Research on whether music facilitates recall has been inconsistent and has lacked a theoretical basis. Three competing emotion-based theories yield differential predictions dependent on arousal levels, mood congruence, and functional relevance of information respectively. The aim of this study was to determine the most informative framework to understand the effect of emotion-inducing music on the short-term recall of information about narratives. Ninety-five participants (range = 18–58 years) were randomly allocated to one of four groups differentiated by the type of music presented to them, which was either happy ( n = 26), sad ( n = 19), fearful ( n = 25), or calm ( n = 25). Participants listened to music, followed by a positively or negatively emotionally-valenced narrative, and free recall of the narrative was tested approximately five minutes later. The results provided strongest support for the mood congruence theory in this context. After exposure to positive music, recall of positive information was significantly greater than recall of negative information. Mood regulation ability moderated this effect, with symmetrical mood congruence observed in participants with a tendency to repair their negative moods. Music may therefore offer an effective means of facilitating encoding of information when the mood induced by preceding music is congruent with the valence of information learnt. While the arousal and function theories may be more informative in other contexts (for instance, when music is played following learning or longer-term recall is tested), the current findings may help to clarify some of the inconsistencies previously observed in the research on music-facilitated recall.