For the past several decades it has been proposed that birds show latitudinal variation in song complexity. How universal this variation may be and what factors generate it, however, are still largely unknown. Furthermore, while migration is confounded with latitude, migratory behaviour alone may also be associated with variation in song complexity. In this paper we review the literature to assess current ideas on how latitude and migratory behaviour may drive large-scale geographical patterns of song complexity. At least seven distinct hypotheses have been proposed in 29 studies of the topic. Four of these hypotheses posit that sexual selection pressures co-vary with latitude and/or migration, resulting in concordant changes in song. Other hypotheses suggest that mechanisms other than sexual selection, such as large-scale changes in environmental sound transmission properties, may be at play. Sixteen studies found support for increased song complexity with increased latitude and/or migration, whereas 13 did not. Relatively few studies exist on this topic, and methodological differences between them and variable definitions of 'complexity' make it difficult to determine whether results are comparable and concordant. At a minimum, it is possible to conclude there is no strong evidence that song complexity increases with latitude and/or migration in all birds. Future work should focus on examining multiple hypotheses at once to further advance our understanding of how latitude, migration and song complexity may or may not be related.