How I Learned to Stop Worrying and Love the Crime Decline

In 2021, many folks paused the COVID-19 doom loop by spending a couple of hours with the drippingly sanctimonious Don’t Look Up. The idea of the film was that terrible catastrophe awaited those who failed to recognize blindingly obvious signals, specifically, the threat of climate change. Smugness aside, Don’t Look Up copies the narrative of oodles of Shakespearean tragedies—Lear, Hamlet, Othello, Macbeth—so it’s a verifiably tried and true formula.

But what about the converse? What happens when we are showered with treasure and fail to manifest abundance? What happens when we collectively fail to recognize success?

That is the enigma of today. Unemployment is, has been, and projects to be historically rare. The stock market froths with an end-of-the-1990s hey, look, we discovered the internet!-like optimism, with repeated all-time highs on the backs of our expectations for a boiling AI future. And, of course, violence has fallen precipitously.

Now, a zillion words have been written on the social psychology of the whole matter. How our perceptions, attitudes, and beliefs are at odds with the truth we fail to embrace. How partisanship, pandemic residue, and distrust in the media all combine into a frothing stew of disbelief. How we live in a post-truth world, poisoned by bad faith arguments, conspiracy theories, and outright propaganda.

To all that, my response is, meh.

In my humble estimation, each new era of the American psyche is special, but not unique. The specific facts today are different than in the past, but that does not make this the end of history. There is a German word for this, because of course there is, Endzeitgefühl, which generally translates to the ‘feeling of end times’.

I do not question that we are immersed in Endzeitgefühl. I, however, doubt that we have entered a post-truth world (which I guess would be Postfaktische Welt, although that doesn’t feel like the German translation adds very much).

Sometimes, it just takes a while for big changes to sink in. And sometimes, it takes a specific thing to happen for big changes to sink in. I’d like to propose here what that specific thing is. And whether it has happened. Let me leave you in suspense about that for a few paragraphs.

The thing I am interested in understanding, specifically, is whether we have reached a level of post-truth that is sufficient for a collective rejection of the idea that we are in the middle of a meaningful and substantial decline in crime and violence. I’ve made this argument several times, but bear with me for a moment because there is new data.

First, it is important to note that we have little data about the latest trends in crime. In part, this is because the US invests less in crime statistics than we do in agriculture statistics. In part, this is because the FBI changed how it collected data from law enforcement in 2021, and as a result, we know almost nothing about crime in 2021 and less about recent years than we know about the more distant past. I am delighted to be a part of the Council on Criminal Justice’s Crime Trends Workgroup, which is releasing our report next week, and so I will have a lot to say about crime data problems after that.

Now, that said, we do have (preliminary) official statistics from the FBI for the first quarter of 2024—January through March. Bloomberg News did a really nice job of summarizing those data.

The data show how the first quarter of 2024 (January through March) compares to the first quarter of 2023 (also January through March). Crime is down by double-digits in all eight of the big crime buckets, what we used to call Part I crime. Also, Bloomberg helpfully shows how these 2023 rates compare to the 2022 rates. Compared to 2022, 2023 crime rates were also down for seven of the eight categories of crime, everything but motor vehicle thefts. The point is that the declines in blue come on top of big declines from the previous year.

One key point to make is that these data are comparing winter 2024 to winter 2023. And I have been getting the question: won’t these declines disappear once summer violence heats up? My answer is that they might, or might not, but it will not have anything specifically to do with it being summer.

Crime last year was seasonal, as it always is, with big increases in the summer months. But crime and violence always go up in the summer. And crime went up in the summer of 2023 compared to earlier in the year, just as it will this year. Since we are comparing the high levels of last summer to the high levels of this summer, the fact that crime will also go up in the summer of 2024 does not mean the big crime declines above will disappear. They will only disappear if this summer’s peaks are closer to last summer’s peak than this winter’s were to last winter’s.

It is worth noting that according to the Gun Violence Archive, nonfatal shootings in the peak summer months last year (June, July, and August) were only down 2.8% compared to summer 2022. So, summer 2024 will be compared to summer 2023 when shootings were still relatively high. The biggest declines in shootings appeared to begin around Labor Day 2023.

Will the summer of 2024 sustain these gains? We don’t know. But more violence every summer does not mean more violence this summer than last summer.

Trends in Violence

Historically, the FBI has been reticent to release national crime statistics until they are quite certain that the data will withstand close scrutiny. The FBI is processing data from more than 15,000 agencies, so that’s a big endeavor. However, the FBI has received a lot of criticism for the 2021 data problems and is under increasing pressure to release data in a timelier fashion, so here we are, with an early look into 2024 trends. Compared to Q1 2023, the official (preliminary) Q1 2024 FBI data show:

• Homicide (-26.4 %)

• Robbery (-17.8%)

• Aggravated Assault (-12.5%)

• Sex Assault (-25.7%)

• All Violent Crime (-15.2%)

My guess is that these data will be revised downward. Those who have looked closely at the data report that they seem incomplete for some cities, and that the data are only partially reported. Thus, when complete data is eventually included, those data will inevitably show more crime and thus smaller declines.

This is fine. Revising earlier data is common in big statistical releases. At the beginning of every month, the government (the Bureau of Labor Statistics) releases two measures of unemployment for the prior month. The Current Population Survey (CPS) interviews 60,000 households and asks the respondents about their employment status. The Current Employment Statistics (CES) survey interviews 119,000 businesses about their employment numbers. These are huge sample sizes, and they generate very precise estimates, even down to small places and small job categories.

Even with these massive samples, these numbers always get revised, twice. Once the following month and once the month after that. Why? Because new information comes in.

So, the idea that the FBI will have to revise its Q1 2024 numbers once new data comes in should not be a surprise, nor a particular concern. It’s pretty clear where the problems are, and how the revision is going to go.

Just a quick digression. You may be asking yourself, hey, why don’t we do that for crime data? Why don’t we invest and conduct monthly surveys of criminal victimization at this kind of scale to allow close to real-time estimation of national crime trends?

We should absolutely do that.

Other Crime Trends Data Sources

Given that there is likely to be some revision in the (provisional) Q1 2024 FBI data, where else might we look to get a sense of what the final number might be?

Three ongoing data collections provide context for the FBI’s numbers. The Major Cities Chief’s Association reports the results of a survey of their membership, which is made up of 70 of the biggest cities in the US. Comparing the first quarter of 2024 to the first quarter of 2023, they find:

• Homicide (-17.3 %)

• Robbery (-3.0%)

• Aggravated Assault (-8.0%)

• Sex Assault (-16.7%)

Jeff Asher reports that homicide through May 30^th is down 18.5% for his sample of 260 US cities. My calculation of the shooting data he reports finds that shootings are down 22% comparing YTD 2024 to YTD 2023 (I believe this to be YTD through May 30^th). Data from the Gun Violence Archive is a little harder to analyze, but fatal shootings appear to be down around 13% for 2024.

Take this all together, and you end up with the graphic below. The data from 2005-2023 are from the FBI, and for 2024 I weight the data; Asher (50%), MCCA (40%), and FBI data (10%).

I estimate that the homicide rate in the United States, as of April 1, 2024, is 4.5 per 100,000. For comparison, the homicide rate was 6.8 per 100,000 at the peak of the pandemic and 5.1 per 100,000 at the beginning of the pandemic. The current level is approaching the lowest level in the last 20 years (4.4 per 100,000), which is also the lowest level in the last 60 years.

Violent Crime

Turning to all violent crimes, I again use a combination of sources to calculate the violent crime rate per 100,000. I use official crime statistics from the FBI for 2005-2022. For 2023, I use the preliminary FBI estimates. For 2024, I use a weighted average of preliminary Q1 2024 data from the FBI (25%) and the MCCA (75%). The graph below multiplies the homicide rate by one hundred for ease of comparison.

Two facts are clear from this graphic. First, while the rise in homicide during the pandemic was substantial, the rise in all violent crime was much more muted. And second, violent crime rates, from a historical perspective, are at really low levels.

Aggravated assaults, which can be domestic or on the street, rose by more than 10% in 2020. But in 2020 robbery, which in large part is a street crime of opportunity, fell almost as much as assault rose. This was not a surprising result given the substantial decline in outside-the-home pedestrian traffic throughout 2020.

In the 16 combinations of one-year decline for each of four years (2021-2024) and for each of the four categories of violence (homicide, aggravated assault, robbery, and sexual assault) declines occurred in every combination, except robbery in 2022, which was up slightly, 1.3% from 2021.

My estimated violent crime rate for Q1 2024, as of April 1, was 323.6 per 100,000, which is the lowest violent crime rate in 55 years, since 1969.

The VIX is in!

Finally, I want to make a quick point about why this news is being treated skeptically. And I don’t just mean partisan opponents of the current administration are skeptical—even those inclined to support it are having difficulty accepting these statistics. I wrote a little about this six weeks ago, on May 4, 2024. In that essay, I argued that a big problem with accepting these numbers is that they hadn’t been officially ordained, blessed by the FBI, and given the imprimatur of official national statistics.

But now they have! On June 10^th the FBI officially released the preliminary numbers described above. And… my sense is that folks still don’t believe it. Partially this is due to concerns raised almost immediately about the comprehensiveness of the data. And partially because the drop, as reported, is so jaw-droppingly huge that it gravitationally defies belief.

That said, I think there is more to it. I think there is a larger force exerting its mass on our belief in good news—a force larger than crime and violence. And that is our collective gross uncertainty about the state of things. The problem is that fear, residual and otherwise, is crowding out evidence and data. But that is changing.

The pandemic unleashed untold existential terrors on the US. Threats of unstoppable disease. Threats of historic unemployment, closed schools and businesses, the economy grinding to a halt. And yes, for some, threats to liberty. As these terrors were vanquished, to some extent, new terrors arose, specifically real and spiraling inflation—for many folks who came of age after the 1980s, this was their first real exposure to inflation and the havoc it wreaks on daily life. That threat too has waned, at least empirically.

The thing about existential threats is that they are, well, existential, and emotional, and thus hard to vanquish with logic and data. We will only really turn the corner on all of these terrors and begin to accept the evidence as presented when we … turn the corner. But again, that seems to be happening.

First, it’s worth putting some historical context on the existential threats we just experienced. Here’s a chart with 80 years of unemployment. 2020 stands defiantly. Wow, that’s a lot of unemployment and it’s right there in the rearview mirror, like Mount Rainier as you head north on I-5. Going all the way back to World War II, you will find no period when unemployment was as high. None.

You might also notice that prior to 2020 it had been a very long time since the last recession of any type (these are the eponymously grey periods in the graph). If you are of the median age of all Americans, 38-ish, you have seen only four recessions in your lifetime, one when you were in kindergarten. That does not make for a very resilient populace. But even that mild rebuke is probably uncalled for given the recent threats.

I could show you a dozen more of these graphics on various economic and social health measures, but they all look like this—or like this upside-down, in a bad way. We know from Tversky and Kahneman that people are risk averse, specifically that people fear losing things of value much more than the benefits of gaining something of the same value. And in moments when that risk aversion is validated, it takes time to re-adjust to normal levels.

So, when a never heard-of before virus combines with catastrophic economic and social loss, hey, it’s gonna take a while to get over that. You can’t just rub some dirt on it and move on.

The thing is though, we are moving on, slowly, slowly, slowly, but definitely moving on. And also quite definitively on a path to continue to move on in the near future.

My thesis here is that as we move away from fearful times, we will be more open to accepting the facts as they are and less buffeted by emotion.

To make this point, I want to borrow a statistical measure from the markets called the VIX, which is the Chicago Board Options Exchange Volatility Index. In the simplest terms, the VIX measures investor fear. From Investopedia, “[v]olatility, or how fast prices change, is often seen as a way to gauge market sentiment, and in particular the degree of fear among market participants.” It’s important to note that this is a near-term measure of fear, over the next 30 days.

Another great measure of fear, or contentedness, is the Michigan Consumer Sentiment Index, which is a survey of consumers about current economic conditions and expectations about future economic conditions. It has become increasingly political, in the sense that people increasingly conflate their economic reporting with their political ideologies—that is, if you don’t want the person currently in office to stay in office that will lead you to state that things are bad (while you would report the opposite if your person was in office). Nevertheless, it has some interesting properties.

The chart is complicated, but the point is simple. In the chart above, consumer sentiment is in red (top) and the VIX volatility measure is in blue (bottom). Now, here’s the takeaway: when VIX spikes, consumer sentiment collapses. And a spike in the VIX never, ever causes a positive change in consumer sentiment. But when the VIX is quiet and short-term fear is low and steady, consumer sentiment increases. Another way to say this is: spikes in the VIX show powerful emotions at work—quiet periods are driven by other forces, presumably logic and facts. How is the Vix doing lately?

This is why I am optimistic that evidence and data have a bright future.

Where’s the bottom of the trend?

The only correct answer is: I have no idea.

But as of right now, there is more reason to be optimistic about public safety than there has been for decades.

Don’t forget to sing when you win.

Musical Interlude – Endzeitgefühl Revisited

Look, America is in a funk. Blue notes abound. The stock market is at an all-time high, violent crime is nearing 50-year-lows, and unemployment is microscopic as well. But we are not feeling it. It’s a vibe-cession.

What’s the cure for a vibe-cession?

Better vibes!

To quote the great 20th-century philosopher Mike Damone, just act like wherever you are, that’s the place to be.

<Looks around>

<Spreads arms wide>

<Nods head>

Isn’t this great?

Ladies and gentlemen, may I present to you, REM.