Joint taxation and the labour supply of married women: Evidence from the Canadian tax reform of 1988, by Thomas F. Crossley and Sung-Hee Jeon

Fiscal Studies, vol.28, 2007, no.3, pp.343-365

A change to the taxation system in Canada in 1988 has enabled Crossley and Jeon to conduct a natural experiment to test the effect of changes in marginal tax rates on labour market participation.

The Canadian tax system treats husband and wife jointly. Before the change, the husband received a tax allowance for the dependent spouse which reduced as the spouse’s earnings rose. This meant that for each additional dollar earned by the woman, tax was paid on an additional dollar of the husband’s income at the husband’s highest current rate. After the change, deduction of the additional dollar of the husband’s income was at a flat rate of 17%: the lowest tax rate. This meant that a low-earning husband experienced no change, whereas a high-earning husband experienced a change of 12% in their marginal tax rate.

The researchers chose two samples of couples similar in all respects except that the husbands in one sample earned $24,568 (standard deviation $37) and the husbands in the other $53,273 (standard deviation $195).

A simulation of average tax rates showed that for the low earners the tax reform created almost no change in net incomes, whereas for the high earners the effective marginal tax rate on the woman’s additional income was considerably reduced, and that the difference was greatest for women earning $5,000. This suggested that the effect of the change on labour market participation was likely to show up most clearly in part-time employment rates.

The researchers found that labour market participation amongst wives of high-earning husbands increased by 10%, which they rightly suggest is both ‘economically and statistically significant’ (p.357); and, as they expected, they found that ‘the principal effect of the reform was to increase incentives for part-time work’ (p.357). Amongst the wives of low-earning husbands there was no significant change.

The conclusion which the researchers draw is that individualising the taxation system would increase labour supply amongst women. This is a valid conclusion to draw for systems in which the jointness of the system imposes high marginal tax rates on spouses of earning husbands. The researchers also rightly suggest that their research should be ‘of interest …. where the unit of taxation is the individual, but where recent trends towards means-tested benefits and tax credits on family income have increased the ‘jointness’ of the system’ (p.362) – as in the UK.

They might have drawn a further and more general conclusion: that lowering marginal tax rates increases labour market participation. This is an important result which should have been highlighted. The study of the effects of tax and benefits changes on labour market participation is still in its infancy, and this article contributes significantly to the field because it shows that increasing marginal tax rates by means-testing benefits or by withdrawing tax credits will decrease labour market participation, and that decreasing marginal tax rates by increasing universal benefits such as Child Benefit and by replacing means tested benefits and tax credits by universal benefits will increase labour market participation.

It is a pleasure to read an article which contributes such unambiguous results to a field with direct application to the reduction of poverty and the efficiency of the economy.