Background: Numerous studies have demonstrated an association between social support and health, almost regardless of how social support and health have been conceptualised or measured. Even so, the issue of causality has not yet been sufficiently addressed. This issue is particularly challenging for mental health problems such as depressive symptoms. The aim of the present study is to longitudinally assess structural and functional aspects of social support in relation to depressive symptoms in men and women, through a series of competing causal models that, in contrast to many other statistical methods, allow for bi-directional effects. Methods: Questionnaire data from the Northern Swedish Cohort (n = 1001) were utilised for the years 1995 (age 30) and 2007 (age 42). Associations were analysed by means of gender-specific structural equation modelling, with structural and functional support modelled separately. Results: Both structural and functional support were associated with depressive symptoms at ages 30 and 42, for men and women alike. A higher level of support, particularly functional support, was associated with a decrease in depressive symptoms over time among men. Among women, there were bi-directional effects of social support and depressive symptoms over time. Conclusion: Concerning social support and health, the social causation hypothesis seems relevant for men whereas, for women, the associations appear to be more complex. We conclude that preventive and health promoting work may need to consider that the presence of depressive symptoms in itself impedes on women’s capability to increase their levels of social support.