Alcohol use is highly comorbid with depression, especially among college students, whose rates of both phenomena are higher than in the general population. The self-medication hypothesis (i.e., alcohol use is negatively reinforced via the alleviation of negative affect) has dominated explanatory models of comorbidity. However, self-regulation has also demonstrated cross-sectional relationships with both depression and alcohol problems and may contribute to the development of alternative comorbidity models. Self-regulation introduces three alternative models that can be tested empirically: (a) a depressed regulation model, (b) a central nervous system depressant model, and (c) a self-regulatory failure model. The purpose of this study was to test the utility of these models (in addition to the self-medication hypothesis) by examining the temporal precedence in the relations between self-regulation, depressive symptoms, and alcohol problems among heavy-drinking college students (N = 393) over 5 assessment points (baseline, 1 month, 6 months, 12 months, and 16 months) using an autoregressive cross-lagged model. Lower self-regulation, and higher alcohol problems and depressive symptoms, prospectively predicted higher depressive symptoms. Higher alcohol problems and lower self-regulation prospectively predicted higher alcohol problems. Only self-regulation prospectively predicted self-regulation. These results were consistent across multiple time points. Findings are consistent with a self-regulatory failure model of depressive symptoms and alcohol problems. Therefore, self-regulation may be an important etiological variable and potential intervention target to reduce both alcohol problems and depressive symptoms among college students. (PsycINFO Database Record (c) 2019 APA, all rights reserved).