Background: Some parents supply alcohol to their children, reportedly to reduce harm, yet longitudinal research on risks associated with such supply is compromised by short periods of observation and potential confounding. We aimed to investigate associations between parental supply and supply from other (non-parental) sources, with subsequent drinking outcomes over a 6-year period of adolescence, adjusting for child, parent, family, and peer variables. Methods: We did this prospective cohort study using data from the Australian Parental Supply of Alcohol Longitudinal Study cohort of adolescents. Children in grade 7 (mean age 12 years), and their parents, were recruited between 2010 and 2011 from secondary schools in Sydney, Perth, and Hobart, Australia, and were surveyed annually between 2010 and 2016. We examined the association of exposure to parental supply and other sources of alcohol in 1 year with five outcomes in the subsequent year: binge drinking (more than four standard drinks on a drinking occasion); alcohol-related harms; and symptoms of alcohol abuse (as defined by Diagnostic and Statistical Manual of Mental Disorders, 4th edition [DSM-IV]), alcohol dependence, and alcohol use disorder (as defined by DSM-5). This trial is registered with ClinicalTrials.gov, number NCT02280551. Findings: Between September, 2010, and June, 2011, we recruited 1927 eligible parents and adolescents (mean age 12·9 years [SD 0·52]). Participants were followed up until 2016, during which time binge drinking and experience of alcohol-related harms increased. Adolescents who were supplied alcohol only by parents had higher odds of subsequent binge consumption (odds ratio [OR] 2·58, 95% CI 1·96–3·41; p<0·0001), alcohol-related harm (2·53, 1·99–3·24; p<0·0001), and symptoms of alcohol use disorder (2·51, 1·46–4·29; p=0·0008) than did those reporting no supply. Parental supply of alcohol was not significantly associated with the odds of reporting symptoms of either alcohol abuse or dependence, compared with no supply from any source. Supply from other sources was associated with significant risks of all adverse outcomes, compared with no supply, with an even greater increased risk of adverse outcomes. Interpretation: Providing alcohol to children is associated with alcohol-related harms. There is no evidence to support the view that parental supply protects from adverse drinking outcomes by providing alcohol to their child. Parents should be advised that this practice is associated with risk, both directly and indirectly through increased access to alcohol from other sources. Funding: Australian Research Council, Australian Rotary Health, Foundation for Alcohol Research and Education, National Drug and Alcohol Research Centre.