Tax and Skills

Earlier this year, the Henry tax review (aka AFTS) commissioned me to write a paper on the tax treatment of education. They’ve now posted it on their website. Here’s the executive summary (click on the title for the full paper).

The Impact of the Tax-Transfer System on Education and Skills in Australia
Andrew Leigh

Human capital theory suggests that in considering whether to undertake education, individuals weigh up the costs and benefits. On the cost side, this includes direct costs (such as tuition and textbooks) and indirect costs (which largely take the form of foregone earnings). On the benefit side, individuals should consider the additional earnings that will accrue from having higher educational qualifications.

The tax-transfer system can affect both costs and benefits of higher education. For example, more generous student income support should increase educational participation rates, while more progressive taxes should reduce educational participation rates. However, the magnitude of these impacts is an empirical question.

With some exceptions, the literature on taxes and educational participation generally concludes that taxation can have a substantial impact on human capital acquisition. However, one of the features about the empirical studies on taxation and human capital is that it consists almost exclusively of simulation studies, which model behaviour according to a set of parameters that are drawn from previous studies. A limitation of these studies is that they generally assume no uncertainty and full information, which may not hold in practice. For example, one survey suggests that a majority of respondents to one United States survey wrongly believed that their country’s income tax system was regressive. It is possible that misinformation may be even higher among the cohort who are choosing whether or not to stay in school, attend TAFE, or complete a university degree.

The literature on subsidies and human capital tends to consist largely of natural experiment studies, which have the advantage that they are estimated from real-world policy changes. These studies tend to suggest that subsidies can affect participation, but that the effects are larger for low-income students, and that the impact of grants is larger than the impact of loans. Since educational subsidies are generally marketed directly to young people, it is not unreasonable to think changes in subsidies may be more salient than the degree of progressivity in the taxation system.

In trying to set optimal education taxes and subsidies, it is useful to have regard to the literature on social returns to education. This suggests that social returns are present, particularly in the areas of crime (from higher school completion rates) and productivity (from higher university completion rates). However, the best estimates of the size of social returns suggest that in the main, they should not be a key driver of policy. By contrast, there is robust evidence that private returns to education are large and significant. Completing year 12 raises gross income by 30 percent (relative to completing year 10) and completing a bachelor’s degree raises gross earnings by 49 percent (relative to completing year 12). Taking taxes and transfers into account lowers these estimates by 11-15 percent, but the private gain from human capital acquisition is still substantial.

What is the cross-country relationship between educational participation and taxes and subsidies? To test this, I look across 27 developed nations, to see whether those with higher public subsidies to education, or less progressive taxes, have higher rates of participation in tertiary education. Contrary to theoretical predictions, I find no significant evidence that more generous subsidies or lower tax rates on the rich have the effect of raising educational participation. One possible interpretation of this result is that the cross-country measure of participation is poorly measured, or confounded by an omitted variable that affects both participation and subsidies/taxes. Another plausible explanation is that, in aggregate, taxes and subsidies have a relatively small impact on educational participation.

Australia’s leading higher education blogger, Andrew Norton, has written two posts discussing the paper, which can be found here and here.

This entry was posted in Economics of Education, Tax. Bookmark the permalink.

1 Response to Tax and Skills

  1. Hersh Sahai says:

    That taxes and subsidies have a small impact on educational participation seems unintuitive. I favour your first conclusion, as more variables are likely to be introduced to the model as each country’s data is accumulated.

    If anything, this kind of result should encourage policy makers to perform randomised experiments within communities where the samples are sufficiently representative of the population. The reason being, to eliminate cross cultural/geo-political variables. This reasoning would favour selecting samples within the ‘same’ country for treatment and control groups.

    Doing so would surely give us more robust conclusions about the efficacy of education policies. My hope is that it is only a matter of time before a government organisation stumbles over the same conclusion.

    Unfortunately, scientific reasoning doesn’t win votes in the political arena… atleast not yet!

Comments are closed.