Category Archives: Higher education

The science bubble finally shows signs of deflating

Since 2009, demand for science courses has been growing strongly. This is leading to a serious labour market over-supply, and so I believe there needs to be a correction. Unfortunately the information flows to prospective students for science contain a lot of misleading signals from STEM boosters, who persist in high-profile, but in my view incorrect, claims that this area of education needs encouragement.

The February 2015 applications statistics show some tentative signs that the market is adjusting to science realities. Through the tertiary admissions centres, there has been a 3 per cent decline in science applications. In reality, it is probably a larger drop than this, as I understand that a course reclassification that led to a big drop in environmental studies applications should have boosted science. Against this trend, however, there has been a 5 per cent increase in direct applications for science courses (for non-Year 12 applicants, there is a trend towards applying directly rather than through tertiary admission centres).

Feb demand

The biggest drop in demand if we don’t consider the environmental studies reclassification has been for education courses, down 9 per cent through tertiary admissions centres. This is probably because with a clear occupational outcome in teaching negative labour market information is transmitted much more effectively. Although this is a sensible adjustment, Kim Carr will still be on the warpath about nearly 900 education course offers to applicants with ATARs below 50.

Total applicants (as opposed to applications) are up 1 per on 2014, indicating a market that is essentially stable.

Pure research at universites increasing absolutely, but declining relatively

I have an article in The Conversation this morning arguing that policy over the last quarter century has made public universities more public than they would otherwise have been. Without external pressures, academics and universities would focus on a narrower range of objectives than governments and public opinion would prefer.

In response, Gavin Moodie observes

I am concerned at the diminution in the importance of pure basic research in Australian universities, which is now very much a minority activity. This is particularly worrying since while there are many other bodies established to conduct applied research, there is no other institution to conduct pure basic research.

While the share of all research that is ‘pure’ (definitions here) has gone down (see yesterday’s post), looking at the absolute spending gives a somewhat different picture. Spending on pure research doubled in real terms between 1992 and 2012, but this is a much lower growth rate than for the more applied types of research.

pure research
Source: ABS

Although I don’t think the trend is as negative as Gavin perhaps does, I have some sympathy for his broader point. While public universities can and should do more than academics are inclined to, their comparative advantage compared to other institutions is likely to be at the more pure end of research. If the government wants more applied research, it should probably steer much of that funding to more specialised organisations with clear commercialisation/practical problem solving goals.

Research commercialisation policy déjà vu

Australian is known as a nation that conducts high-quality basic research, and the Government wishes to maintain this reputation. … The Government considers, however, that a greater proportion of such research should be in fields that have the potential to improve the nation’s competitive position. …. We have a poor record in translating the results of basic research into effective application.

- John Dawkins, Higher Education: a policy discussion paper, December 1987

Overall, the Australian research sector is highly productive, internationally connected, and recognised globally for high quality research. … Despite this strong performance in producing excellent research, our ability to translate publicly funded research into commercial outcomes lags behind comparable countries.

- Christopher Pyne and Ian Macfarlane, announcement of a new research commercialisation strategy, May 2015

There are various perpetual critiques of Australian higher education, and the idea that we don’t do very well in commercialising research is one of them. As the ABS figures reported in the chart below show, there has been a significant shift in research towards the applied end of the spectrum. But it remains the case that Australian businesses infrequently report universities as a major direct source of innovation.

Research type

Perhaps some good will come of the latest round of policy initiatives, but I doubt it will be enough to stop similar analyses being offered 25 years from now. Universities just aren’t particularly well suited to producing commercially-oriented research. They attract people whose main interest is curiosity-driven research, not people who want to make money. Academics are much more likely to apply for grants that don’t involve collaboration with industry than those that do, even though industry grants are easier to get. The university status system is oriented around publications in theoretically-inclined high-prestige international journals, not solving the practical problems of Australian industry. While incentives for particular types of research activity matter at the margins, they are unlikely to change university culture.

No need to spend more than $2 billion promoting STEM subjects

Unfortunately Labor’s promise to write-off the HELP debt of 100,000 science, technology, engineering and maths graduates suggests that they have learned little from their previous mistakes in this area. Following a 2007 election promise, to boost science and maths they cut student contributions and introduced a HECS-HELP benefit, under which around $1,700 a year of HELP debt is written off if graduates work in specified occupations related to their degree.

The cut in student contributions was strongly promoted, and there has been on-going advocacy for STEM disciplines from the Chief Scientist, Professor Ian Chubb. There has been a big increase in science demand and domestic undergraduate enrolments – up 35 per cent between 2008 and 2013, or more than 21,000 full-time equivalent places. By far the largest increase has been in the biological sciences, which made up nearly 40 per cent of the total. Engineering, which did not have a cut in student contributions, increased by 32 per cent over the same period, with more than 8,000 additional full-time equivalent places. Science demand kept growing in 2013 and 2014, despite student contributions being put back up again.

As I have long argued, there has never been any evidence that we need a significant boost in bachelor-level science graduates. The latest employment data confirms that the surge of completions in science is only leading to serious un- and under-employment among science graduates, who have been hit especially hard in the general graduate employment downturn. So it is hard to argue that there is any general problem to solve in the first place.

sci take 3

Possibly there are still some niche employment issues in say secondary maths and science teachers – although they have fallen off the skills shortage list. But a promise to write off a few tens of thousands of dollars in student debt is unlikely to change how many people see a teaching career. Even for financially motivated students, the cost of university is not high relative to career earnings for full-time professionals. Perhaps the main thing that will drive graduates to teaching is that they may have few other options, thanks to the over-supply of graduates.

Course and career choices are primarily about interests and aptitudes, with long-term earnings a factor. These can be influenced – people have multiple interests and are not necessarily aware of all the suitable course and career opportunities. But this influence can be achieved without writing off more than $2 billion in student debt (we get similar numbers to the government). A few million dollars in marketing expenditure would probably have the same effect, if this was a desirable outcome – which it is not. Labor’s latest policy is, unfortunately, only likely to to encourage people to make choices that put them at high employment risk.

The Budget significantly understates likely student-driven higher education spending

Compared to last year, for higher education the 2015-16 Budget is uneventful. But the problem with the Budget papers is that they assume that several things that are unlikely to occur will happen. In particular, they assume that Commonwealth contributions will be cut, domestic undergraduate student fees will be deregulated, and that students at private universities and non-university higher education providers will enter the publicly-funded system. While I think only fee deregulation is completely impossible with the current Senate, on the government’s current legislative strategy it is not clear how any of these proposals will become law.

The government’s reform package not passing the Senate will have both positive and negative effects on the Budget forward estimates. The reform package had two significant costs: added subsidies from more students becoming eligible for Commonwealth-supported places, and added doubtful HELP debt from fee deregulation. From this perspective, the Budget is over-stating likely higher education spending.

However, the Budget is also assuming that per student subsidies will be lower than in fact they will be, as the government is still pursuing per student funding cuts averaging 20 per cent. It is also, I think, assuming an efficiency dividend (I can’t see it mentioned but I have not seen an announcement that it is no longer being pursued), and a change to the grant indexation system (although this change would not deliver any significant short term savings). From this perspective, the Budget is under-stating likely higher education spending.

Calculating the net effects of these changes is difficult on the information presented in the Budget. We have used varying scenarios about public university-only student numbers, indexation and the efficiency dividend. Based on these scenarios, the forward estimates understate likely government spending on the Commonwealth Grant Scheme by between $2.5 billion and $3.7 billion. In other words, universities will receive between $2.5 billion and $3.7 billion more than the Budget papers suggest.

On currently available information, we cannot provide a sensible revision to projected HELP costs. Compared to information on HELP published late last year, HELP’s costs have been revised down. The 2015-16 Budget papers suggest that VET FEE-HELP borrowers will shrink from 225,500 to 128,000 between 2014-15 and 2015-16. If so, that is a remarkable response to current efforts to clean up the vocational education industry. They are also predicting fewer HECS-HELP and FEE-HELP borrowers.

The ‘performance indicator’ for the proportion of new HELP loans that are expected not to be repaid has been reduced by two percentage points across the comparable forward estimates years, so that by 2017-18 it is anticipated to be 21 per cent instead of 23 per cent. While an improvement would flow from fewer VET FEE-HELP borrowers, these proportions still look to be on the optimistic side.

Corrected post: Most HELP debtors still owe less than $20,000, but big debts are increasing

The ATO personal income statistics provide information on how much debt students and former students are holding. Although debt levels are increasing, the vast majority of HELP debtors (71 per cent in 2012-13) still owe $20,000 or less.

The modal amount of debt is below $10,000, with more than 40 per cent in this category in all years. However, this is a poor guide to what the typical student will end up borrowing, as it includes people who are early in the borrowing phase and towards the end of their repayment phase (plus probably quite a few people who did not borrow much in the times when HECS was much cheaper, but have not repaid).

HELP debt levels

The share of HELP debtors owing more than $40,000 has increased from 3.5 per cent to 5.6 per cent between 2010-11 and 2012-13, or just over 100,000 people in 2012-13.

Fewer people are repaying their HELP debt

A few months ago I argued that flat graduate incomes and an initial threshold that was indexed to average weekly earnings was going to mean fewer graduates making a repayment. The 2012-13 taxation statistics that came out today shows that this is already a problem. The total number of people who made a HELP repayment that year dropped by over 2,000 compared to 2011-12, while the total number of debtors increased by more than 142,000.

Help debtors and repayments

The surge in enrolments since 2009 means that it is inevitable that a lower percentage of debtors will make a repayment, since it takes time for people to finish their courses and enter the workforce. But this is only the second time since HECS/HELP started in 1989 that the absolute number of people making a repayment went down, other than due to a deliberate policy shift (increasing the threshold for 2004-05). Total repayments did go up by $24 million, or about 1.6%.

The falling number of debtors making a repayment highlights again the need for a lower threshold and measures to reduce the manipulation of HELP repayment income.

Would restoring or increasing discounts for up-front student contribution payments improve’s HELP’s finances?

The Australian this morning is giving a lot of attention to this paper by Neil Warren and Richard Highfield on HELP repayment.

It is a good paper, analysing ‘bunching’ of HELP debtor incomes below important repayment thresholds. The significance of this (for both the government and the debtor) is that if the debtor’s income crosses one of the thresholds they have to pay a higher percentage of all their income in repayment.

This has turned out to be a valuable aspect of HELP’s repayment system compared to those of other countries, where students pay at the margins (eg 9% of income above the threshold). In Australia, anyone who consistently earns above the initial threshold (about $53K this year) is likely to eventually repay all their debt, as they will pay back $2,000+ each year. But in other countries, earnings slightly above the threshold result in only very small repayments. This is one reason that, bad as Australia’s student doubtful debt figures are, they are not as bad as England’s.

However, because crossing a threshold has a high cash cost there are temptations to keep income below it. I have to admit that long ago I considered turning down a pay increase because I thought it would reduce my take-home pay more than I can afford, as someone who at the time had a very tight financial situation. In the end, what was to my boss a generous increase boosted my take-home pay by some token amount (and of course sped up my HECS repayment). What Highfield and Warren present is evidence that some people are using deductions for education, work and charity to bring their taxable income below thresholds.

Their policy suggestion is to change the definition of HELP repayment income so that some or all of these deductions are not included. This is a strategy the government has adopted before to tackle negative gearing and fringe benefits. I need to think more about the education and work deductions, but support removing charitable deductions (the taxpayer would still get the deduction for income tax purposes).

Another suggestion is to revisit the issue of discounts for paying student contributions upfront. The discount was cut from 25 per cent to 20 per cent in 2005, and then to 10 per cent in 2012, with a bill to eliminate it entirely stalled in the Senate. Money never lent will never cause repayment problems, so this has an obvious attraction – if the savings from the reduced lending outweigh the discount’s cost.

However, I am not convinced that the discount is having a big enough effect in encouraging up-front payment to warrant keeping it. The proportion of students paying upfront is in long-term decline, as seen in the chart below. The halving of the discount in 2012 did cause the biggest year-to-year decline in upfront payment in this time series, but it really only sped the trend up, rather than marking an obvious major turning point. We are talking perhaps about 1 to 2 extra percentage points of students paying upfront if they had a stronger financial incentive.

HECS-HELP discount

Cutting the discount has been based in part on the assumption that many of the people paying upfront are not principally motivated by a financial calculation. They are parents paying for their kids, employers paying for their staff, or individuals who don’t want to hold debt even on very low interest rates. Overall, the government is probably better off trying to collect the full student contribution amount via HELP than giving windfall gains to people who will pay upfront anyway.

While it has always made sense to borrow under HECS or HELP, due to the very favourable lending conditions, I am not entirely sure why we are seeing such a strong trend. An increasing share of low SES students might explain some of it, but I suspect there is more going on.

Public opinion on special admission standards for Indigenous university applicants

In the United States, racial preference in university admissions is a highly controversial issue. But in Australia universities have long had special admissions programs for Indigenous applicants, with little obvious controversy. So far as I am aware the latest ANU Poll, on Indigenous affairs, is the first to ask the general public what they think.

As the chart below shows, a small majority of respondents, 54%, favoured special programs and admission standards for Aboriginal people. This was lower than support for governments helping Aboriginal people find employment (69%) or who think the private sector should do more to employ Aboriginal people (66%).

It’s hard to explore the reasons for these results from within this survey. However there are common ideas around minimum entry standards (as seen in the annual January low ATAR debate), and using ranked prior academic performance to allocate scarce places, that would influence views on university admission more than staff hiring practices.

Indigenous

Years to repay student debt as a way of setting student contributions

In The Australian this morning, Melbourne University VC (and my former boss) Glyn Davis has an op-ed saying:

As the 2011 Lomax-Smith report made clear, there is no consistent principle guiding public and private contribution to university study. A law or economics student pays about 83 per cent of the cost of their education while students in other disciplines enjoy a much larger public subsidy. This is not fair.

I have no dispute that the current system of setting overall funding rates and student contributions is untidy. It is the accumulated result of ad hoc decisions going back 25 years. There has been no careful empirical work to ensure that either funding rates or student payments reflect clear funding principles.

That said, I do not believe that the current principles for setting student contributions are inherently ‘unfair’. When different student contributions for different courses replaced previously flat HECS rates in 1997, the stated rationale was the new fees would reflect a mix of private benefit and course costs.* The higher the private benefit, the higher the student charge, with some but not primary acknowledgement of course cost. If course cost is not the major factor in setting the fee, it is not clear that calculations based on course cost denominators tell us much. The implied denominator for a private benefits approach is future earnings or some other measure of private advantage.

The reason law and economics student pay most of their course costs is their high private benefits/low course costs combination. But the idea behind differential HECS was to get a more even outcome on a student costs/graduate private benefits calculation.

I think one interesting way to look at this is to calculate how many years it takes graduates in different disciplines to repay their HELP debt. Based on 2011 census data, we think it would take a median male graduate about 10 years to repay their HELP debt. The chart below shows that even though law students pay high fees, they are estimated to take less time than average to repay, because their salaries are higher than average. Business graduates are estimated to take 11 years, a little above average.

median years reapy

From an egalitarian perspective, something like this system means that graduates across the disciplines put in more similar work effort to repay their debts than a system in which their subsidies are a more consistent percentage of course costs. For example, if we had a 50:50 public/private funding system law graduates would get higher subsidies and take less time to repay, despite already being on the shorter side of the median. Science students would get lower subsidies and take more time to repay, despite already having an above-median repayment period. It was counter-intuitive outcomes like these that sunk the Lomax-Smith recommendation of a 60:40 public/private split.

From a public benefit subsidy perspective, the current system also has some merit. In this analysis, students make some calculation of private benefits (financial and non-financial) from a course and compare these with the costs (also financial and non-financial). If we want to encourage more people to take education generally or particular courses via subsidies we can alter private financial benefits to make them more attractive. But if private benefits are already high, we don’t need subsidies at all or to the same extent. This means that we can charge students in high private financial benefits courses more, regardless of cost.

To reiterate, this is not a defence of the detail of the current system. But there are good reasons not to be too worried about what numbers we get from a student contribution/total course costs calculation, and to look at other rationales for setting student charges.

—–
* I am using ‘course costs’ a bit loosely here; the calculations are actually student contribution/Commonwealth-supported student funding rate. The funding rate is typically around the actual costs, but this varies between disciplines and institutions.