The Barro-Becker model of fertility has three controversial predictions: (i) fertility and schooling are independent of family income; (ii) children are a net financial burden to society; and (iii) individual consumption is negatively associated to individual income. We show that introducing credit frictions into the model helps overturn these predictions. In particular, a negative relationship between fertility and individual wage income can be obtained when the intertemporal elasticity of substitution is larger than one. The credit constrained model can also explain the quantity-quality trade-off: individuals with higher wage income choose more schooling and fewer children.