She was right.
Federal investigators identified seven former complainants willing to provide interviews. One Black hotel employee said Mercer stopped him walking home after midnight and held him for forty minutes because he became nervous.
A Latino mechanic said Mercer pulled him from a vehicle after he asked why officers wanted to search his trunk. A Black college student described being detained outside his own apartment after a neighbor reported a suspicious man near mailboxes.
The facts varied. The rhythm often did not.
Mercer started with vague suspicion. A person asked why.
The question became attitude. Attitude became noncompliance.
Noncompliance became officer safety. Then force or search followed.
Investigators requested five years of Mercer’s discretionary pedestrian and vehicle-stop data. Harbor Glen stored information across several poorly connected systems, so civilian data analyst Priya Desai spent days cleaning duplicates and matching officers with comparable assignments.
The pattern was difficult to dismiss. Black residents represented about thirty-eight percent of the population in Mercer’s usual patrol area.
They accounted for nearly seventy-two percent of his discretionary stops. His consent-search rate was more than twice that of similar officers, yet his searches produced contraband less often.
Statistics could not tell investigators what Mercer believed in his private thoughts. They could show repeated choices.
Priya found something else. Mercer used phrases like furtive motion, bladed posture, appeared nervous, and for officer safety far more frequently than officers working similar districts.
Then she compared reports. Twelve contained entire sentences copied almost word for word from older incidents.
One recurring line read, The subject repeatedly glanced toward an apparent avenue of escape while making nervous hand movements. The same sentence appeared in reports involving a parking lot, a bus stop, a vehicle stop, and an apartment hallway.
Even the punctuation stayed identical. Priya called Agent Ruiz.
“He isn’t documenting unique observations.” “What is he doing?”
“Using language that already survived review.” Investigators then checked who approved those reports.
Sergeant Douglas Tate’s name appeared repeatedly.
Tate was the same supervisor who advised Mercer after Maya’s stop. Federal investigators expanded preservation requests.
Inside City Hall, Police Chief Marcus Bell reviewed the history and realized the department’s problem could no longer be described as one officer losing control at an intersection. Several complaints contained warning signs.
Several settlements existed. Yet no system forced supervisors to compare them.
The department had stored history without learning from it.
A week later, former Internal Affairs commander Olivia Chen contacted Agent Ruiz through counsel. Four years earlier, Olivia had supervised the Professional Accountability Unit and repeatedly objected when senior command softened findings before final approval.
She was later transferred into policy planning. Olivia brought dates and file numbers, not stolen evidence.
She explained the process. Investigators would write findings based on video, witnesses, and policy.
Then command review meetings sometimes altered language. Unreasonable force might become tactics inconsistent with best practice.
False statement might become insufficient evidence to determine memory accuracy. If policy failed to address conduct precisely, managers often treated the gap as protection for the officer instead of a reason to improve policy.
Olivia said Mercer was viewed as productive. He generated arrests.
Volunteered for difficult shifts. Wrote large numbers of citations.
Few commanders asked how many of those stops actually produced valid enforcement outcomes.
Agent Ruiz asked about Devin Shaw. Olivia’s expression changed.