Self-efficacy (confidence in one’s abilities to execute behavior to reach ones goals) has been reported to reduce the impact of stress on symptoms of depression and anxiety. However, findings have been inconsistent regarding the two different conditions and the temporal resolution (individual situations [micro-level] vs. general disposition [macro-level], cross-sectional vs. longitudinal). By analyzing available data from an epidemiological cohort study, we explored the impact of self-efficacy, stress, and their interaction on symptoms of depression and anxiety in N = 1072 adolescents (age 14–21) from the general population. We conducted questionnaire assessments (macro-level) and ecological momentary assessment (EMA; micro-level). Multiple linear regressions and hierarchical linear modeling were applied to explore cross-sectional and longitudinal associations on both levels. On the micro-level, cross-sectional results indicate that self-efficacy lowers the impact of stress on depression, but not anxiety, during everyday life. Longitudinal effects were not found. On the macro-level, cross-sectional analyses revealed a buffering effect of self-efficacy against stress for anxiety, but not depression. Longitudinally, only direct effects could be observed for self-efficacy on anxiety and for stress on depression and anxiety. It appears that the adverse impact of stress on depression might be too strong to be buffered by self-efficacy in general, but only on a micro-level. The macro-level protective effect of self-efficacy regarding anxiety on the other hand might not be retrievable during everyday life. Results imply that therapy of depression might benefit from strategies to lower stress, whereas treatment of anxiety might focus on increasing self-efficacy to reduce avoidant behavior.