Speeding penalties: fewer repeat tickets, but signs of liability shifting

A reform with two kinds of punishment

Do harsher speeding penalties discourage repeat offending, or encourage people to shift responsibility? In the preprint Sanctions, Evasion, and Deterrence: Evidence from Speeding Enforcement, Romain Gauriot and Maria Wisniewska of Deakin University investigate a Queensland policy change that lets them examine both possibilities.

On 1 July 2022, Queensland increased speeding fines across speed categories. One narrow category also received extra demerit points: drivers caught travelling 11–12 kilometres per hour above the limit saw their fine rise from AUD 183 to AUD 431, while their points increased from one to three.

That distinction matters because points accumulate toward licence suspension. A fine is an immediate financial cost; additional points can also make future violations more consequential. But camera tickets initially go to registered vehicle owners, who can nominate the actual driver instead. That legitimate correction process also creates opportunities for dishonest transfers.

Comparing drivers around the change

The researchers use administrative speeding records and compare offences occurring before and after the reform. Their main method, regression discontinuity in time, estimates whether outcomes changed abruptly at the implementation date, allowing for separate time trends on either side.

The main analysis covers 150 days before and after the change, with 149,755 fixed-camera offences and 438,590 mobile-camera offences after exclusions. Reoffending means another recorded speeding offence within a year. The main measure excludes the first 14 days, allowing time for drivers to receive their initial notice.

This is not a randomized experiment. Its causal interpretation depends on drivers caught around the cutoff being comparable, with no simultaneous change in enforcement or other influences that could explain a jump in outcomes.

Camera type is therefore important. The authors report that fixed cameras provide clearer frontal identification, whereas mobile cameras may photograph vehicles from behind. Observable driver characteristics were broadly balanced at the cutoff for fixed-camera tickets, but changed for mobile-camera tickets. The researchers consequently use fixed-camera offences to estimate repeat-offending effects and mobile-camera offences to investigate transfers.

Where repeat detections fell

Across fixed-camera offences, the preferred specification estimates a 1.6-percentage-point reduction in the probability of another recorded speeding offence within a year.

The larger result concerns drivers initially caught 11–12 kilometres per hour above the limit. Their estimated reduction was 4.6 percentage points, compared with a pre-reform reoffending rate of 34.8%—roughly a 13% relative decline. For other speed categories, which received higher fines without additional points, the estimated reduction was smaller and not statistically significant.

Several checks support this pattern. The authors test alternative time windows, examine artificial reform dates in earlier years, and compare the affected category with drivers caught 9–10 kilometres per hour above the limit. They also find a reduction when subsequent offences are restricted to fixed-camera detections, reducing concern that mobile-camera transfers explain the result.

However, the reform did not vary fines and points independently. The 11–12 category received both extra points and a larger fine increase than nearby categories. The evidence therefore supports an effect of the combined sanction increase, not a clean estimate of the effect of points alone. Nor does an insignificant estimate for other categories prove that higher fines never deter speeding.

Transfers increased, but dishonesty is unobserved

For mobile-camera tickets in the 11–12 category, the event-study estimates suggest transfer rates rose by approximately 30–60% relative to their low pre-reform level. Comparable increases were not evident for fixed cameras or other speed categories.

The statistical evidence is less decisive than that headline suggests. The immediate discontinuity estimate for the affected mobile-camera category was not conventionally significant at the preferred window, with a p-value of 0.112. Significance varied across alternative windows. An additional comparison with the nearby speed category showed a more persistent increase over subsequent months.

Most importantly, the records identify final fine recipients and whether a transfer occurred, not who actually drove. More nominations could reflect dishonest blame shifting, more accurate identification of drivers, or changes in who was detected. Strategic evasion is an interpretation consistent with the pattern, not an observed fact about individual transfers.

What the findings can establish

Mobile-camera detections also declined around the reform. The data cannot distinguish increased caution or camera avoidance from changes in deployment intensity. That uncertainty further limits interpretation of the transfer results.

The study measures recorded repeat speeding, not every speeding episode, crash, injury, or death. It therefore does not directly establish a road-safety benefit or predict what another jurisdiction would achieve.

As preprint evidence, its central lesson is conditional: cumulative penalties may strengthen deterrence, while opportunities to reassign responsibility may generate another response. Penalty severity and reliable attribution need to be considered together.