Panel data from the South African National Income Dynamics Study, a nationally representative sample of households (years 2008, 2010 and 2012), were used to examine the longitudinal association between religious involvement and depression risk. Approximately 89.6-91.8% identified themselves as religiously affiliated, while 88.0-90.3% perceived religion to be important in South Africa during the observed study periods. A short-term association between religious involvement and significant depressive symptomatology was not detected, but logistic regression models that accounted for the clustering of repeated observations within participants indicated that, over time, those with religious affiliations (aOR 0.85, 95% CI 0.76-0.96) were at a lower risk than those not affiliated. Furthermore, individuals who reported religious activity as being important (aOR 0.81; 95% CI 0.73-0.91) were at less risk of significant depressive symptomatology over time than those rating religious activity as not important. Our study points to the potentially important role of religious involvement as an emerging area of investigation toward improving mental health at a population level in resource-limited settings.