.net 98 degree

1 Introduction

Public funding of higher education is an important policy decision for government. Although a number of factors will affect this decision, key among them is the desire to maximise the return to society from investment in higher education. Returns derive from several sources, perhaps the most important of which is the increased productivity (and hence earnings) of graduates. These returns accrue primarily to graduates. However, additional benefits to society as a whole may also occur, for example because it leads to increased taxation revenue, reduced welfare dependence and a lower propensity to commit crime. Associated with private and societal benefits are calculations of private and societal rates of return. While, from an efficiency perspective, the government is properly guided by the societal return in determining the level of support, the government is also concerned about the effect of policy on the government budget at both the current point in time and into the future. In this paper, we focus on the effect of government higher education spending on its own "balance sheet".

The framework adopted considers, from a government perspective, the revenue and cost streams in each year that are associated with university education. Costs are primarily determined by each year's funding decisions, while revenue depends on the taxation revenue derived from the higher earnings of graduates and on the repayments of Higher Education Contribution Scheme (HECS) debts. However, while current-year costs relate to the current cohort of students, current-year revenues depend on the current earnings of previous student cohorts. As a corollary, the revenue benefits of current funding arrangements are not derived in the funding year (for the most part), but will be distributed over future years. This suggests two approaches. In the first, which we term the "current.net benefit" approach, current costs are compared to current returns, where the returns are related to degrees earned in the past. In the second approach, which we term the "cohort.net benefit" approach, the.net contribution of a specific s tudent cohort is considered.

The current.net benefit is evaluated by considering, in each financial year since 1989-90, the contemporaneous cost and revenue streams associated with university study. Costs are assumed to comprise the current-year expenditure on universities (both current and capital accounts) and total HECS debts acquired by students in the current year. Benefits are assumed to comprise the higher taxes paid on income and consumption that result from the increased earnings facilitated by higher education, as well as repayments of HECS debts. Estimates of the total increase in earnings due to higher education are based on the level of employment of university graduates, the actual earnings of graduates compared with non-graduates by discipline and an assessment of the value of the difference in innate ability between the graduates and non-graduates.

An advantage of this approach is that benefits and costs are contemporaneous and no judgements concerning the time value of money are required. The main limitation is that the results cannot be attached to a particular cohort or set of policies, since revenue in a particular year reflects the cumulative effects of investments in all previous student cohorts (who are still in the labour force). Nevertheless, the results can be interpreted as the current budget position with respect to higher education, and that is the motivation for this part of the analysis.

Measurement of the cohort.net revenue benefit involves consideration of the costs and revenues attached to a particular cohort of students: those beginning higher education in 1999. As with the estimation of the current.net benefit, the university education costs to the government of a specific student cohort are compared to the revenue stream remitted in additional tax payments and HECS repayments. However, revenue and cost calculations relate to all years (the entire "life") of a particular cohort, as opposed to all student cohorts in a particular year. It follows that we can interpret the higher education cost of a cohort as an investment by the government that yields a return in the form of higher taxation revenues in the future. Consistent with this interpretation, the comparison of the costs and benefits is undertaken by calculating the (internal) rate of return to the government.

In addition to producing aggregate current and cohort.net benefit estimates, separate estimates are also produced for each of eight broad fields of study. The rationale for this is that it would seem inappropriate to treat all university education as homogeneous - a stance supported by our results - and our approach facilitates comparisons across the fields of study.

Although the costs and benefits taken into account in our analysis are perhaps the most significant ones, we should note that they do not constitute all costs and benefits to the government that are associated with higher education. We do not consider potential external benefits of higher education (for example, a lower crime rate), potential indirect costs and benefits of higher education (for example, lower per capita health care costs to the government), nor the potential implications of higher education attainment levels for the welfare system (which may be substantial). For this reason, we suspect our estimates may understate the true.net benefit to the government of higher education. The available data also limit the ability to accurately measure the costs and benefits we do include, and to a large extent dictate the methodology we use. We consequently discuss our methods with reference to the data to be used.

Our attempt to ascertain the fiscal position of the federal government with respect to university education, taking into account the earnings effects of higher education, appears to be somewhat novel. Studies of the costs and benefits of higher education typically focus on private costs and benefits--examples in Australia include Blandy and Goldsworthy (1975), Chapman (1977), Miller (1982, 1984), Chapman and Chia (1989), Chia (1991), Maglen (1994) and Chapman and Salvage (1997). Three exceptions are Cabalu et al (2000), Johnson and Lloyd (2000) and Larkins (2001). Cabalu et al (2000) attempt to measure the importance of higher education to the Australian economy in terms of the "direct income and employment generated" by the higher education sector, the "enhancement of the nation's human capital through the education of university graduates" and the "creation of wealth through the spillover effects of research". However, this paper makes no attempt to isolate the costs and benefits of the sector to the Austr alian government. Larkins (2001) examines both the private and social rates of return to Australian bachelor and research degrees in the sciences, humanities and social sciences, and in the process also estimates the cost and revenue to the federal government of each degree type. Larkins (2001) contribution in this area is similar to the cohort.net benefit component of our paper, but he considers the.net benefit to the government assuming a 4 or 5 percent discount rate, rather than calculating an internal rate of return. Johnson and Lloyd (2000) present estimates from the RED99 model developed by NATSEM, estimating the internal rate of return to the government from a representative student--analogous to the cohort.net benefit estimates obtained in this paper. Johnson and Lloyd report a private return of 13.5 per cent for a prospective science graduate and a return of 9.9 percent to the government.

As with the Australian research, the international literature similarly focuses primarily on private returns. However, the OECD (1998) report "fiscal returns" for seven countries, including Australia, where a "fiscal return" has a similar interpretation to our cohort.net benefit. They report a 10 percent fiscal return for Australia in 1995.

In this paper we extend both the Johnson and Lloyd (2000) and the OECD (1998) analyses by estimating returns for each of eight discipline areas, and by bringing the estimates up-to-date using contemporary (2001) data which incorporates the effects of the policy changes that occurred in the late 1990s.

2 Current.net benefit--Methods

2.1 Costs