New Study’s Staffing Model Calls for 108 New Officers

Austin Police Department Downtown headquarters (Photo by John Anderson)

A new study funded by the Greater Austin Crime Commission, an outside private support group for local law enforcement agencies, finds that if the Austin Police Department wants to maintain an average response time of 6½ minutes to 911 calls to meet community expectations, the city should budget for 882 patrol positions, which is 108 more than are funded in the current city budget.

Giovanni Circo, a professor at the University of New Haven (a private school where more than a third of the student body is majoring in criminal justice), is the study's lead author. Circo developed a machine learning model – basically, a complex algorithm that analysts can use to work with huge data sets – to analyze 6 million officer responses to 2 million calls for service between 2016 and 2020. Even when only looking at the most urgent calls – those coded by 911 dispatchers as P0 or P1, typically meaning an imminent threat to life and safety – that's still nearly 346,000 encounters; less urgent calls were also examined to better understand how officers use their time.

More than half of the P0 or P1 calls were categorized as either "check welfare" or "disturbance," which are broad categories for calls that require a police response. As APD's Chief Data Officer Dr. Jonathan Kringen describes it, the goal is risk prioritization. A "disturbance" call may include a neighbor smoking a joint in public (a lowest­-priority P3 call), behaving erratically (a higher risk level), and appearing to hold a weapon (P0). Maybe when the police arrive, officers ascertain quickly that the neighbor is actually holding a spoon, which they had no way of knowing in advance. Perhaps the 911 caller made up the part about a weapon, knowing that claiming someone is waving a gun will force the police to respond to what's actually a nuisance call. (This is roughly the sequence of events that preceded the April 2020 killing of Mike Ramos by APD Officer Christopher Taylor, now under indictment for murder.)

In all P0 and P1 calls, first responders are faced with two choices and four outcomes: respond quickly or not (mostly determined by staffing), and find either an urgent situation or a false alarm. By drilling down into the outcomes of each of a huge number of calls, with special attention to those that resolve public safety threats (like arrest, or recovering a firearm), the study concludes that to respond appropriately to all high­-priority calls, APD needs to hire more police officers.

"What the study shows is staffing creates time and availability," Kringen said in an interview with the Chronicle. More patrol officers will improve response times, but can also produce the 35% of work hours dedicated to "uncommitted time" that APD has targeted for patrol to spend on pro­active policing and community engagement, instead of rolling out to one 911 call after another. But are there ways to attain those goals without spending more money to hire more officers according to a staffing formula? Even if that formula is more sophisticated and data-driven than the arbitrary 2.0 officers per 1,000 residents championed by the Austin Police Association, which would translate into twice as many new hires?

Justice advocates say yes, and that Austin's ongoing, if slowing, effort to "reimagine public safety" is exploring just such solutions. One point of focus for reformers is officer-initiated (OI) calls, which Circo's study design deliberately excludes. A 2020 analysis by the city's Innovation Office found that 35% of patrol time is already being spent on activities officers choose to investigate on their own. About 94% of the 636,906 OI calls reviewed in that study (from 2017 to 2020) were P3, the lowest priority, which includes the 24% that are traffic stops. Only about 14% of OI calls resulted in a police report, about half as often as do 911 calls.

Just Liberty's Kathy Mitchell, a member of the City-Community Reimagining Public Safety Task Force, argues that City Hall should focus on making APD patrol more efficient before simply adding $10 million to its budget to hire 108 officers. "All the areas we can make change in now are ineffective uses of officer time [that] put more time on the tally sheet," Mitchell explained, adding that the study "assumes if there is a shortage of officer time, the only way to solve it is to hire more officers." Austin remains a low-crime city where most types of crime steadily declined during the time frame of the GACC and Innovation Office studies, so Mitchell thinks policymakers should not be falling back on historical responses. "We can either hire more police and do more of the same, or we can try a more innovative approach, like taking police out of situations they shouldn't be responding to," Mitchell said.

Though the primary recommendation of the GACC study is to hire more officers, it also suggests that a more efficient allocation of officers could improve response times as well. "Staffing based on anticipated demand, like how a power company might turn on more generators during the winter," also helps to improve response times, Circo told us. "If we can forecast when we might need more officers to respond to crime," the time it would take to respond to urgent calls would drop.

Circo also tells us that he welcomes the calls from local policy analysts to make more of the data and code he used to form the model public. "I think releasing some of it in the future could be a possibility," the professor told us. But some of the findings will be submitted for peer review, and there is commercial value in the model – something that could be sold to other cities as a way of determining appropriate staffing levels there. Ultimately, he says, the data belongs to APD and the study to GACC, so it is up to them to determine how much and when to release it.

At a Jan. 11 press conference, APD Chief Joseph Chacon described the GACC study as "one of several tools to help us achieve the right balance for the department and community" in terms of staffing levels. He did not rule out asking Council for a midyear budget amendment to help pay for the new officer positions; after meeting with Council members to discuss the reports, he plans to present to City Manager Spencer Cronk a road map for how to implement the report's recommendations "in a sensible manner that fits with our city budget."

That response highlights another problem: APD can't fill the vacancies it already has. As of Jan. 19, the department has 227 vacant positions, according to APA Pres­ident Ken Casaday; a new cadet class of 58 is slated to graduate from the Austin police academy in two weeks, and APD has funding for two more classes this fiscal year (ending Sept. 30). Increasing the number of vacant positions to 335 would not lead very quickly to more patrol officers on the street; Chacon says he's developing a "multi-pronged effort" to address attrition through recruitment and retention. "It's important that we recruit the right type of folks to be police officers in our community," Chacon said. "We have a very robust and diverse recruiting unit that is working on doing that very thing for future classes."