We examine the causal effects of family income on child outcomes and households’ educational spending using panel data of children matched to their parents. Our identification strategy relies on the largely exogenous, discontinuous changes in the Child Allowance Policy in Japan that occurred between 2010 and 2012. We examine whether an exogenous variation in family income due to policy changes in the payment schedule has any causal effects on children’s cognitive outcomes and households’ educational spending. Our ordinary least squares (OLS) and first-differenced (FD) results show that, in most cases, family income is positively correlated with children’s cognitive outcomes and family’s educational investment. Our FD instrumental variable (FD-IV) results, using exogenous changes in child allowance payments as an instrument, show that family income does not have any causal impacts on child outcomes in the short run. This suggests that the positive income effects on cognitive outcomes in OLS and FD models are not causal effects. In comparison, we find some evidence of positive income effects on households’ educational spending. To examine the heterogeneous effects, we estimate FD-IV regressions for various population subgroups: those divided by parental education, income levels, children’s age, and gender. We find that family income does not have statistically significant impacts on children’s cognitive ability, whereas it has significant positive impacts on educational spending for high-income families and girls.