BACKGROUND: Long-term population-representative data on motor fluctuations and levodopa-induced dyskinesias in Parkinson's disease is lacking. METHODS: The Cambridgeshire Parkinson's Incidence from GP to Neurologist (CamPaIGN) cohort comprises incident PD cases followed for up to 13 years (n = 141). Cumulative incidence of motor fluctuations and levodopa-induced dyskinesias and risk factors were assessed using Kaplan-Meyer and Cox regression analyses. RESULTS: Cumulative incidence of motor fluctuations and levodopa-induced dyskinesias was 54.3% and 14.5%, respectively, at 5 years and 100% and 55.7%, respectively, at 10 years. Higher baseline UPDRS-total and SNCA rs356219(A) predicted motor fluctuations, whereas higher baseline Mini-mental State Examination and GBA mutations predicted levodopa-induced dyskinesias. Early levodopa use did not predict motor complications. Both early motor fluctuations and levodopa-induced dyskinesias predicted reduced mortality in older patients (age at diagnosis >70 years). CONCLUSIONS: Our data support the hypothesis that motor complications are related to the severity of nigrostriatal pathology rather than early levodopa use and indicate that early motor complications do not necessarily confer a negative prognosis. © 2019 International Parkinson and Movement Disorder Society.