Neighborhood Network

November 17, 2021

Dear Neighbor,

I’m thankful for the dedicated first responders that keep our businesses, churches, neighborhoods, and schools safe.

As we look ahead to a new year, the Central Texas Public Safety Commission is focused on the future. We’ll work on patrol staffing, long-term budget planning, and EMS, fire, and police contracts.

When the Austin City Council voted to restart the training academy this past summer, it was clear the police department needed a new evidence-based staffing model. Austin was already experiencing patrol staffing challenges. Worsening response times and vacancies required moving officers from specialized (investigative) units back to patrol.

The Austin Police Department’s patrol staffing crisis won’t improve until we:

  1. Implement patrol staffing model recommendations
  2. Increase training capacity
  3. Reduce attrition

In May, the Central Texas Public Safety Commission pledged to work with the Austin Police Department and business groups to recommend a new patrol staffing model. Initial findings from that research are available now. A final report is due by December 31.

Researchers at the University of New Haven developed a state-of-the-art machine learning model to analyze patrol response and staffing. The program analyzed more than five years of patrol data, including 5.8 million calls for service.

Machine learning recommends responding to urgent calls for service (crimes in progress) within 6 minutes and 30 seconds. (Currently, response time averages exceed 8 minutes and 30 seconds.) The model projects a staffing level of approximately 704 working patrol officers to meet that target. The police department is authorized 774 patrol positions, but has a lower staffing level due to vacancies and leave. Patrol should have 859 positions to maintain 704 working officers given vacancies, which requires 85 additional positions.

Three important notes:

  1. Patrol staffing recommendations are based on machine learning. A large-scale survey by Texas State University is underway to offer additional data about how residents prioritize police services. Community input may adjust model variables.
  2. Analysis only applies to patrol, not specialized (investigative) units and administration. Separate modeling will determine appropriate staffing for those functions.
  3. Researchers continue to explore patterns in calls for service related to day of week, time of day, and sector to seek further efficiency.

Just adding positions doesn’t solve the problem. By the time the current cadet class graduates in late January, the police department will likely have more than 250 vacancies. If the class graduates 70, that leaves at least 180 vacancies. Next year, two full academy classes and a modified class might add 200 officers, but the department will continue to accrue more vacancies.

Again, implementing research recommendations, increasing training capacity, and reducing attrition are the most important first steps to solve the patrol staffing problem.

Happy Thanksgiving!



P.S. First responders tackle the toughest jobs in our community. Help tell the stories of everyday heroes. Send an email to [email protected]. I’ll share as many as I can through the Neighborhood Network and social media.