Is HELP doubtful debt likely to increase?

In February last year, when introducing legislation to change HELP repayment thresholds and rates, along with other HELP reforms, the government said that:

The fiscal challenge for the government is that HELP repayments have not kept pace with HELP lending growth. From 2010-11 to 2016-17, the level of new debt not expected to be repaid increased from 16 per cent to 25 per cent.

But unfortunately there is a disconnect between the stated purpose of the legislation and its likely effects. The thresholds and rate changes are unlikely to reduce doubtful debt , and indeed may increase it. Using  ATO tax data,  at the Grattan blog Will Mackey estimates that the new HELP repayment thresholds will produce slightly lower total annual student debt repayment than the previous thresholds.

The reason is that although some low and high income earners will repay more each year, most of the more numerous middle-income earners will repay less, producing a total repayment estimate of about $50 million a year less than if the previous thresholds and rates had been retained.

Read More »

Is HECS a tax?

My use of the word ‘lent’ in the chart below was disputed on Twitter, on the grounds that payments of HECS or HELP are tax levies. Although not spelled out in the Twitter comment, this point is often more than just a semantic one. It is part of a larger argument about how student/graduate-sourced funding of higher education should work.HELP total debt

One potential system for funding higher education is a graduate tax. The idea here is that graduates pay a proportion of their income above a threshold for a period of time after they complete their degree. With a graduate tax,  higher education is free but extra taxes are paid by financially successful graduates. The revenue could go into general government funds or be set to recover what the government thinks should be the student contribution to total higher education expenditure. But there are no specific charges for subjects or courses and there is no loan. The language of ‘lent’, ‘borrowed’ or ‘debt’ would not make sense conceptually or legally.

Veteran Labor (and current QUT) higher education policy adviser John Byron has argued for thinking about HECS in something like these terms:Read More »

1996 Cabinet papers: HECS ideas pursued and rejected

This year’s National Archives Cabinet papers release includes material related to the 1996 Budget changes to HECS.

The most important of these were replacing flat HECS rates with ‘differential HECS’, so that rates were based on subject disciplines, and lowering the HECS repayment thresholds, so that debtors began repaying earlier and repaid more at each income level (historical thresholds are at page 47 of this document).

The main submission released today does not have these final decisions, but outlines different views within the government and bureaucracy about how to proceed.

In public statements, differential HECS was justified by reference to both course costs and the expected future income of graduates. Neither Treasury nor Finance were keen on using future income. Finance noted, as others have since, that it varies a lot between graduates. Treasury thought that it was unfair that students in some disciplines would end up paying a much larger share of costs than others.

The idea that students should pay a share of course costs has regularly resurfaced since, most notably in the 2011 base funding review. But in the Cabinet submission we see an early version of why this idea has been consistently rejected. In the draft differential HECS rates based on cost recovery, law ends up in the cheapest band 1 (of 5; there were 3 in the end), while nursing is priced in the middle. Nurses paying more than lawyers is not an easy political sell. In the final announced decision, law was in the highest-priced band and nursing in the lowest-priced band.

The Cabinet submission also has a pricing rationale of expected demand that was not, so far as I know, used in public statements.  If demand already greatly exceeds supply, prospective students are less likely to be price sensitive. But politically that raises the possibility that other students would be price sensitive, which the government wanted to downplay.

Capping access to subsidised higher education to one degree or to a time period was considered; the Fraser government had tried something similar. In the final policy this was sort-of implemented by concentrating funding cuts on postgraduate coursework places. A fuller version of the idea arrived with the 7-year learning entitlement under Brendan Nelson, which started in 2005.  It was later abolished by Labor.

While mainly about course charges, the submission also mentions means-testing access to income-contingent loans by linking it income support thresholds. That would have been the most radical conceptual departure from current policy in the submission if it had been approved. There is also the Department of Finance’s usual attempt to get real interest on student debt, which wins the prize for the most-suggested change to student loans that has never been legislated.

One omission is interesting in light of subsequent policy concerns. Although there is mention of the fact that (by design) not all HECS debt will be repaid, there are no estimates of how significant this is. Perhaps some numbers were in other submissions we have not seen yet, and could explain the big reduction in repayment thresholds.

In 1996 government accounting conventions struggled with income contingent loans, as they still do. The submission mentions which changes will and won’t count towards the politically-salient Budget deficit. Because expected losses from student doubtful debt are not counted in the deficit/fiscal balance, this biases policy towards cutting direct grants to universities, which do count.

Fortunately, however, accounting conventions did let 1996 policymakers see that selling the HECS debt was a bad deal for taxpayers. Another Cabinet submission makes this clear. This possibility was raised again in 2013, with the same eventual conclusion.

As these submissions show, many ideas around HECS/HELP recur repeatedly over time.

 

Many graduates will repay less per year, and maybe less in total, under the new HELP thresholds

The Government had a rare higher education Senate victory this week, passing various amendments to the HELP loan scheme.

These include a series of changes to HELP repayment thresholds. Most of the political attention went to the initial repayment threshold, below which no repayment is required. It will drop from the current $52,000 to just under $46,000 in 2019-20. At that point, debtors will have to repay 1 per cent of their entire income.

In principle, I support this step in the direction of better aligning HELP with other government income support thresholds. This 2016 Grattan report supported a lower initial threshold.

Unfortunately, another key recommendation of that report, of consistent percentage increases between each threshold at which the repayment rate increases, was not strictly followed.

For most of the higher thresholds, each is 6 per cent higher than the one before it. But there is a 15 per cent gap between the first and second thresholds.

Combined with starting the repayment percentage at just 1 per cent,  this radically changes the nature of the threshold reform. It is not now something that we can assume will significantly alter HELP doubtful debt.

One intention of the original Grattan proposal was to move debtors more quickly through the repayment rates.  This was partly to recover more HELP debt before female full-time labour force participation drops from their late 20s, as shown in chart 1 below.

Chart 1: Female bachelor degree graduate labour force status, 2016

female labour

Read More »

Should HELP debts be capped?

The Council of Australian Postgraduate Associations has a paper out today opposing the government’s legislation that would cap HELP debt at $104,000 for most students, or $150,000 for students taking the high-end health courses. A media summary is here.

There is already a similar cap for students borrowing under the full-fee student HELP program, FEE-HELP.  The legislation would count HECS-HELP borrowing towards the total – this is the loan scheme used by the vast majority of domestic undergraduates paying student contributions for Commonwealth supported places.

Contrary to the impression given by the media article, adding in HECS-HELP would be prospective, only applying to debt accrued after 1 January 2020.

CAPA is right that a cap will tip more students into an at least partial up-front fee market. But given HELP’s overall design, this is not a knock-down argument against it.

As I have long argued, HELP has substantial costs in doubtful debt. And although the big debtors tend to be in fields with relatively good earnings prospects, the more someone owes the greater the risks for taxpayers – both because of the total amount owed, and because of the danger that the debtor will not spend long enough earning an income above the repayment threshold to repay in full.

Especially when it is hard to control costs by reducing repayment thresholds or abolishing the deceased estate HELP write-off, that creates pressure to reduce high-risk debt by limiting what students can borrow. The government can’t lose what it doesn’t lend. We are already seeing this in other parts of the system.

Loan caps can also usefully serve as soft fee caps in the fee-deregulated parts of the system. CAPA notes that even the current FEE-HELP cap isn’t enough for some JD courses. But if some universities are charging exorbitant fees for law courses that isn’t something public policy should encourage, especially when the fees end up being subsidised through HELP debt write-offs. There are plenty of much cheaper law courses out there.

There is room for debate about exactly what the cap should be. But $104,000 would allow most students to do an undergraduate degree and professional development postgraduate courses. In an amendment proposed by the government, debtors could replenish their cap by paying off some of their existing debt.

If we completely re-worked HELP, we could take a more actuarial approach to lending at higher levels – allowing it for low-risk borrowers, declining for high-risk borrowers. But that would be a radical conceptual change to HELP, with discipline, age, and gender all likely to significantly influence actuarial risk.

Staying within HELP’s current conceptual basis, we need general rules that support reasonable amounts of study but protect taxpayers from courses with excessive fees and from perpetual students. A cap on outstanding HELP loans is one such rule.

 

 

 

 

Has abolishing the discount for upfront payment of student contributions made a difference to upfront payment rates?

An article in The Conversation on incentives around student contribution payments made me wonder what difference the 2017 abolition of the 10 per cent discount for paying upfront was having.

The discount had a cost to taxpayers, since universities were compensated by the Department for upfront payment discounts. If students pay upfront, the risk of the remaining debt not being repaid is removed, as is the interest subsidy for the time that repaid debt is outstanding. The discount is only worthwhile from a taxpayer perspective if it induces upfront payments on a sufficient scale to reduce doubtful debt and interest subsidies by more than the cost of the compensation to universities.

At the time, I supported the decision to abolish the discount, because I doubted that it was generating a net financial benefit for taxpayers.

One reason for this is that various sources of evidence over the years suggested that upfront payments were coming from sources unlikely to be very sensitive to discounts. These include parents wanting their kids to be free of debt, employers, and scholarships. In the 2012 student finances survey, for example, 9 per cent of undergraduates reported receiving money to pay tuition fees. Read More »

A high HELP repayment threshold increases pressure to restrict or deny access to HELP

The Department of Education and Training’s 2016-17 annual report announced the first public use of a project to link up the ATO’s HELP repayment data with the Department’s enrolment data:

In 2016–17 the department worked with both the Australian Taxation Office and the Australian Government Actuary to create a database that links education courses with
income and occupation information. In 2018, the QILT website will publish graduate income data sourced from this database, which will inform students of the earning potential in their study area.

This will, of course, be very interesting, and could include income by university attended, as well as by study area. This has been done in the UK, although their QILT equivalent does not use it at this point.

But this data linking work is being done to better understand HELP debt and which factors affect repayment. It has potential uses well beyond student advice, uses that would fit a pattern of the government trying to reduce its risk of bad HELP debt.

The most obvious of these is to restrict access to courses with high rates of non-repayment. This was a feature of the VET Student Loans reform so that

loans are only being provided for courses that are closely aligned to the skills employers need in their workplace, thereby enhancing the opportunities for graduates to work, and to repay the money lent to them by taxpayers. [emphasis added]

Read More »