Positive moods are thought to restore self-control resources following depletion. However, it is not well understood whether this effect is due to affective valence (pleasantness), arousal (activation), or a combination of both. Across four studies, we set out to investigate the role of positive moods on cognitive and behavioral measures of self-regulation in an ego-depletion paradigm. In studies 1 and 2, we independently manipulated affective valence and arousal and assessed self-regulation with a Stroop task. Results did not suggest a restorative effect of either on cognitive resources. In study 3, we employed both behavioral (the 'handgrip task') and cognitive (Stroop) assessments of self-regulation. Again, no significant effect of mood was observed on the Stroop task. Additionally, participants did not persist significantly longer on the handgrip task following a positive mood induction. Finally, in study 4, high vs. low states of arousal were manipulated and self-regulation was assessed via pre- and post-manipulation Stroop performance. In study 4, Stroop performance improved slightly more across time points for those in the high arousal condition than for those in the low arousal condition. Therefore, across four studies, we failed to find a consistent pattern of results suggesting that positive moods restore cognitive resources.