[Note: This article previously appeared in Michigan Police Chiefs, Summer 2024]

Police agencies face myriad staffing challenges. Among these are changing preferences for police work, training officers for increasingly complex environments, diversifying police workforces to reflect the communities they serve, and ensuring officer wellness. Perhaps the most pressing of these challenges, as reported by agencies themselves, is the need to maintain force levels by balancing recruitment and retention.

Citing decreasing numbers of applicants and increasing numbers of retirements or other departures from the field, many police agencies report being understaffed. This, in turn, raises the questions of how we know an agency is understaffed and what, ideally, its staffing levels should be. Answering these questions can be more complex than it appears.

Understandably, agencies, community leaders, and other stakeholders seek indicators and rules-of-thumb to serve as benchmarks for their target staffing level. Benchmarks such as staffing ratios, peer-agency comparisons, and allocation levels are frequently used because of their ease of identification, historical and widespread use, and ease of understanding. Yet their use can lead to myths about their appropriateness that hinder agencies as they seek to provide police services as efficiently as they can. In this article, we discuss these common benchmarks, perils of their use, and an alternative way forward.

Staffing ratios

Staffing ratios determine an ideal number of officers based on a simple formula, such as ensuring a specified number of officers per unit of population. This or that ratio of officers per capita is often put forth as a standard, but there appears to be no reason for this aside from ease of calculation and access to data—for example, the FBI’s yearly data on staffing levels and rates for communities, broken out both individually and in various categories such as city size, county, state, and region. While the per capita method is easy to calculate and apply, it does not account for how officers spend their time nor for community conditions and expectations. It also cannot guide decision makers on how to deploy their officers, such as by time, geography, or function.

Some of the shortcomings of the per capita method are evident when comparing communities with similar number of officers. In 2022, municipalities nationwide reported 1.09 officers per 1,000 population, while those in Michigan reported 1.66. The Michigan municipalities with the highest numbers of officers per 1,000 population in 2022 were Mackinac Island (10.22 per 1,000) and Mackinaw City (8.18), both popular tourist destinations with seasonal peaks. Grosse Pointe (4.18), Detroit (3.80), and Kalamazoo (3.28) have roughly similar ratios, but Grosse Pointe and Kalamazoo have public safety officers providing both police and fire services, boosting their staffing numbers but also expanding their responsibilities. Nearby communities with similar populations and agencies can vary in their ratios. For example, Warren and Sterling Heights are adjacent communities north of Detroit, each with about 135,000 population, but Warren in 2022 had 224 officers (1.63 per 1,000 population) and Sterling Heights had 156 (1.18). Simply knowing the size of a community reveals little about its preferences for service style, need for service in form or quantity, geographical constraints, or other related work factors.

The disadvantages in using per capita ratios have led the International Association of Chiefs of Police, among others, to advise against using them in determining police staffing.

Peer agencies

Another method police agencies may use to determine their appropriate staffing levels is by comparing themselves to a peer agency. Yet this method, too, has problems. The first of these is identifying what truly is a peer agency. Warren and Sterling Heights, for example, may appear to be peer communities, but they differ on some characteristics. Sterling Heights has a larger foreign-born population and more non–English speakers, while Warren has several large employers (e.g., the General Motors Technical Center). Differing community compositions and levels of economic activity can lead to needs for different types and levels of police services.

Even if an agency could identify a truly peer agency and community, there is no guarantee that the peer agency has identified the ideal number of police officers for its community or that it’s effectively maintaining its target amount. For example, Livonia and Clinton Township are both suburban Detroit communities with more than 90,000 population and other similar demographic and socioeconomic characteristics. Both communities in 2022 had Part I crime rates of a little less than 20 per 1,000 population, both below the rate of 24 per 1,000 population across all Michigan municipalities. Livonia accomplished this with 1.44 sworn officers per 1,000 residents while Clinton Township had 0.94. Given what appears to be similarly successful efforts at controlling crime, it is difficult to say which should be the “peer” standard for the other.

Similarly, Grosse Pointe Farms and Grosse Pointe Woods, adjacent suburban Detroit communities, have similar demographic and socioeconomic characteristics, though Grosse Pointe Woods has about 5,000 more residents. Part I crime rates in both were a little more than 10 per 1,000 population in 2022. Both have consolidated public safety departments, but Grosse Pointe Farms reported 3.34 sworn officers per 1,000 residents while Grosse Pointe Woods reported 1.67. Again, communities need to consider for themselves what level of policing they need and why, and need to consider their own unique characteristics more than what seemingly similar communities do.

Allocation levels

Perhaps the most common method for determining staffing need is comparisons with the authorized allocation level. This is the number of staff for which an agency is budgeted. While this authorized level may be determined through a formal staffing assessment, it is often driven by staff ratios and comparisons with what are thought to be peer agencies, and can be the result of resource availability, politics, history, and incremental change. It typically does not reflect any identifiable criteria such as demand for service, community expectations, or efficiency analysis. It can also be difficult to determine what exactly is the authorized level, particularly in large departments which at any given time may have substantial numbers of officers unavailable because of leave or other limited capacity.

Because the authorized level is often derived independently of workload considerations, an agency may be able to meet workforce demand and even other performance objectives with fewer officers than authorized, especially if ways to improve efficiency exist. The authorized level can become an artificial benchmark for need, creating the perception that the agency is understaffed and overworked if the actual number of officers does not meet the authorized level. It can become a false ruler for gauging staffing needs and deficiencies. This, in turn, can adversely affect morale, strategic planning, and other organizational issues. Furthermore, unless an agency staffs above the authorized level, fluctuations in recruitment, selection, training, and attrition may lead to actual staffing levels falling below authorized levels.

The way forward

A more accepted and comprehensive method for determining police staffing need is to consider the actual workload of an agency. Workload-based analyses derive staffing indicators from demand for service. A workload-based analysis is the only approach to systematically analyze and determine staffing needs based on demand (e.g., calls for service), service style or performance (e.g., how much discretionary time an agency wishes officers to have), and other agency characteristics.

The workload approach estimates staffing needs by modeling levels of police activity. While there is no universally accepted method for conducting workload-based assessments, such approaches typically incorporate work schedules, backup personnel needed, benefit time for officers, and training time to estimate allocation and deployment across shifts and districts.

Workload-based assessments are generally applied to allocating patrol officers, the “backbone” of the force, but they can be applied at every level of an agency for all key functions. Workload-based assessments can be modeled to facilitate a discussion between what a community wants or desires and what it can afford. The importance of the workload-based approach to staffing is demonstrated by it being codified by the Commission on Accreditation for Law Enforcement Agencies.

The workload-based approach does have some limitations. It relies on average rather than extreme demand in producing estimates. Response times can vary by community and can be lengthy in large jurisdictions, which can affect workload calculations. Finally, workload-based estimates generally work best for communities with at least 15,000 resident calls for service per year.

Even with these shortcomings, staffing models and decisions based on actual workload and performance objectives are preferable to other methods that are more mythologized than reflective of environmental and agency-specific characteristics. No single metric or benchmark should be used as a sole basis for determining an agency’s staffing level. A workload-based approach, combined with consideration of other metrics an agency and community agree reflect actual demand and desires for policing services, can help place the discussion of police staffing in an appropriate practical context. It can help each community determine what policing services it needs, what additional services it may want, and how best to achieve these levels through staffing or other means.

Jeremy M. Wilson, Ph.D.
Director, Police Staffing Observatory
Professor, School of Criminal Justice, Michigan State University

Clifford A. Grammich, Ph.D.
Director, Birdhill Research and Communications

Tweet