Social support is increasingly acknowledged as an important resource for promoting wellbeing. We test whether social support changes around retirement. We also examine whether social support moderates dynamics in mental wellbeing around retirement and consider both own and spouse's retirement. Using longitudinal data from Australia, we find little effect of own or spouse's retirement on social support. However, in fixed-effects models, dynamics in mental wellbeing are significantly different between those with low/high social support. Using pension eligibility as an instrument, we find that own retirement causally improves mental wellbeing for women (weaker evidence for men) and by a similar degree for those with low/high social support. We also estimate responses to life satisfaction and find evidence that spill-over benefits from spousal retirement are much larger for individuals with low social support.