BACKGROUND: The use of private motor vehicles places a considerable burden on public health. Changes in workplace car parking policies may be effective in shifting behaviour. We use a natural experimental design to assess whether changes in policy were associated with differences in commute mode. METHODS: We used cohort data from participants working in Cambridge (2009-2012). Commuters reported their trips and travel modes to work over the last week, workplace car parking policy and socioeconomic, environmental and health characteristics. Changes in policy were defined between phases (1608 transition periods; 884 participants). Using generalised estimating equations, we estimated associations between changes in parking policy and the proportion of trips that (i) were exclusively by motor vehicle, (ii) involved walking or cycling and (iii) involved public transport at follow-up. RESULTS: 25.1% of trips were made by motor vehicle, 54.6% involved walking or cycling and 11.7% involved public transport. The introduction of free or paid workplace parking was associated with higher proportions of motor vehicle trips (11.4%, 95% CI (6.4 to 16.3)) and lower proportions involving walking or cycling (-13.3%, 95% CI (-20.2 to -6.4)) and public transport (-5.8%, 95% CI (-10.6 to -0.9)) compared with those with no workplace parking. Restrictive changes in policy were associated with shifts in the expected direction but these were not statistically significant. CONCLUSION: Relaxation of parking policy was associated with higher proportions of trips made by motor vehicle. Further longitudinal and intervention research is required to assess generalisability of these findings.