Although it is well established that workplace demands and culture can affect employee well-being, to what degree might these same factors have lingering implications on individual well-being after employees retire? To begin to answer this question, in this article we propose and test a model explaining how retiree alcohol consumption may depend on pre-retirement contextual conditions. Specifically, we propose and test a moderated-mediation model in which two ambient work-unit characteristics – work-unit stress climate and work-unit drinking norms – moderate the indirect effects of retirement, via distress, on modal alcohol consumption (i.e. the typical quantity and frequency of alcohol consumed). Using a prospective study design and a multi-level, zero-inflated negative binomial model for predicting modal alcohol consumption, our findings lend partial support for the proposed model. We found retirement (vs continued employment) to be associated with a heightened probability of being an abstainer after retirement eligibility (i.e. at Time 2), regardless of the hypothesized unit-level moderators. Still, retirement had mixed effects on the level of modal consumption among those not abstaining at Time 2, with these effects being partially mediated by distress and contingent upon unit-level stress climate and unit-level drinking norms.