cycles

One helpful mechanism for separating truth from nonsense is to ask yourself, when the history of this event is written, will the version I’m being spoon fed right now hold up? Occum’s Razor is widely misunderstood as counseling the selection of the simplest story over the nuanced one. What it says instead is that given two equally evidenced explanations, you should select the simpler one. Only when many explanations are equally evidenced should you prefer simplicity—it does not counsel you to run about chasing lightning bugs thinking you have a discovered a means of sustainable energy.

I’m going to break with my five-year-old tradition in this newsletter of not being self-referencing or self-congratulatory and embrace both, for just a moment. Three years ago, I was genuinely curious why the gun violence spike happened and why it had reversed. I know genuine curiosity is frowned on in a content-driven, Mister Beast dominated-world but there it is.

What I ended up being convinced by was data showing an inverse correlation between local government employment and violent crime. As local government employment rapidly declined in the spring of 2020, violence exploded. As those jobs came back, violence receded. When local government employment began to reach and exceed new highs, violence hit 50-year lows. And it’s not just correlational data—there is a compelling causal mechanism here, the direct interaction between local government employees (teachers, counselors, mentors) and young men at risk of violence and victimization.

But here’s the key point—what happened over the last three years is not just that there were big federal investments in local anti-violence programs, there was an order of magnitude larger investment in broad prevention, delivered by ordinary people in classrooms, community centers, churches, and libraries. The criminological literature is clear that the more people who care about a young man, the less likely they are to engage in violence. In the last three years, we funded more people than ever, who could spread more care to more young men. Occum’s Razor.

If Crime Cycles, it Follows That There Are National Forces at Work

As the sub-header title bluntly states, the violence spike and decline appear to have been mainly national not local. That idea requires a serious explanation because it turns it that is quite a novel way to look at crime. It’s not ‘novel’ in the sense that no one ever thought of this before, as obviously lots of people have thought of this before. It’s novel in the sense that the whole criminal Justice edifice pushes us not to think this way.

We have been conditioned by local media (and now social media), that crime is episodic (not routine), that violence is rare and thus newsworthy (sadly this has never been true), and that the local police are our one and only defense against crime and what they do, or don’t do, makes all the difference.

This last point is really critical. Ted Kennedy said all politics is local, and we’ve been conditioned that this is true of crime as well. It is not.

A much more productive way to think about crime cycles is to think of national crime as the surface of the sea. The current and the swells can roll in from faraway storms. Huge domes of high pressure can smooth everything for thousands of square miles. While local storms can create chop and whitecaps, that’s on top of whatever currents and swells are moving about. If you’ve ever been out on a boat on a calm day when a big storm rolls in—you’ve seen how fast the sea can roil. But you’ve also experienced a rapid calm. Local effects are everywhere soon overtaken by broader forces.

So, crime moves up and down mainly with tide and current, with local events pushing and pulling on those effects. Your local police task force or new technology can push crime below the national trend, or apathy and an insistence on outdated ways can push your community above it. Your local police have a lot of control over where you sit relative to the trend but not the trend itself.

So why are there so many voices trying to convince you otherwise? Because there are many voices on the local side and few on the national side. Quantity has a quality all its own.

There are 19,000 local law enforcement agencies in the US and 3,100 counties with county courts, and thousands of jails and state correctional institutes. Each has a strong reason to convince you that their work is critical to solving the crime problem. Their budgets and thus their scope and scale itself require a broad public belief that they are consequential.

By contrast, national policing is extremely limited in the US, at least compared to peer nations. The FBI and the federal prison system are tiny when compared to their state, county and local counterparts. Thus, the narrative is skewed towards local effects and away from national effects.

One other reason why there may be limited interest in the idea that national forces drive crime, is that few macroeconomic indicators are correlated with national crime. GDP and unemployment, for example, seem at best to be lightly related to crime levels. Inflation has some correlation to property crime, but little correlation to violence. This lack of correlation with the most important macroeconomic indicators may have led to a broad tendency to believe that macro forces overall have negligible effect on crime. This is a bit of a selection effect. Just because the amount of currency in circulation or the number of people employed doesn’t explain changes in crime doesn’t mean other important national measures—like how many teachers are cumulatively employed by local government—have no effect.

This same selection effect causes us to believe that there are few national levers to pull if crime soars. The logic is that while we can raise or lower interest rates to address inflation and unemployment, since those forces are not correlated with crime, there must not be any federal mechanisms that can be adjusted to help reduce crime. So mostly, when crime goes up and down there are a lot of fingers pointing but not much action.

But there are a lot of other things that can be done. National policies like child poverty tax credits can have a substantial effect on crime. Easing stress on poor families turns out to be very helpful at reducing the pull of crime.

So that’s the big point I want to make here. I’ve written about the need for a marocriminology discipline to think specifically about these challenges and opportunities.

But getting there requires a mind shift that is tough to pull off given the volume of voices telling you not to look in that direction.

I wrote this essay on the Notes app on my phone, with my thumbs, while shopping with my kid. I am ambivalent about the experience. For non-digital natives like me, thumb-typing on your phone is the new cursive—it is so slow and awkward that you have to think carefully about each word—is it really worth the effort to mistype this sentence and fix the autocorrect, or can I gently place it on a raft and set it out to sea? Anyway, and with any luck, I won’t have to do that again soon. The next essay was written more conventionally: hunched over, elbows-in, four finger typing, laptop in lap, on a way-too-small commuter jet on a way-too long flight, while being commoditized by the McAirline industry.

One the Decline in Disorder

One day, in a moment of sweet reflection, you may find yourself asking: what is the relationship between crime and disorder? Crime and disorder are not the same thing, though they are often confused for the other. I wrote about this last year and offered some thoughts about how crime and disorder are different, and how confusing the two often leads to ineffective or even iatrogenic policy. But it is sufficient here to say that it is a classic endogeneity problem: does crime cause disorder, or does disorder cause crime?

This is a problem. When crime and violence go up, there is a reflex to fight disorder: implement curfews, crackdown on encampments, arrest the fare evaders. When there is an increase in disorder, there is a reflex to punish criminals more severely: mandatory minimums, three strikes, truth in sentencing.

But here’s the thing: we don’t actually know how, or how much, crime and disorder are related.

It’s important here to say what I think disorder is. I define disorder as rule-breaking that does not rise to the level of criminal conduct—social disorder like loitering boisterously, playing loud music, driving without care, and physical disorder, such as the presence of trash and graffiti. Professor Ralph Taylor at Temple University labels disorder as ‘incivilities’ which I find to be the right framing, and argues, “[i]ncivility indicators are social and physical conditions in a neighborhood that are viewed as troublesome and potentially threatening by its residents and users of its public spaces” (the Taylor article begins on page 65).

Now, there has been a lot written recently about the fall of crime, and increasingly some embrace of my theory that local public sector losses and then gains in employment are the key driver. But what is happening with disorder? While there was a flurry of articles when disorder appeared to be spiking, I have not seen anything written about whether disorder is declining, or if it is, why? (Spoiler: disorder is declining).

I don’t think the disorder concern is, or was, a vibes thing, I think it was much more tangible than that. Some of the disorder spike was media-created. There is little evidence that smash and grabs were anything more than great content for social media, and, it is not clear whether shoplifting really spiked or retailers just used that perception to lock up merchandise rather than hire staff and close some underperforming stores.

But most of it was real, if not always normative. The number and scale of encampments of un- and under-housed people increased as did their visibility, at least in several cities. Reckless driving increased during the pandemic. Drug use was higher and its consequences—particularly psychopharmacological effects (the effects of being under the influence)—increased as well. There were others but these along with retail crimes seem to have been the ones that really grabbed the public’s attention.

So, what’s the situation now? Well, as with many social problems, it is hard to measure. The Urban Institute has a helpful guide to public disorder measures and what is clear from this list is that finding national statistics is a challenge.

Source: Urban Institute, Preventing Public Disorder.

But a number of influential people have written about disorder, for example, at the end of November in 2024, Matt Yglesias wrote about A Crisis of Disorder. Yglesias notes that crime spiked and declined, but also that, “what is true, though, is that a lot of lower-level disorder that spiked alongside shootings in 2020 never went back to normal.”

There is a slight end-of-history vibe to this statement. Never is a strong word, and implies some level of prognostication, some nodding inclination that disorder will never return to normal. There is little data on disorder, but the data we do have show much more profound declines than Yglesias suggests. Yglesias only identifies two measures that have timely administrative data: unruly airline passenger reports and car accident fatalities. It is about six months since Yglesias wrote this, let’s see where we are today.

Unruly passengers

As Yglesias demonstrates, in mid-November 2024, the FAA reported that disorderly passenger reports were on a trajectory to be unchanged from 2023 and about double pre-pandemic levels. I give Yglesias judos here, because I do this is a valid and clever measure of disorder. But I don’t think it is a reliable one, meaning that it always measures the same thing. I would speculate that it was not just the behavior being reported on that changed after 2020, I suspect reporting behavior changed as well. That is, I wonder whether flight attendants were unwilling to suffer as much grievance after the spike as before the spike and thus were more willing to report disorderly behavior they would have gritted their teeth and ignored in prior years. That there was a spike in unruly passengers in 2020 and it continued for some time is undeniable. That all of the difference between 2024 and 2019 is unruly behavior is probably an overstatement.

Anyway, here's the same graphic through the middle of 2025.

Unruly passenger reports are down about 20% in 2025, and are on pace for 1,600 for the year. Thus, if that trajectory continues, we are headed for relative normalcy soon.

Car Accident Fatalities

Yglesias cites the chart above from Consumer Shield to show that there was a spike in fatal car accidents during the pandemic. Again, there is no disputing the spike. And here, there is no dispute as to the reliability of the data. They are reliable. But the data do show an obvious trend toward fewer fatal accidents. So, where are we in 2025?

It does not appear that Consumer Shield, which produced the original chart, has updated it. But the preliminary final 2024 data from the US National Highway Transportation Safety Agency is available and they estimate there were 39.3 thousand total fatalities in 2024, about the same as 2020. They also provide this helpful context, “the quarterly fatality declines that began in the second quarter of 2022 also continued, with the fourth quarter of 2024 marking the 11th consecutive quarterly decrease in traffic fatalities.” If this trend continues, fatal car accidents should be back around 2015 levels by the end of this year.

And finally, although he has no data, Yglesias notes that, “the question of to what extent shoplifting has actually increased in sharply contested, but at a minimum, retail stores have clearly felt compelled to lock up much more of their merchandise as an anti-shoplifting countermeasure.”

Now, this happens to be something I know quite a bit about, as I testified on the before Congress last year. There are essentially five sources of retail theft, and the retail industry helpfully shares no data on any of it. Anyway, the five categories are:

• Cargo theft (this is stealing a container in transit, or the like)

• Fraud (for example, an employee writing down that you received five packets of Skittles when they received six and eating the missing item)

• Employee theft (this is the employee just grabbing some Skittles off the floor and eating them. I would note that the anti-theft shelves also make it harder on employees to take a five-finger discount)

• Minor shoplifting. These are impulsive crimes of opportunity which (statistically speaking you have engaged in at least once in your life)

• Organized shoplifting

Economists talk about principal-agent problems, where there are mismatched incentives. In this case, the principal-agent problem is that retail operators have a lot of incentive to report that organized criminal rings are robbing them blind. It’s embarrassing to admit that their own accounting failures, mischievous employees, or lack of investment in security are a problem. Better to blame you, the customer, and organized rings (which they hope will cause the police to patrol their stores for free on the public’s dime rather than them having to pay for it (economists call this rent-seeking)).

Anyway, it’s not at all clear that the shoplifting spike was real.

In summary, there was a spike in disorder. And that disorder spike has receded. Perhaps not as much as gun violence, but in a meaningful amount, nonetheless.

To Conclude, a Slight Digression

Now, since I am a fan of long-term trends as a way to understand multi-causal outcomes like Americans fear of crime and perception of high rates of disorder, let me show you one graphic that I think is helpful in understanding the last couple of decades. Beginning in 1996, homeownership rates steadily increase for a decade, then plummet for a decade. Rates have steadily climbed since 2016, but millions of Americans who have previously owned homes do not currently own one.

How does that correlate with perceptions of disorder? In a couple of ways. First, when you own a home, your perception of your own stability tends to be higher as compared to renting. When you go from owning to not owning, that puts you back in an unstable setting, or worse, the transition itself can be traumatic and contribute to additional perceptions of instability. Second, when you rent and live in denser spaces, your tolerance for disruptive or uncivil behaviors in others may be lessened, especially for former homeowners or older renters. The point is that, all else being equal, this change in status may cause you to perceive the world as more disordered even if it is the same as it was before.

I bring this up just to point out that there are likely to be structural issues that affect disorder that are different than the structural issues that affect crime and violence. There is no apparent correlation between homeownership and violence, but there is a credible hypothesis of a relationship with disorder, or at least perceptions of disorder.

And that’s another reason why it is important to think about national forces affecting crime.

Musical Interlude

My kid says you can see us in this picture. I’m wearing a Phillies ballcap. See? Right there. Just out in the middle there, like halfway back.

Anyway, very fun. Lewis Capaldi joined Noah Kahan for Northern Attitudes for us, but I can’t find a good video of that. So here he is from earlier last week in Glastonbury.