We study how beneficiary identity affects investment decisions. Using Danish administrative data linking parents and children, we compare how the same adult simultaneously manages their own account and their child’s account. Investments made for children are safer: child accounts take less systematic and idiosyncratic risk, rely more on mutual funds, and are less likely to hold lottery-like stocks. Despite being less profitable, child portfolios earn higher risk-adjusted returns and exhibit less pronounced behavioral biases. These differences cannot be fully explained by taxes, monitoring, or investment horizon. Instead, our findings point to purpose-specific investing within households.