Using harmonized micro data from 17 European countries, we investigate the causal impact of being socially active on old age cognition. To this effect, we employ nonparametric bounds-based partial identification methods that require fairly weak assumptions. We find that social activities have a strong positive effect on cognition, as the narrowest identification regions of the treatment effect exclude zero. This result is more likely to hold for females
than for males. We also provide evidence, again using bounds-based estimation, on the economic significance of social participation by examining the beneficial effect of higher cognition on households’ economic well-being.