This paper presents the results of an empirical investigation of the economic consequences of divorce which uses the recently developed Longitudinal Administrative Database (LAD) constructed from Canadian tax files to track individuals leading up to, at the point of, and following marital breakup. This is the first study to use Canadian longitudinal data to address the question, while the LAD-derived sample is unequalled anywhere in terms of its size and representative nature. It is found that women's family income drops roughly one-half, and men's declines about one quarter in the first year of divorce, while using income-to-needs ratios to adjust for family size indicates a smallish rise in economic well-being for men, versus drops of just over 40 per cent for women. These initial income changes are followed by moderate rises for both men and women in the following years. Poverty rates jump from .16 to .43 for women in the year of divorce, and then drop off slowly, while for men they rise only a couple of points at the breakup. Women's labour market activity does not change in the pro-divorce period, but increases moderately at the split, while for men it is stable throughout. These figures do not include social assistance income, nor are child support payments deducted from the payer's side, but figures are presented to show that the principal results would not change qualitatively with their inclusion. Several implications of the results for the understanding of divorce, marriage, and the standard neoclassical economics model of the family are discussed.