BackgroundPenalties are a key component to improve road user safety, but previous studies suggested that they might not be successful in reducing crashing in offending drivers. However, these studies were not able to consider important crash risk factors in the analysis that might confound the results. Using data from a large prospective cohort study of young drivers in New South Wales, Australia, we explored if novice drivers with driving offences have a higher rates of car crash and if these differences are explained by established crash risk factors.MethodsWe used data from a 2003/2004 Australian survey of young drivers, linked to police reported offence and crash data, hospital data and deaths data up to 2016. We used Poisson regression models adjusted for confounders to estimate the association between driving offences during 2003–2006 with car crash during 2007–2016.ResultsThe study cohort comprised 20 781 young drivers of whom 7860 drivers (37.8%) had at least one driving offence and 2487 (12.0%) were involved in at least one crash. After adjusting for confounders in the regression model, drivers with three or more driving offences had 2.25 (95% CI 1.98 to 2.57), 2.87 (95% CI 1.60 to 5.17) and 3.28 (95% CI 2.28 to 4.72) times higher rates of any crash, crashes that resulted in hospital admission or death and single vehicle crashes compared with drivers with no driving offences.ConclusionMeasures that successfully mitigate the underlying risk factors for both, crashes and offences, have the potential to improve road safety.