What could Chicago do with 600 more police officers?

Part 2 of the Workforce Allocation Study

---

Before we get started, a plug for two upcoming events:

1. Join other City That Works readers Wednesday, May 13th at Midwest Coast Brewing from 5:30-7:30pm. We’ll be joined by local Alderman Walter “Red” Burnett, who’s been fighting for more housing in the 27th Ward. Sign-up here to let us know you’re coming.

2. Richard will be back at the Hideout with State Rep. Eva-Dina Delgado and the Indivisible Chicago Crew on Monday, May 18th from 7-8:30pm. We’ll be talking about Gov. JB Pritzker’s Build Plan. Doors (and the bar) open at 6. Register here.

---

We’ve spent a lot covering the way the Chicago Police Department allocates its officers. In short, we aren’t putting officers where they’re needed most, and don’t provide them with consistent and capable support. The result is that our Police Department does less to protect Chicagoans than it should – both from violent criminals, and from officers themselves.

To address these issues, the Chicago Police Department contracted with Matrix Consulting to conduct a comprehensive study of CPD’s workforce allocation. Two weeks ago, we covered the most striking conclusion from that Matrix study: that more than 600 police officers are working desk jobs instead of serving in roles that require a gun, badge, and sworn arrest powers. This week, we’ll look at how we could better deploy those officers.

As of April 2026, CPD has roughly 11,587 sworn officers across the department. Our focus today is the regular patrol units: the 3,950 regular patrol officers who roll around the city in blue shirts and marked cars. Matrix also had a lot to say about other units in the Department, including tactical teams and detectives, which we’ll get to in the coming weeks.

Matrix Workforce Allocation Study

Matrix thinks we need 273 more beat cops*

Patrol units need to do a mix of two things: respond to calls for service, and spend time on ‘proactive policing,’ which can include building community relationships, understanding crime patterns in a neighborhood, and being positioned in high visibility places to deter crime. If officers are stretched thin, they lose the opportunity to do that proactive work, which is a necessary part of building a safer city in the long run.

The Matrix team set a target for the share of officer time spent on proactive policing at 40%. Then they did the math on the total number of available working hours by district. They also looked at the total number of calls for service by district, and the average time it took to respond to those calls. That allowed them to calculate the total time requirement, and see how much proactive time was left over by district. You can see the results below.

Matrix Workforce Allocation Study

Notably, in total across the department, right around 40% of time is spent on proactive policing today. According to Matrix’s calculations, we could just move officers between districts, and add just 7 additional officers to get to a 40% rate everywhere.¹ But that’d be difficult to execute in practice – both because Alders (and residents) likely don’t want to see cops pulled out of their neighborhoods, and because once officers bid for a position in a district, under the rules of the labor contract, CPD cannot re-assign them without their consent.

So instead, the Matrix team suggests holding the number of officers in better staffed districts constant, and adding 273 positions in the districts with low levels of proactive time. The graphic below outlines what those changes would look like today:

Crucially, however, they’re not designed to be static. As the Matrix team stresses, it’s important that CPD regularly revisit staffing ratios and move officers to where they’re needed next.² If we don’t, over time we’ll gradually see a divergence in where officers are deployed vs. where they’re needed, and be back in the same position we are in today.

These changes are designed to complement a couple of other reforms recommended by the Matrix team. I’ll spare you the full details here, but by adding officers and simplifying the current shift rotation schedule, CPD should have an easier time maintaining the consistent staffing and oversight requirements required by the consent decree. And by assigning officers to the CPD’s 66 sectors, instead of 269 beats, it will be easier for CPD to match officer headcount to workloads.

What should we make of the Matrix findings?

Broadly speaking, these changes are very good. It’d be valuable to ensure officers receive consistent and adequate supervision. And we’d all be better off with fewer officers behind desks, and more cops out on the streets.

But it seems weird to me to index this off of calls for service. That’s because calls for service are not a true reflection of crime in a neighborhood – they’re a reflection of both need and a community’s willingness to call the police in response. Nationally, just 40% of violent crimes are reported to law enforcement. In neighborhoods where police don’t respond promptly or appropriately to calls, those rates can be much lower.

How can we tell? Look at homicides. They generally come with a body, and are matched against death records. As a result, they are almost always captured in crime statistics. That means that they’re one of the better tools researchers can use to estimate overall crime or disorder.³ It’s hard to imagine that a neighborhood has a high number of homicides, but low rates of non-fatal shootings, assaults, or robberies, for example.

This comparison reveals some pretty stark differences across the city. In 2024, Chicago’s 3^rd District on the South Side, ranked dead last in officers per homicide (4.44). One in every 961 calls for service was for a homicide. In contrast, the North Side’s 20^th District had more patrol officers per homicide than any other in the city (77.33). But in 20, there were ten times as many calls for service per homicide (9,788) as there were in the 4^th District. In short, District 20 residents are calling the police far more often, for far less serious issues than residents in District 4 are.

This pattern repeats itself across the city – relative to the number of homicides (our best indictor of the actual prevalence of violent crime), Black neighborhoods are much less likely to call the police.

Why is that the case? It’s complicated – CPD’s long history of misconduct and mistrust in Black communities is likely part of the problem. But another issue is likely simpler: According to the Matrix data District 4 ranked dead last in response time for low-priority police calls (53.5 minutes on average). That’s more than twice as long as the response time in the 20^th District (25.3 minutes). If the cops take forever to show up, why bother calling in the first place?

But if residents of high violence neighborhoods aren’t calling, they aren’t getting included in the Matrix data. And a staffing analysis that allocates officers based on calls will magnify allocations to high-trust, low-crime neighborhoods where residents already call the police, and minimize the number of officers routed to the lower-trust, higher-violence neighborhoods where they’re needed most.

If you look at the ratio of officers per homicide, instead of officers per service call, you see a very different picture. While the 12th and 8th Districts (where residents call police relatively frequently) could still certainly use more officers, the biggest shortages are in the 3rd, 6th and 10th Districts - which are higher violence areas on the South and West Sides - but less well staffed than the highest-violence areas like Englewood and Garfield Park.

You obviously can’t ignore calls for service – cops need to show up when they’re called, and it’s important to have a baseline level of police protection regardless of the number of murders that occurred the prior year. It’s not hard to see why leaders on the Southwest side have been calling for additional resources. But we should probably be targeting a higher share of proactive time in high-violence neighborhoods where trust in police is lowest.

As Jack Maple, the architect of New York’s Compstat revolution put it way back in 1999:

In the old way of policing, the poor people got [underserved], but maps don’t know the difference between a poor person and a rich person. The dots are the same size regardless. A robbery is a robbery. Those 10 dots tell you where to put your cops. The dots don’t say, “This affected Donald Trump; it’s a press case.” That is one beauty of mapping.

If you go by the maps, you see where the crime is and you deploy there. That crime goes down. Then those people that need the help the most get the best service.

Let’s be clear about what we’re trying to accomplish

Again, I don’t think you should take this as a knock against the whole Matrix study – if we implemented the workforce allocation recommendations today and updated staffing decisions regularly the City would be in a much better position.

But we should be clear about what we’re trying to do. The real reason you should care about this year-long, thousand page study, isn’t because CPD needs to hit a 40% benchmark for proactive police work. It’s not even because we have to come into compliance with a consent decree. It’s because if Chicago is going to become a sustainably safer city, we desperately have to break out of the “old way of policing” that Maple referred to 27 years ago.

That will require reforms that go far beyond the requirements of the consent decree. We can start by putting officers where they’re needed most.

1. Notably, the study didn’t explain how they got to the 40% target.

2. That means CPD will likely need to either negotiate to update the bidding process in the next contract, or be very quick to close out open positions in districts where calls for service are falling.

3. They’re not perfect – in particular it’s worth noting that because there are fewer of them, homicides have more year over year variance. There are also likely types of calls that occur at similar rates regardless of the level of homicides in a neighborhood (like wellness checks, for example). Ideally, it’d be great to compare all this data on a multi-year timeframe. I can say that I’ve cut the data for 2025 and it’s broadly similar to the 2024 numbers discussed here. But I don’t have the Matrix call-for-service data for 2025, so I can’t do a full multi-year comparison.