This paper investigates the impact of occasional grandchild care on grandparent healthusing the pan-European dataset SHARE on elderly people. We find a small but statisticallysignificant positive correlation between grandchild care and physical health, cognitivefunctioning and mental health. These relationships are robust to controlling for sociodemographiccharacteristics, social engagement, and intergenerational support. Usingpropensity score matching, we can show that the findings are also robust to the underlyingfunctional form. To see whether they can be interpreted as causal, we exploit the panelcharacter of the data and use an instrumental variable approach. Applying these methods,the effects seem to be smaller than the OLS correlations suggested.