As shown in Figure 1, the
distribution of speeds ticketed by the Florida Highway Patrol between 2005 and 2015
shows substantial excess mass at speeds just below the first fine increase. Meanwhile, a
remarkably small portion of tickets are issued for speeds just above. We take this bunching
as compelling evidence that officers systematically manipulate the charged speed, commonly
charging speeds just below fine increases after observing a higher speed, perhaps
to avoid an onerous punishment for the driver.3 We also find that a substantial portion of
officers throughout the entire state exhibit no such bunching (Figure 2), suggesting that
the aggregate bunching reflects officer discretion rather than drivers choosing to bunch
below a fine.
By studying officers’ choice to discount a speeding ticket, this paper makes four contributions.
First, we document the presence of racial bias in officers discounting. We
show that this bias is robust to accounting for all relevant contextual factors, including
past driving history, and persists after accounting for racial differences in true speed.
Second, and most importantly, we generate officer-level estimates of lenience and bias
by comparing each officer’s behavior towards whites and non-whites. We find that 25%
of officers are biased against black and Hispanic drivers and reject that this number is
greater than 30% of officers. Third, we correlate these estimates to officer demographics
and show that officers tend to favor their own race, older officers are more racially
biased, and women and college educated officers are less biased on average. We show
that officers who are more lenient in general are less likely to receive complaints or use
force on the job. Officers who are more biased against minorities seem to use more force,
though standard errors are too large to say conclusively. Fourth, we explicitly model each
officer’s decision process, accounting for potential racial differences in speed, allowing us
to perform various counterfactuals and evaluate which personnel policies may be useful
for ameliorating the treatment gap. Directly firing biased officers and hiring more women
and minorities reduces the gap, but not substantially. More effective is to re-assign the
most lenient officers to heavily-minority areas

Its only 2.5 pages in FFS. The officers fudge the numbers because they can. I've had it happen to me all four times I've been pulled over for speeding. In both a bright yellow 240SX with airbrushed flames and a Honda delsol si. Each time the officer found me "right under" the threshold for a serious fine (when I very clearly wasn't). Once I was given a $7 insurance processing fine, once I was fined for the lower threshold, and the last 2 times I was let go when they asked about the ID badge i had on my console.