This paper examines the effect that working for pay and volunteering has on the mental health of older Irish women and men. Data from four waves of The Irish Longitudinal Study of Ageing (TILDA) are used. Three measures that capture different dimensions of mental health are considered. Ordinary least squares regression estimates suggest that both working for pay and volunteering have statistically significant and substantially large positive effects on mental health. However, these effects are less well defined when fixed effects regression is used. The analysis also suggests that combining working for pay with volunteering is more beneficial in terms of mental health than either working for pay or volunteering on their own. That is, there is something "extra" from engaging in both activities. The estimates also suggest a possible trade-off between working for pay and volunteering in terms of mental health benefits. Volunteering may be a "good mental health substitute" for working for pay. The extent of this substitutability is particularly important amongst older people, since participation in paid employment decreases while volunteering increases in older age. Higher levels of volunteering may compensate for the mental health loss associated with lower levels of working for pay. If this is the case, policies that promote volunteering may be cost-effective if they result in higher levels of self-sufficiency amongst older people.