Yes, of course crime is way down

The people telling you crime isn't falling are feeding you a line.

Photo by Josh Hild on Unsplash

It’s election season, so the narratives are flying fast and furious. One of the MAGA side’s narratives is that America has become a violent, chaotic, and ungovernable place under Biden, and that Trump will restore order. This is a replay of a narrative they successfully used in 2016, when it was sort of true; in fact, it was the theme of Trump’s convention speech. It’s natural for the MAGA folks to want to re-up this golden oldie. Their story of a world in chaos (thanks, of course, to weak and hapless Democrats) encompasses border security, the wars in Ukraine and Gaza, and, most importantly, spiraling crime rates.

But there’s at least one big problem for this narrative: Crime is falling fast in America.

When one dares to speak this truth on social media, one is typically deluged by angry MAGA types who deny that crime is falling. Here is just a tiny sample of those replies:

An unfortunate lesson from the past few years is that if Americans don’t want to believe a plain and simple fact, they simply won’t. No amount of data, careful studies, or reasoned argument will convince them. And so a whole lot of Americans will march to the polls in November absolutely, utterly convinced that crime has soared under Biden.

A whole lot…but maybe not all. I stubbornly believe that there is still a large subset of Americans who do respond to evidence, and so all I can do is present that evidence and hope that someone out there cares.

So let’s take a look at the data. First, let’s talk about the FBI’s estimates. The FBI has a data set called the Uniform Crime Reports, where they try to estimate the number of crimes in the U.S. based on what the police report to them. This data shows crime falling across the board in 2023, and falling at an accelerating pace in early 2024:

Source: Justin Fox

Now, as soon as you present this data, a bunch of MAGA people will pop up to claim that the FBI data is fake. These claims are either grave misunderstandings or willful misrepresentations of a real thing that happened in 2021. In 2021, the FBI changed the system by which city police departments report crime. The new system was hard to use, so almost a third of police departments didn’t send in their numbers that year.

Some people apparently think this means that the FBI simply assumed the number of crimes went down because fewer crimes were being reported to them. That’s just not how it works. The FBI tries to estimate how many crimes happened in the cities that don’t report, based on the numbers for the cities that do report. The totals they show include these estimates of the unreported crimes.

Now, those estimates might be off! If the cities where crime has increased the most are also the ones that fail to report, the FBI could easily miss this, and its estimates could be too low. But the FBI is not simply tallying up all the crimes that police departments report to it. They are not simply saying “Durrrr, fewer cities told us their crime numbers, so I guess there are fewer crimes happening in America!”. The FBI are not that dumb, nor do they think you are that dumb.

But anyway, all of this is a little beside the point, because by 2022, the reporting problem had basically been fixed. Jeff Asher, a renowned criminologist who studies this data for a living, has a good post explaining how by 2022, police departments had started reporting again:

The FBI estimates crime counts to fill in gaps when estimating national figures when agencies fail to report. It's usually not a big deal when only 5 or so percent of crime data, but it's a big problem if you're estimating 35 percent. This led to…a ton of questions about the nation’s crime figures for 2021. If you can see 90 or 95 percent of a picture then you can guess what that picture is with much better accuracy than if only 65 percent of the picture is visible.

Fortunately, the FBI backtracked in 2022 and allowed non-NIBRS agencies to submit data and coverage improved dramatically to just about 94 percent in 2022 - a level that was comparable to historical reporting norms…[M]ore than 85 percent of the US population was covered by a NIBRS-reporting agency by the end of 2023 (including 9 of the 10 largest cities in the country). Adding in non-NIBRS agencies should mean normal coverage again for 2023 data.

Here’s his key graph:

So the FBI data for 2022 onward should be pretty good.

But if it were only this one data source telling us that crime is falling, I would still tell you to be skeptical. There are plenty of cases in which even the most carefully gathered data turns out later to have big problems that change our entire understanding of the facts. If it were only the FBI reporting that crime fell, there’s a possibility that a few years down the road, some careful data sleuths would discover some error showing that the MAGA people were right after all, and crime really did soar under Biden.

Fortunately, though, we actually have a bunch of data sources on crime — especially on murder and other violent offenses. And although there are slight disagreements about how much crime has fallen over the past two and a half years, there’s broad agreement on the direction of the trends.

First, Jeff Asher has his own crime data, collected by his company, AH Datalytics. AH Datalytics gets this data by calling up police departments and asking them for their crime numbers. It shows a dramatic drop in murder in 2023 and early 2024, to levels lower than what prevailed in 2016-2019. Here’s a chart from Justin Fox, combining FBI numbers through 2022 with Asher’s numbers for 2023 and early 2024:

Source: Justin Fox

Another data source is the Council on Criminal Justice, which looks at a sampling of about 38 U.S. cities and tries to track crime reports for those cities. They find drops in most types of crimes since 2021, including murder, although the drop is less than the one Asher finds:

Source: CCJ

Another data source is the Major Cities Chiefs Association, which includes 70 big city police departments that pool their data on crime. Although this only captures a fraction of crime in America, it’s clear that crime rates are falling:

Source: Major Cities Chiefs Association

Two other data sources include the Gun Violence Archive, which collects open-source data on crime, and the CDC, which collects mortality data from the health care system. Both of these sources show murder (or fatal shootings) falling steadily since 2021, though not yet back to the level of 2019:

Source: Jeff Asher

So we have six data sources, using six different methodologies, all telling us the same story about crime: the FBI, AH Datalytics, the Council on Criminal Justice, the Major Cities Chiefs Association, the Gun Violence Association, and the Center for Disease Control. All of them agree that murder peaked in 2021, and those that report other types of violent crimes agree that they also peaked in 2021.

One data source could have some catastrophic error that renders it effectively fake. But six different data sources, using six different methodologies? No way.

Now, there is one other data source that shows a rise in crime in 2022. This is the National Crime Victimization Survey, collected by the Bureau of Justice Statistics. This survey asks people whether they have been victims of a crime. Obviously murder is not counted in these numbers, since people who have been killed tend not to answer the phone and say “Yup, I was murdered.” But the NCVS does show a small uptick in other types of violent crime in 2022, which is the most recent year available:

Source: BJS

But note a couple of things from this graph. First, even with the uptick, the NCVS shows violent crime at lower levels than it was in the early 2010s — or pretty much any other period in our recent history. Second, the NCVS shows no increase in violent crime in 2020 or 2021, the years of the big crime wave — in fact, it shows that those years were record lows for violent crime. So that data is, frankly, pretty suspect.

In any case, the overwhelming message from the various data sources remains the same — crime has fallen a lot since 2021. The MAGA narrative of rising crime under Biden is not based on facts; it is based on willful fantasy.

Now, that fact still leaves two big questions unanswered:

1. Why did crime soar in the first place, and how much was Trump’s fault?

2. Is Biden actually responsible for the drop in crime, or did he just get lucky?

Here we leave the realm of simple fact, and enter the murkier world of complex theory and supposition. I can’t give you reliable answers, but I can give you my best guesses.

First of all, it now seems obvious to me that the recent crime wave was the result of social unrest. There were no obvious policy changes, economic factors, etc. that could have caused the rise, and popular long-term explanations like childhood lead poisoning and abortion rates just changed way too slowly to have caused such a sudden rise.

Instead, social unrest — most commonly referred to as the “Ferguson Effect”, after the 2014 Ferguson protests in the wake of Michael Brown’s death — is the overwhelmingly likely explanation for the rise in crime. In fact, the “Ferguson Effect” is really two effects — 1) police scaling back their policing effort, and 2) people being mad about the incidents that sparked the unrest.

There’s a decent amount of academic evidence that this explanation explains the modest crime wave of 2015-16. For example, here’s Cheng and Long (2022):

We conduct a single-city analysis of St. Louis and a multi-city analysis of 60 large cities. Our regression discontinuity and difference-in-differences estimates provide consistent and strong evidence that those high-profile killings reduced policing activities, including police self-initiated activities and arrests. The estimated reduction in self-initiated activities shows that police officers proactively withdrew from law enforcement activities, providing direct evidence of de-policing.

And here’s Premkumar (2019):

The effects differ by offense type: Reduced police effort yields persistently fewer arrests for low-level offenses (e.g., marijuana possession) but limited changes in arrests for violent or more serious property crimes. I show that decreased interaction with civilians through police stops may be driving the results. However, the increase in offending is driven by murders and robberies, imposing significant crime costs on affected municipalities. The effects only occur after there is broad community awareness of the incident…I find that the reduction in low-level arrests corroborates public scrutiny as the causal channel. Finally, I provide evidence that the increase in offending is driven by both a response to the reduction in policing effort and a reaction to the police killing itself, suggesting that measures to reduce use of force should be prioritized.

Note that this is a more nuanced story than most conservatives or MAGA people will tell you. Yes, decreased policing effort after an anti-police backlash is a factor driving up crime. But anger at the police killing that initially prompted the backlash also seems important.

Basically, we have a good idea what events caused the crime wave of 2015-16, and that probably applies to 2020-21 as well. But that doesn’t settle the question of whom to blame. If you’re a conservative, you’ll tend to blame rioters and anti-police activists for cowing the cops into passivity, and for overreacting to police killings. If you’re a progressive, you’ll tend to blame cops for riling up the populace and increasing local distrust of law enforcement by shooting too many people in the first place. So this is a bit of a Rorshach test.

The same thing applies to whether you think Trump was partially responsible for the huge crime wave of 2020. If you don’t like Trump, you’ll claim that his very presence in the White House exacerbated unrest, making people more likely to rebel against the police because they associated cops with the MAGA movement. (This seems plausible to me, but maybe that’s just because I don’t like Trump.) If you’re a Trump supporter, you’ll say that people who get riled because of politics and riot and commit crimes are responsible for their own actions, and that it’s never acceptable to do these things even if you don’t like the President. (I can see this side of the argument as well.)

In any case, we’re not going to settle the question of who’s to blame for the crime wave, because blame is in part a moral judgement rather than a purely factual one.

As for how much Biden has contributed to the crime decline since 2021, that’s less of a moral question and more of a factual one, but it’s still subject to huge uncertainties. Maybe Biden’s calming, grandfatherly presence in the White House helped cool the passions unleashed by the Trump era — but if so, we’ll never really know.

Biden is unquestionably a tough-on-crime President. From the very first moment that “Defund the police” became a rallying cry of the BLM protesters, Biden came out strongly against the idea. Early in his presidency, he released a plan to put 100,000 more cops on the streets of America.

It’s not clear, though, how much Biden’s tough-on-crime attitudes have translated into policy at the local level. We do know that increased policing does reduce crime by a significant amount. Biden’s Safer America Plan never made it past Congress. He did manage to increase spending on the COPS program and a few other federal initiatives to support local cops, but federal spending on policing did fall as a percentage of gross domestic product in 2021 and 2022. And other than funding law enforcement, the federal government does relatively little to affect crime trends at the local level, where policing policies and budgets are mostly determined.

Biden did, however, do one thing that might have made a substantive difference: He gave money to local governments. Justin Fox argues¹ that this allowed local governments to sustain or even increase their spending on policing, at exactly the time when pro-police sentiment was rising:

In Bloomberg Opinion’s chart-based assessment of Biden’s tenure in March, I concluded that his American Rescue Plan’s $350 billion in aid to state and local governments, which he strongly encouraged them to spend on law enforcement, arguably did play a role in reducing crime…

On crime, my theory is that it was Democratic local and state politicians, many of them elected in the wake of and in reaction to Trump’s 2016 victory, who were pushing counterproductive policies, especially after George Floyd’s murder by a Minneapolis police officer in May 2020…The election of Biden, a Democrat who opposed calls to defund the police, paved the way for a reconsideration in many cities. Minneapolis voters rejected a plan to replace the city’s police department. San Franciscans threw out their progressive district attorney. New Yorkers elected a former cop as mayor. And so on.

So perhaps Biden did help bring crime down — he certainly wanted to, despite opposition from some in his own party. But how much he contributed is hard to know.

In any case, this doesn’t answer the question of whether Biden or Trump would do more to tamp down crime. I suspect that Biden’s combination of soothing grandfatherly attitudes and quiet increases in police funding would be much more effective than Trump’s flamboyant authoritarian rhetoric. But regardless, it’s clear that crime has fallen for at least two and a half years under Biden, and that’s a trend I would very much like to see continue.

---

1. Fox also argues that Biden’s stimulative macroeconomic policies sustained an economic boom that reduced the incentive for crime. But given weak correlations between crime levels and economic factors, I consider this explanation to be very tenuous.