Simultaneity and selection in financial hardship and divorce

While the correlation between financial hardship and divorce is well-documented, the causality remains unclear: it is plausible that divorce causes hardship, that hardship encourages divorce, or that unobserved factors produce both outcomes. We specify a model that nests these possibilities and estimate it using the National Longitudinal Survey of Youth 1979. Structural estimates indicate divorce reduces the income/needs ratio in women's households by 0.35 standard deviations, though this is partially offset by apparent anticipatory labor supply responses. We also find a negative structural error correlation between divorce and income/needs ratios, but no evidence that a change in hardship causes divorce.