Background and aimsPrevious research indicates that compared with individuals with lower socio-economic status (SES), individuals in higher SES groups are more often drinkers but those who drink report drinking smaller amounts more frequently. We aimed to decompose trends in self-reported alcohol consumption in Germany into age, period and birth cohort effects and examine whether these effects varied by SES.DesignAge-period-cohort (APC) analysis using data from eight waves of the cross-sectional German Epidemiological Survey of Substance Abuse (ESA) collected between 1995 and 2015.SettingGermany.ParticipantsThe analytical sample included n=65821 individuals aged 18-64years reporting alcohol use within the last 30days.MeasurementsAlcohol measures included drinking prevalence, alcohol volume and prevalence of episodic heavy drinking (EHD). Educational attainment was used as an indicator of SES. A series of generalized linear and logistic regression models, including both main and interaction effects of age, period and cohort with SES, were estimated.FindingsRegression models revealed significant interactions between APC effects and SES on two alcohol consumption measures. Higher SES was consistently associated with drinking prevalence across age (P<0.001), period (P=0.016) and cohort (P=0.016), and with volume of drinking in younger cohorts (P=0.002) and 50+-year-olds (P=0.001). Model results were inconclusive as to whether or not APC effects on EHD prevalence differed by SES.ConclusionsIn Germany, there are positive associations between socio-economic status and alcohol consumption during the life-course, over time and among birth cohorts. Three groups appear vulnerable to risky drinking: high socio-economic status young birth cohorts who drink high average quantities, low socio-economic status young birth cohorts who show a risky drinking pattern and high socio-economic status adults in their 50s and older who increase their drinking volume beyond that age.