Why did crime decline in 2023?

Because local government went back to normal

Last month I wrote that the huge spike in violence in America in 2020, 2021, and 2022 was the result of pandemic-related losses in staff and functionality within local governments. Today, I write that the reason crime has declined so much in 2023 is because local governments have mainly returned to normal.

As the calendar flips over to 2024, I am reading endless stories about the great crime decline of 2023. What these stories seem to have in common is that they are all about the trees, but there’s nothing about the forest. Yes, there have been big investments in police and a renewed focus on specific efforts to fight gun violence. Yes, there have been big investments in community-led strategies, particularly violence interruption programs, where private citizens with lived experience directly intervene after a shooting to break the cycle of retaliation. Both approaches have been productive in reversing the 2020 COVID-19 violence spike.

But focusing on the programs rather than the broader landscape is like focusing on the weather and ignoring the climate—you only get so much of the picture with a zoomed-in lens. So, let’s zoom out—way out.

What almost all of the explanations for the crime decline have in common is that they are activities supported by local government. After a terrible period of underfunding and under-staffing caused by the pandemic, local governments have, by most measures, returned to pre-pandemic levels.

Zooming all the way back to see what local governments are doing allows you to also see all the myriad ways local government fights crime. Through grant-making to CVIs and support for the police, bailiffs, correctional officers, and probation and parole agents (who are all, of course, local government employees) local government directly fights crime and violence.

But local government also employs social workers, public health workers, teachers, and librarians who, by the millions, battle every day to keep our young people safe and headed in the right direction. Their ranks were also decimated by COVID-19, but they are back. And their work matters enormously in explaining the crime decline.

The Crime Decline, 2023 Edition

Crime declined in 2023. We won’t know exactly how much until next fall when the FBI releases the official crimes reported to police numbers for 2023. But others have filled the void, estimating national trends from city data that is publicly available.

The Council on Criminal Justice (CCJ) Crime Trends Workgroup released a report earlier in December that notes that the US is on pace for “the largest single-year homicide reductions in the era of modern record keeping.” Jeff Asher, writing on Substack, reports that, “every type of Uniform Crime Report Part I crime with the exception of auto theft is likely down a considerable amount this year relative to last year.”

It is pretty clear what happened in 2023. But the ‘why’ question remains. Why did crime and violence decline in the US in 2023? Many of the explanations are asymmetric—they explain why crime and violence increased in 2020 but do not explain why crime rates came back down. Huge increases in firearms sales, de-policing, and declining legitimacy of the police all explain the rise in crime but not the fall.

Changes in local government staffing and funding, by contrast, explain both the rise and fall. This explanation does not just fit the data because the timing up and down is the same—it explains the data because there are tangible causal mechanisms within local government activities that explain the crime spike and decline.

Crime and violence declined in 2023 because local governments finally returned to pre-pandemic employment levels.

Local Government Staffing Crashed in 2020 then Slowly Returned to Normal

In February 2020, there were 22,871,000 people employed by government at all levels- federal, state, and local. By May 2020, there were only 21,396,000. That’s a loss of 1.5 million jobs in just two months. It took until September 2023 for those jobs to return.

I made this point last month, but I really want to dive into it here. There are three levels of government—federal, state, and local—and pandemic job losses were not evenly distributed. Each level of government’s role in fighting crime is not evenly distributed either.

When people say ‘government’ they usually mean Washington (or when they say ‘Washington’ they mean the federal government). But the federal government has very little to do with fighting crime and is actually pretty small compared to the other levels of government. Just like Harvard University is just a hedge fund that holds classes on the side, the federal government is mainly just a check-writing machine, which doesn’t require a ton of employees.

Of the three levels of government, the federal government is the smallest, with about three million employees out of twenty-three million at all levels. Occasionally, it gears itself and does something big and tangible, like fighting a war, and hires a ton of people. Also, every ten years the federal government runs the Census, and those big spikes in employment levels in 2010 and 2020 are the hundreds of thousands of boots on the ground necessary to do an accurate count (among other reasons, because some places don’t get mail). But otherwise, as the chart below shows pretty clearly, the COVID-19 pandemic did not affect federal hiring.

State government is almost twice the size of the federal government, in terms of employees. And it was hit hard by COVID. The CARES Act provided about $175 billion to states beginning in March 2020, which mitigated (temporarily) some of the job losses from the pandemic (although with the expiration of those funds in 2021, job losses returned to low levels in early 2022). Hiring began again in early 2023 and is now returning to early 2020 levels. Overall, about 250,000 state government jobs were lost.

The level of government that gets the least attention in policy and political circles—local government—is the biggest, by far. Local governments employ 60 percent of all government staff, between 13 and 15 million employees, which is about five times as many employees as the federal government. And local government was massively disrupted by COVID-19—more than one million jobs were lost. Put another way, 75% of government jobs lost during the pandemic were in local government.

Unlike the federal government which hired Census workers and the state governments which were supported by the CARES Act, there are no discernible turning points in the local government’s return to normalcy. All of the federal aid helped, but it trickled down through states. What really changed local government fortunes was a return to economic normalcy.

The return to normalcy in staffing required that revenues returned to normal. Explaining how local governments are funded requires its own essay, but excellent primers are available from the Urban Institute and the Government Finance Officers Association. But it is relatively easy to see all the ways COVID-19 choked off local government revenues.

Local governments get a lot of funding from user fees—which include user taxes on things like parking meters (which were way down for a long time). They also get a lot of money from state governments and the federal government. While few local governments rely directly on income and sales tax for revenue, they do rely enormously on transfers from their state government and the federal government, who in turn get a lot of their revenue from income and sales tax, which declined a lot during the pandemic. And it took a long time for all of that revenue to start flowing again and from there to start flowing to local governments. 

Local government staffing was also directly affected by large numbers of employees who also got sick from COVID-19: about 3.5 million Americans were hospitalized from COVID and more than one million died. And millions of Americans with public-facing jobs quit. And it took a long time to fill those vacancies. Since much of local government revenues is indirect and difficult to predict, it was far harder for local governments to fill vacancies than it was for other levels of government.

But slowly and steadily over three years, the number of people working for local government returned to pre-pandemic levels.

Why Does Local Government Staffing Matter in Understanding the Crime Decline?

Police

The most direct way local government fiscal health affects crime and violence in America is through law enforcement as local government employs almost all of the police in America. We do not yet know exactly how much law enforcement was affected by the pandemic. For this essay, I extracted “Employed full time, Wage and salary workers, Police and sheriff's patrol officers” from the Bureau of Labor Statistics Current Population Survey and the most recent data are through 2019.

But, we can reasonably speculate that the effect on police staffing was substantial.

In a typical year, about 200 law enforcement officers are killed in the line of duty (mainly traffic accidents). However, in 2020, more than 400 died and in 2021, more than 600 died. The Department of Justice COPS office estimates that at least 68% of officer deaths in 2020 were COVID-related and other media report that COVID was the leading cause of death in 2021 as well. In addition to the COVID-related deaths, many officers were disabled. The University of Illinois Chicago Law Enforcement Epidemiology Project uses workers’ compensation claims to estimate officer disability. In 2021, they estimate that there were 73,560 injuries among police officers and court personnel, with 1/3^rd attributed to COVID-19, or about 25,000 COVID-related injuries in a single year.  

So the direct impact on police staffing was substantial.

Changing Police Practices to Protect Officer Health and Safety

It is also clear that there were substantial changes in how police patrolled during the pandemic. In particular, proactive policing was curtailed. According to a survey on behalf of the International Association of Chiefs of Police:

By March 23, 2020, 43% of responding agencies stopped or significantly changed their responses to CFS, 57% reported a decline in CFS, 61% implemented policies to reduce proactive stops, and 73% limited community policing activities. As of May 10, 2020, a second survey wave indicated that 53% of the responding agencies continued to have policies that limited proactivity, and 64% were still limiting community-oriented policing activities—both slight decreases from wave 1.

But police patrols are only a part of local government functions around health and safety. Local government also funds court personnel, local jails, and probation and parole. All of these services were directly impacted by COVID-19. Again, from Lum, Maupin and Stolz (2022):

The significant impact of the pandemic on every aspect of life included unprecedented changes in the criminal justice system (for an overview, see National Commission on COVID-19 and Criminal Justice, 2020). For example, because of social distancing guidance and the risk of contagion and death, court services attempted to reduce person-to-person contact and postponed hearings and trials,⁵ leading to extensive downstream backlog impacts and changes in court operations (Baldwin et al., 2020; Chan, 2021; Godfrey et al., 2022; Jurva, 2021; Witte & Berman, 2021). Correctional systems also adjusted, attempting to restrict intake and hasten release to reduce the incarcerated population and alleviate the spread of COVID-19 in confinement (Carson et al., 2022; Hawks et al., 2020; Marcum, 2020).⁶ Parole and probation supervision and treatment services were also affected, with officers and providers unable to meet with their clients (Schwalbe & Koetzle, 2021; Viglione et al., 2020).

Data on local government staffing and my own repeated interactions with people involved with the criminal justice system from jurisdictions across the country suggest that returns to normalcy took years, resolving only recently, or still remaining in flux. Things are a lot better today than they have been at any point in the last three years, but recovery is ongoing.

If you believe that certainty of detection, swiftness of response, and severity of punishment have an effect on violent crime (and almost everyone believes that at least certainty matters), then it is not hard to see how the disruptions from COVID would lead to more violence.

And, it is equally straightforward why returns to normalcy in policing, courts, jails, probation, and parole would reduce violence.

Finally, I would like to point out that almost all of the narratives around policing during COVID-19 have confounded changes in policing to protect officer health and safety with changes in policing in response to the George Floyd protests. The evidence above is overwhelming that most departments substantially changed their policies to protect their officers. The evidence that similar changes occurred during (or after) the protests is far thinner.

Education

But that is just the tip of the iceberg on what local government does to keep people safe. The data below from the Census Department makes it clear that by far the largest role for local government employees is in education. In fact, as many people work in education in local government as work in all of state and all of the federal governments put together!

The evidence that improvements in educational attainment reduces crime and violence is overwhelming. “Empirically, an increase in educational attainment significantly reduces subsequent violent and property crime yielding sizeable social benefits” and three-quarters of state prisoners do not have a high school diploma (Hjalmarson and Lochner, 2012). The mechanism is obvious: more education improves expected returns from legal work and thus increases the cost of criminal activity.

Thus, it should be no surprise that huge reductions in teaching staff were accompanied by big increases in crime and violence in 2020 and 2021, and bug declines in 2023 after a full year of full-staff in-person learning.

But the other big categories of local government employment matter as well. Hospital staff, health staff, and public welfare staff routinely come into direct contact with victims of crime. Without their support and links to services, victims’ risk of revictimization or retaliation is much higher. An under-staffed court system that holds remote court hearings and limits visitation does little to administer certain, swift, and effective justice.

There is much more to be said about the role of local government in fighting crime and violence. One last point may help to drive this idea home. In 2017, NYU professor Patrick Sharkey and colleagues at NYU published an influential study that found that community-based non-profits with a mission to fight local crime and violence had an outsized impact.

Drawing on a panel of 264 cities spanning more than 20 years, we estimate that every 10 additional organizations focusing on crime and community life in a city with 100,000 residents leads to a 9 percent reduction in the murder rate, a 6 percent reduction in the violent crime rate, and a 4 percent reduction in the property crime rate.

But these community-based organizations do not exist in a vacuum. They are often funded by the local government. Local governments link them to training and technical assistance, create peer networks where they can learn from other direct service providers, and participate in convenings where they can learn new strategies and tactics. Local government links them to researchers who can create performance management systems—so they can repeat their successes and avoid repeating their mistakes—and learn about evidence-based best practices.

All of this ground to a halt in 2020. All of it is back up and running in 2023.

In Summary

A simpler way to say all of this is just to say that the explanation for the great crime decline of 2023 was a regression to the mean, and local government was the mechanism.

What does that mean going forward? It means that a continuing crime decline of the decade-long sort America experienced in the 1990s is very unlikely. Crime and violence in America was going up (slowly) from 2014 to 2019 so if we are regressing to the mean, we are regressing toward an upward-trending mean.

It means that we have to embrace different strategies if we are going to make meaningful strides toward a safer America. It means we have to think beyond the police and the courts and prisons. It means we have to think about violence prevention as something teachers and social workers and hospital staff and librarians do and not just something the police do. It means we have to think beyond individual risk factors and think about risk conditions that are imposed on whole communities by concentrated poverty and loose gun laws.

It means we have to think about the forest.