FUNDING HIGHER EDUCATION IN AN AGE OF EXPANSION 

Submission to the Review of Higher Education and Policy

by

Nicholas Barr and Iain Crawford

London School of Economics and Political Science

July 1997

Authors' note. As British commentators, we would not presume to make specific recommendations about reform in Australia. However, the systems of higher education in the two countries grow from common roots, and there are significant strategic similarities in the problems they face. We hope, therefore, that this paper (which was a submission to the UK National Committee of Inquiry into Higher Education) might assist in the formulation of a reform strategy, though clearly the implementation of that strategy in Australia would in parts be different.

FUNDING HIGHER EDUCATION IN AN AGE OF EXPANSION

Nicholas Barr and Iain Crawford

1 The Strategy

1. The central fact of British higher education is the increase in the participation rate from an élite 5 per cent in the early 1960s to a mass 30 per cent by the mid 1990s. So far as funding is concerned, this expansion has three principal implications (discussed more fully shortly):

• the need for a wide-ranging system of student loans;
• flexibility to allow price differentiation by universities to reflect their differential costs; and
• a move away from central planning of higher education.

2. This paper explains why — in the absence of a return to an élite system — these implications are inexorable, and offers practical solutions for dealing with them. It is taken as common ground that higher education is important — important for economic growth, important for the transmission of cultural values, important in sustaining individual freedom. The policies discussed here are all instrumental in the sense that they are concerned mainly with the how of higher education finance.

3. Implication 1: the need for a wide-ranging loan system. Public funding of a high-quality system is possible — albeit regressive — for a 5 per cent system. It is not possible for a mass system. A mass system therefore requires public funding to be supplemented on a significant scale by private funding.

4. Private funds can derive from three (and only three) sources: the student’s family; his/her earnings while a student; and his/her future earnings.

• Family resources. Though family resources will always be part of the picture, it is well known (Barr and Low 1988) that the system of parental contributions works badly. It is, in addition, grossly inequitable to rely on family resources as a major source of finance.
• Student earnings whilst a student. Students already support themselves partially through paid work during their student years. But such activities have to compete with study time and leisure activities. This approach, though useful, is limited.
• Students’ future earnings. It follows that in a mass system the only potential funding source which is (a) large and (b) not grossly inequitable is a system of loans which allows students to borrow against their future earnings. The source of private funds is student repayments. These will accrue slowly if students borrow from government (since in the early years, government will be lending large amounts but receiving only a small flow of repayments). If students borrow from the private sector, however, the injection of private funds into higher education is immediate and substantial.

5. Objectives. Specifically, what is needed is a system of loans which:

• Promotes access, both for equity reasons and for efficiency, to minimise the waste of talent in an increasingly competitive global market. Access has at least two major dimensions: students should have an adequate standard of living while students; and student loans should not deter applicants.
• Restores quality, which has been eroded by particularly sharp expansion (from 14 per cent to 30 per cent participation) since 1989 with little increase in total resources going to universities. An increase in resources for universities is an essential component in restoring quality. Another is the development of a system which allows universities to respond to student demand.
• Contains taxpayer cost. This is an objective for several reasons. The taxpayer cannot shoulder the major burden of financing a mass system. It is not realistic to expect significant increases in public spending on higher education under any party, not least because of other educational demands, for example pre-school education (see National Education Commission 1993), and the prospective claims of an ageing population.
• Liberates private funds immediately. As a consequence of the previous objective, the issue is not just to bring in private funding through students’ eventual repayments, but to bring it in fast by devising a loan mechanism which allows students to borrow from the private sector.

6. So far as funding is concerned, therefore, the central consequence of the move to a mass system is the inescapable need to draw in private funding which, in practice, means a system of student loans.

7. Implication 2: price differentiation. With an élite system it was possible, at least as a polite myth, to assume that all universities were of equal quality (this was known as parity of esteem), that degrees were worth the same whichever university conferred them, and hence that universities could, broadly, be funded equally. With a mass system this myth is no longer sustainable. The characteristics, the quality and the costs of different degrees at different institutions will vary much more widely than hitherto. To illustrate the argument, assume that there is a strong positive correlation between quality and cost. Assume further, for simplicity, that there are three types of institution: internationally competitive, with high costs (not least because of the need to pay internationally competitive salaries); average institutions with average costs; and institutions with low costs (not least because staff costs do not cover time spent on research). The problem, if all universities receive broadly equal funding is the following.

• If all institutions are funded at the level of the internationally competitive universities, the result is inefficient, wasteful, unaffordable and politically unsustainable.
• If institutions are all equally funded at an average level, the quality of the internationally competitive institutions cannot be sustained. Lower quality institutions, in contrast, are overfunded which is both wasteful and gives them no incentive to improve.
• If institutions are all funded equally at a low level, the best universities, a fortiori, are disadvantaged, and the average quality also suffers.

8. The argument, in short, is that the move to a mass system leads inevitably to the need for a loan system which achieves the objectives discussed above; and, given the increasing diversity (much of it desirable) in what institutions offer, those loans need to be compatible with a funding régime in which institutions can charge differential prices to reflect their differential costs. The only coherent alternative — though in our view a highly undesirable one — is to return to an élite system.

9. Implication 3: central planning of higher education is no longer feasible or desirable. Though many people — including us — mourn the loss of many aspects of the 1960s, the past is no longer on offer. Forty years ago, universities in England and Wales offered a fairly standard package of three-year, full-time degrees in a fairly limited range of subjects for 18-21 years olds. This model, it can be argued, was appropriate for a more static and less-technological era. Today, in contrast, technology requires that more people receive more education and training, and the labour market requires learning to be a lifelong experience. Thus there are many more students; the training they require is much more diverse; and it is changing and will continue to change. The task has become vastly too complex for central planning any longer to be possible.

10. Nor is central planning desirable (Barnes and Barr 1988). Market forces work best when consumers are well-informed. It is our firm view that students are very capable of making the necessary choices (which they make already) and will in future be capable of making more complex choices provided they have the necessary information. In recent years there has been considerable improvement in the provision of market information, particularly by the broadsheet press.

11. If students in future pay a significant fraction of the costs of their higher education (implication 1), their behaviour will change and that, in turn, will create demands for greater variation in what universities provide. Some students will demand more time-efficient courses, e.g. if the content of a three-year degree is taught over a two-year period without lengthy vacations, students have to finance only two years of maintenance and can enter the labour market a year earlier. Others will demand more flexible part-time options which can be combined with continuing labour market activity. Both effects will be accentuated by the knowledge that they will almost certainly have to return for additional training at a later stage in their career. Such strong changes in demand will require universities to respond in ways which are wholly impossible within a centrally-planned funding mechanism. Universities have to be free to decide the prices they charge, the types of courses and the number of places. Quantity, quality and price would result from the interactions of universities, students and employers within an appropriate regulatory framework. Student demand will be more attuned to continually evolving employer demand than central planning ever could. If government wishes to influence outcomes (e.g. to encourage more students to study engineering) they would do so by providing financial incentives for relevant courses.

12. The rest of the paper focuses mainly on implication 1, the need for an effective system of student loans, with less detailed discussion of implications 2 and 3. Many equally important issues are not discussed, including different ways of organising degrees; action to restore quality; and the maintenance of academic freedom in a market environment (though note that diversifying sources of income has significant benefits for academic freedom, since there is no single, dominant piper to call the tune). Section 2 discusses the design of a system intended to achieve the objectives set out earlier. Section 3 looks at how that design might be implemented, including discussion of the mechanics of administration and of the privatisation of the bulk of student debt. Section 4 examines outcomes, in particular the overall effectiveness of the proposed arrangements; the broad magnitude of the resulting saving in public expenditure; and repayments from a student perspective; the section concludes with a brief discussion of the possible impact on the incomes of universities, though this is not a topic about which much can be said. The final section looks at the way ahead, in terms both of further work and political aspects of implementation.

2 A New Design for Student Loans

13. Why a new scheme? The present student loan scheme was introduced in 1990. Loans are disbursed by the Student Loans Company (SLC), which also collects repayments. Loans carry a zero real interest rate, i.e. the loan incorporates an interest subsidy. People who have borrowed while a student (henceforth, for short, graduates) normally repay in 60 equal monthly instalments, i.e. repayments are organised like a mortgage or bank overdraft. Repayment can be deferred for 12 months at a time if the graduate’s income is below 85 per cent of national average earnings. The source of funds is the Treasury, i.e. the taxpayer.

14. The argument against this scheme is simple. It fails totally to achieve any of the objectives set out earlier.

• It has mortgage-type repayments, harming access. When the scheme is fully phased in, the loan will provide half of student maintenance; the resulting repayments will absorb about 10 per cent of the income of someone earning just over the 85 per cent threshold. Further expansion of this type of loan would put graduates at a disadvantage in buying a house.
• It provides no additional resources for universities, and hence does nothing to restore quality.
• It is hugely and unnecessarily costly in public expenditure terms: students borrow public money; loans carry an interest subsidy; deferments slow the repayment flow; defaults are a growing problem; and administrative costs are high.
• It mobilises no private funding at all.

All this was both predictable and predicted (Barr, 1989, Ch. 4; see also Barr, 1993, and Barr and Crawford, 1996, section 2).

2.1 Description

15. Collecting repayments. Though there is controversy about charging fees at an undergraduate level, there is now virtual unanimity (including the Committee of Vice-Chancellors and Principals and the National Union of Students) that policy for financing higher education should include student loans with two characteristics: repayments should be income-contingent (that is, should take the form of x per cent of the borrower’s subsequent earnings); and, to the maximum extent possible, the loans should derive from private-sector sources.

16. Barr and Crawford (1996), drawing on earlier work, explain how to put into place both legs of the strategy:

• Income-contingent repayments should be implemented by piggy-backing student loan repayments onto National Insurance Contributions (NICs). Repayments are collected by the National Insurance Contributions Agency. This approach automatically gives full income-contingency, is administratively cheap, and is secure.
• Because repayment is secure, it can be spread over an extended period, making it possible to charge students a positive real interest rate. The combination of a positive real return plus security makes such loans interesting to the private sector.

17. A number of characteristics of the scheme should be stressed.

• It is a loan, not a graduate tax. Nobody who has borrowed while a student repays more than he/she has borrowed, plus interest (the one possible exception is discussed in section 4.1). This is deliberate. A central problem with a graduate tax (i.e. an additional tax for life or until retirement) is that students borrow from the Treasury rather than the private sector; thus a graduate tax releases no additional resources to restore quality except in the longer-term.
• Because the scheme is fully income-contingent, it is possible for low earners to make repayments. Specifically, repayment starts at the lower earnings limit for National Insurance Contributions (£61 per week in 1996-7); at that level of income, a graduate whose loan repayment takes the form of an add-on to his/her National Insurance Contribution of 3p in the pound would have repayments of under 20p per week.
• Because repayment is spread over an extended period, weekly repayments are low, so that students can afford to pay a market interest rate which covers the costs both of funds and of administration. Thus the scheme contains no interest subsidy.
• Graduates have the option at any time to accelerate repayments, including repayment in full. Since there is no interest subsidy, graduates face no artificial incentive to prolong loan duration.

18. Privatising the source of finance. The mechanism for bringing in private funds is for the Student Loans Company, itself or through an intermediary, to sell student debt to the private sector. This technique, known as securitisation, is widely used, for example by credit card companies. Suppose the SLC wishes to sell £1 billion of student debt. A simulation study by Barr and Falkingham (1993), discussed in more detail in section 4.1, suggests that in the long run between 15 and 20 per cent of total student borrowing will not be repaid because of low earnings, emigration, premature death, etc. Thus, provided the loans pay a satisfactory interest rate, the private market would buy £1 billion of student debt for about £800 million. Throughout, the paper deliberately uses this conservative estimate to illustrate the argument. The Treasury pays only the remaining £200 million, rather than the whole of student borrowing, as at present. The new owners of the debt would be entitled to the loan repayments of the graduates. The financial market bears the risk of loan repayments falling short of £800 million.

19. Section 3.2 discusses ways in which securitisation might be implemented and section 4.1 suggests a possible way to increase the fraction of total borrowing which is returned to lenders from 80 per cent to 90 per cent, or possibly even to 100 per cent, in which case Treasury involvement in loan finance could be reduced further or even eliminated.

2.2 Advantages

20. These arrangements, in sharpest contrast with the current scheme, achieve the objectives described earlier.

21. Promoting access. NICs are a user-friendly form of repayment, which most graduates would actively prefer to the present arrangements.

• In contrast with the current scheme, repayments are fully income-contingent and hence automatically affordable. The threshold at which repayment starts is the lower earnings limit for NICs, a figure which — in contrast with the 85 per cent threshold of the current scheme — takes account of incentives and poverty traps. The repayment period varies with the student’s ability to repay, becoming shorter as his/her earnings increase. All this happens automatically, without any application from the student and without the need for complex enquiries and frequent reviews by the SLC.
• There is a ceiling on repayments, since there is an upper earnings limit for NICs. This eliminates any major disincentives for higher earners.

22. Repayment through any other mechanism inevitably harms access. The SLC, as permitted by current regulations, goes to great lengths to avoid students getting on to credit blacklists. Despite the SLC’s best endeavours, however, credit agencies blacklist any student against whom a summons is issued. Undoing the damage is extremely difficult, even where the student was guilty of nothing worse than inefficient paperwork: credit agencies look at each other’s lists and a name, once entered on one list, tends to be perpetuated on others, blighting applications for credit cards, to say nothing of mortgages. Current regulations allow the SLC to sell its debt to private debt collection agencies, but impose no constraint on the methods used by such agencies. The publicity surrounding private debt collection activities (it is easy to imagine a mistaken midnight raid on two young women students the night before an important exam) would simultaneously deter access and create political embarrassment for Ministers.

23. For these and other reasons, the emerging consensus — that maintenance should be financed by loans rather than grants — depends critically on loans being income contingent. Any other approach would immediately destroy the consensus.

24. Restoring quality. The NIC scheme is cheaper to administer than the current scheme, allows larger loans and brings in private funding. In sharpest contrast with current arrangements, the scheme frees resources which could be used to restore quality. This could happen in either or both of two ways: with less public spending on maintenance, there could be more public spending on the universities themselves; and/or part of the loan could be used to pay fees.

25. Taxpayer cost. Savings arise for several reasons.

26. Accelerated repayment stream in comparison with the current system. The NIC arrangement performs better than other loan schemes for two separate reasons. First, unlike current arrangements here and in the USA, repayments are not organised like a mortgage but are tied into the enforcement of NICs, and hence have a low default rate. Second, because repayments are income contingent, they can start at a much lower level of income, thus collecting repayments from a much denser part of the income distribution. Specifically, graduates start to repay as soon as their income exceeds the lower earnings limit for NICs. The current scheme, in contrast, collects no repayments from anyone whose income is below a threshold of 85 per cent of national average earnings (making the threshold currently about £300 per week). By collecting at least some repayment from people earning between £61 and £300 per week, the NIC scheme has a faster repayment stream (Barr and Falkingham, 1993, Figure 1; Harding, 1993, provides similar findings — for similar reasons — in the Australian system). The point is important: currently nearly half of all graduates defer repayment; when repayment is organised as an add-on to NICs deferment is (a) automatic below the lower earnings limit and (b) applies to vastly fewer graduates.

27. There are other reasons why the repayment stream is faster. Because NICs operate on a weekly basis, it is possible to collect repayments from the graduate’s first pay cheque onwards. Under current regulations, in contrast, a student graduating in July in effect has an automatic deferment until the start of the new tax year the following April. Second, recovery through NICs captures earnings from all work, for both employed and self-employed, and for however short the period, irrespective of the number of concurrent or consecutive employers involved. Third, NICs more easily capture income from work abroad for UK employers — important in the EU context.

28. No interest subsidy. Since repayments are fully income-contingent, it is possible to charge students a positive real interest rate. As discussed in section 3.2, the NIC mechanism is sufficiently secure that students would be able to borrow at only a fraction above base rate.

29. Low default rates. The default rate for NICs is around 1½ per cent.

30. Administration entirely paid by graduates. Administrative costs would be paid by graduates through their interest payments (as with any other loan), thus largely or wholly eliminating the taxpayer cost of administration.

31. Saving in SLC administrative costs, though no longer primarily the concern of the taxpayer, are immensely relevant to graduates who have to pay for them. The potential savings relative to the current scheme are discussed in detail by Barr and Crawford (1996, Appendix 1 and Table A1). Despite early criticisms, the SLC is now well on top of its administrative task. Discussion in our earlier paper relates to criticisms not of the SLC but of the task. The great advantage of the NIC system is that it automatically adjusts to fluctuations in weekly/monthly income, in contrast with the current scheme, which requires manual calculations and adjustments and only responds to yearly income.

32. Liberating private funds rapidly. The taxpayer savings listed in the previous paragraphs are those which arise if loans are not privatised. It is precisely those advantages (strong repayment stream, no interest subsidy, a low default rate and low administrative cost) which attract the private sector. Their absence explains the conspicuous lack of private-sector interest in the current scheme.

33. If loans are privatised, the savings to the taxpayer are larger and sooner. The SLC estimates that in the 1996-97 academic year it will lend about £1 billion. Securitisation, by bringing in (say) £800 million, would thus produce immediate savings in public spending of that amount per year.

34. Further expansion is possible. It would be possible to allow students to borrow up to an amount equal to the present grant plus loan plus parental contributions. It might be desirable to increase further the amount each student can borrow — to allow students a somewhat higher living standard than under current arrangements and/or to allow for additional fees charged by universities.

35. Section 4.2 discusses the broad impact of such changes on public spending.

36. In sum, these arrangements have the following advantages. They allow larger loans, an extended repayment period and market interest rates. They bring in private funds. They are administratively cheaper. They have major political benefits: student support would no longer be means tested; the ‘compulsory’ parental contribution would be abolished; and graduates would rarely be involved in Court cases, would not appear on credit blacklists and would not be pursued by private debt collectors. The scheme is also flexible in that, as discussed in section 4.3, it can readily be expanded to postgraduate students, part-time students, students in non-advanced further education and/or to students doing vocational training.

3 Implementation: The Mechanics of the Scheme

37. Sequencing. Implementing policy in the right order — sequencing — is critical but frequently overlooked.

38. Loans must precede additional fees. Loans can be introduced whether or not students are charged an additional fee. Additional fees, however, cannot be effective unless they are supported by a wide-ranging loan system. Fees without loans would be a policy with as few redeeming features as the current student loan scheme.

• They would be inequitable, socially divisive and inimical to access.
• Nor, without an effective loan scheme, could such fees be very high. Thus they would bring in relatively little extra income to universities.
• Fees without loans would also require universities to run an extensive system of fee remissions. The high administrative costs involved would further reduce the resulting net income to universities.

39. Repayment through NICs must precede privatisation. Privatising the source of loans cannot take place at any sensible price until repayment through NICs has been put into place. It is precisely the security of the NIC repayment mechanism which makes it possible to sell student debt for around 80 per cent of its face value rather than, say, 30 per cent. A critical (and novel) feature of the NIC proposal is its use of a public repayment mechanism to liberate private funds. Attempts to implement the privatisation part of the package without repayments through NICs would fetch a very low price, at enormous and continuing waste in public expenditure.

3.1 Administering Repayments

40. It is useful to divide administration into five logical functions: disbursing loans, keeping records, administering debt collection, administering deferments and finalising recovery. Barr and Crawford (1996, Appendix 1 and Appendix 2) discuss these in more detail and rebut official criticisms of the proposed administrative changes.

41. Disbursing loans. The SLC would set up an account for each new borrower and disburse loan payments broadly as at present. In addition, it would transmit the borrower’s name and National Insurance number to the Contributions Agency so that a revised National Insurance number could be issued, most simply by adding a suffix ‘E’ to his/her National Insurance number for the duration of the loan. The purpose of the suffix would be to inform employers of the need to withhold the add-on to NICs.

42. Keeping records. The SLC needs to maintain two sorts of record: about the borrower, and about the loan.

43. Records about the borrower. The key information is the graduate’s name and National Insurance number. The SLC would also need the graduate’s address to send annual statements of loan balances, as required by the Consumer Credit Act, and, in most cases, to send a final refund when the loan had been repaid. This task of record keeping, however, is smaller and easier than currently. First, under the present scheme, a missing address can mean a missing repayment; with the NIC scheme it merely means that the graduate will not receive his/her annual statement. Thus the repayment stream is independent of SLC knowledge of addresses. Second, the graduate has an incentive, rather than a disincentive, to keep in touch with the SLC in order to receive statements and his/her final refund. Third, if necessary, addresses are generally available from the social security database. Thus the NIC scheme has a double advantage: it is not necessary to have a current address in order to collect repayments; but if an address is needed, it is generally available from social security records.

44. Records about the loan. The SLC would set up and maintain graduates’ repayment records, ascertaining when the debt was repaid and then closing the record. For most graduates, recording debt and repayment would be done electronically by monthly payments en bloc from the Contributions Agency, coupled with annual electronic returns from the Contributions Agency notifying the SLC of the details of each graduate’s repayments during that year.

45. The task of record keeping would be cheaper than under the current system (a) because it would be smaller (less detailed records on borrowers) and (b) because the recording of debt and repayment would be streamlined, computerised, and would exploit the Contributions Agency’s administrative economies of scale.

46. Administering debt collection. Under current arrangements, the SLC is supposed to collect repayments. Its first task is to remind students of their entitlement to defer. It formerly sent up to eight reminders advising the graduate that court action will follow if he/she failed to repay or to establish an entitlement to defer. The SLC is currently moving to an accelerated collections policy under which only four letters are sent prior to a summons being issued.

47. Under the NIC scheme this time-consuming and costly function would almost totally disappear. For the great bulk of graduates, compliance would be through enforcement of NICs. Employers would implement enhanced National Insurance deductions automatically through payroll computer deductions or deduction tables. Under current arrangements, employers transfer withholdings of income tax and NICs to the Inland Revenue, which strips out the income tax and passes the NICs on to the Contributions Agency. The only change is that the Contributions Agency would now receive both NICs and student loan repayments. The latter would be electronically transferred to the SLC on a monthly basis, a cheap operation from the SLC’s perspective, since repayments would come mostly through this single channel. Development work is needed for this last task. Note, however, that this is a once-and-for all cost.

48. Administering deferments. Under present arrangements the SLC (a) informs students that they are entitled to defer repayments if their income is below 85 per cent of national average earnings, (b) seeks evidence, usually from pay slips, which confirms the graduate’s earnings, and (c) follows up on graduates who have neither made a repayment nor applied for deferment. This involves a large, sophisticated, and highly expensive computer/telephone operation. Under the NIC scheme this function would totally disappear, leaving only a small service function to respond to student queries.

49. Finalising recovery. Once a loan was fully repaid, the SLC would ask the Contributions Agency to remove the suffix E from the graduate’s National Insurance number. It would also be necessary to refund any overpayment. Small overpayments would be frequent, since the Contributions Agency remits repayments to the SLC monthly, but transmits information about individuals only at the end of the tax year.

50. Feasibility and precedent. No major new administrative task is involved. The National Insurance system already collects repayment at different rates for different people. In the past, married women could choose to restrict their contribution to industrial injury insurance, and there are still some remaining claimants of this option. In addition, anyone who continues to work after pensionable age pays a zero employee contribution. Existing software is designed to cope with these differential contribution rates.

51. Nor is any precedent broken. There is no issue of principle in using the NIC system to collect private debt, since the National Insurance system already does so. Under present arrangements, people can (a) belong the state earnings-related pension scheme (SERPS), paying NICs at the full rate, or (b) contract out and belong to an occupational scheme, paying NICs at the (lower) contracted out rate, or (c) contribute to a personal pension. In the case of (c), people pay NICs at the full rate. The personal pension component is stripped out once a year by the Contributions Agency and passed on to the personal pension scheme to which the person belongs. This is an example of public collection on behalf of a private firm. Many of the private pension schemes would not exist if they did not have the guarantee of the income through the National Insurance contribution.

52. These administrative arrangements are not only feasible; if properly implemented they could have a number of significant advantages for current and future students. First, a list of the names and addresses of UK graduates of any generation is very valuable. The SLC could include appropriate advertising with each graduate’s annual statement. The money thereby raised would, wholly or in part, offset administrative costs, thus lowering the charge to students. It is important that the list of graduates remains under the control of the SLC to ensure a continuing benefit to student borrowers.

53. Second, the standard loan agreement should include a requirement that the student would complete a series of questionnaires after graduation. Such questionnaires (which would probably be organised on a sample basis), would provide universities, employers and future students with labour market information. By focussing on outcome rather than input, this would assist well-informed choices by subsequent students of their course and institution. Information is vital to informed consumer choice; it also creates pressures to maintain quality.

54. The central point is that the administration of loan repayments through an add-on to NICs is feasible, will be cheaper and more effective than current arrangements and raises no issue of principle or precedent. The SLC is well equipped to undertake the task in co-operation with the Contributions Agency.

3.2 Securitisation

55. The mechanism for bringing in private funds rapidly is to securitise student debt. There are two broad approaches to putting the idea into practice.

56. Method 1: current public spending with no guarantee. This would work as follows.

• The SLC decides it needs £1 billion to finance the next tranche of student loans, and offers £1 billion of debt for sale. Private buyers pay (say) £800 million to buy the future repayment stream and the Treasury pays the SLC the shortfall of £200 million. Thus the SLC has £1 billion to lend to students.
• The total amount of debt being sold would be offered in small chunks, say of £10 million, deliberately relatively small to make it easier for smaller financial institutions to participate, hence making the process more competitive.
• The scheme would go well beyond the commercial banks. Bidders would also include merchant banks, life insurance companies and, importantly, pension funds.
• Such offers would be made periodically (i.e. once or twice a year) as the SLC needed to replenish its coffers.
• Bidding would be through sealed bids or by auction. Each bid would specify two things: the number of units being bid for, say, 10 units (£100 million), and the offer price per unit.
• Each bidder could make more than one bid, e.g. 5 units at £9 million each, 10 units at £8 million each, etc.
• The units would be sold to the highest bidders. The bidder offering the highest price would get as many units as it had bid for, and similarly for the next highest until all the units had been sold.

57. This approach costs the taxpayer a one-time payment of £200 million. That payment would be current public spending in the year in question. There would be no guarantee to private markets, and hence no contingent liability. These arrangements completely fulfil the Private Finance Initiative’s criterion of a total transfer of risk to the private sector.

58. Method 2: limited guarantee. This approach, which privatises matters even further, differs from that above in two main ways. First, the Treasury would not make a cash payment but would instead offer a guarantee on the first £200 million of losses incurred in the portfolio. With such a guarantee, the market would buy the debt for around £1 billion. In this way the government never has to pay more than £200 million, but may benefit if losses are below the expected 20 per cent. This may have advantages in meeting the Treasury’s value-for-money criteria.

59. Second, the SLC does not sell debt directly to a range of financial institutions, but to a single institution. It has been suggested by financial market practitioners that government involvement of this type would make it attractive for one underwriter to buy the entire issue. The underwriter then sells bonds to individual institutional investors who have the option of buying in larger or smaller amounts, possibly as small as £10,000. In effect, the SLC does not act as retailer of debt but sells the debt wholesale.

60. From the perspective of financial markets, a £1 billion transaction launched with government sponsorship rather than a cash discount, and with a 100 per cent credit wrap by a AAA-rated institution, would be very well received. The view of the market is critical: the more favourable, the lower the interest rate necessary to sell the debt for its face value of £1 billion.

61. Advantages. Quite separate from its advantages for higher education (discussed shortly) securitisation has major advantages in terms of financial markets:

• It is genuinely competitive, thus reducing the cost of funds. The interest rate would be slightly higher than the government bond rate, less for reasons of risk than because of the transactions costs of large numbers of small repayments.
• The asset would be tradeable.
• Because bidding is annual (or more frequent) financial markets can build up experience of the new asset.
• It allows participants to enter on a small scale if they wish, a particularly desirable feature in the early years of the scheme.
• There would be little short-run fluctuation in the volume of bonds sold. In contrast with government bonds, these sales are not cyclical. Thus the market can predict how many such bonds are likely to come on to the market over the next 5-10 years.
• It allows diverse actors to participate in the market for student loans.

62. The financial asset, being long term, would be of particular interest to pension funds. Student debt could, for example, be held by a miners’ pension fund, benefiting retired miners whose pension would, in part derive from the human capital of younger graduates, and also benefiting younger people by providing the loan capital which finances part of the costs of their higher education.

4 Outcomes

63. Outcomes are of two sorts. Macroeconomic effects include the overall performance of the scheme (section 4.1) and its impact on public expenditure (4.2). Microeconomic effects include the impact on students (section 4.3) and universities (4.4).

4.1 The Overall Performance of the Scheme

64. A central issue in earlier discussion is the fraction of total student borrowing which will be repaid. This question was investigated with the LSE Welfare State Programme’s LIFEMOD, which contains synthetic life-cycles for 4000 individuals, based on 1985 data, including 1985 probabilities of marriage and its breakdown, of having one, two, three, etc. children, and of entry into and exit from unemployment or ill-health. Barr and Falkingham (1993) investigated in some detail a loan of £1000 (1985 prices) for each year of a three-year degree, i.e. a total loan of £3000, repaid by an add-on to NICs of one penny in the pound under various assumptions about (a) the real rate of growth of graduates’ earnings and (b) the real interest rate paid by graduates on their loans. Barr and Falkingham (1996) tested the sensitivity of the earlier simulations for add-ons between 1p and 2.5 pence for a total loan of £3000 (1985 prices).

65. The two papers investigated four variables: the percentage of individuals who repay their loans in full; the percentage of total lending which is repaid; the average duration of repayment; and loan repayments as a proportion of gross earnings. This section looks mainly at the results for the second variable — the fraction of total lending which is repaid — shown in Table 1, though where relevant it also refers to the other variables (for the full results see Barr and Falkingham 1993, Tables 1A, 1B and 1C).

66. Repayment through an add-on of 1p in the pound. In Table 1 students are assumed to make loan repayments as an add-on to NICs of 1p in the pound for each £3000 (1985 prices) they have borrowed.

67. The benchmark case (zero earnings growth, zero real interest rate) shows the very different repayment dynamics of mortgage repayments and income-contingent repayments. Taking men and women together, the fraction of borrowers who repay in full (not shown in the table) is seven out of ten under the current scheme, but only five out of ten under the NIC scheme. In sharp contrast, as shown in the top part of Table 1, there is virtually no difference in the fraction of total lending which is repaid under the two schemes, each repaying about 79 per cent.Table 1 about here

68. This result merits explanation. Under the current scheme, repayment is zero for people with incomes below 85 per cent of the national average (around £300 per week); otherwise it is 10 per cent of the loan. In contrast, the income-contingent loan, precisely because it is income-contingent, allows people above the lower earnings limit for NICs (£61 per week) to make at least small repayments. Because repayments are lower, more people do not repay in full. But because the NIC scheme collects repayments on earnings between £61 and £300 per week, total repayments match those of the current scheme. Under the assumptions of Table 1 (in particular zero real earnings growth), both schemes could operate with a government contribution of slightly over 20 per cent of total lending to students.

69. The average repayment period (not shown) is 17 years for the current scheme and 28 years under the NIC scheme. The longer repayment duration under the NIC scheme is, in many ways, a point in its favour. In efficiency terms, it should be possible to spread repayment over the life of the asset. Thus repayment should last three years for new car, while for a university degree or similar long-lasting qualification the option of a longer repayment duration is efficient.

70. Earnings growth. Earnings growth of 1½ and 3 per cent is investigated in the middle and lower parts of Table 1. With mortgage-type schemes, repayment depends little, if at all, on earnings; thus repayments are higher and/or faster under the government scheme only to the extent that more people creep over the 85 per cent threshold. In contrast, earnings growth has a dramatic effect on the performance of income-contingent schemes. With a zero real interest rate, the combined repayments of men and women rise to 85 per cent of total lending with 1½ per cent earnings growth, and to nearly 90 per cent with 3 per cent earnings growth compared, in both cases, with about 79 per cent for the government scheme. It should be noted that real earnings growth of 3 per cent is not implausible: individual earnings rise over time both because the overall structure of earnings increases and, because of life-cycle effects, as individuals move up their pay ladder. The Government Actuary’s estimate of long-run overall annual earnings growth is 1½ per cent is entirely compatible with 3 per cent growth in individual earnings.

71. Alternatively, with 3 per cent earnings growth, loans could pay a real interest rate of 2 per cent (i.e. an interest rate two per cent above the rate of inflation) and still repay 76 per cent, nearly as much as the current scheme, which charges no interest.

TABLE 1:  PROPORTION OF LENDING REPAID: CURRENT LOAN SCHEME AND NIC SCHEME WITH A REPAYMENT RATE OF 1P PER POUND OF EARNINGS^a

SOURCE: Extracted from Barr and Falkingham (1993, Tables 1A, 1B and 1C)

Note: ^a Simulations are for full-time students in advanced further and higher education.

^b NIC = scheme in which loan repayments take the form of an add-on to borrowers’ National Insurance Contributions.

72. Differences between men and women. The second and third columns of Table 1 show the very different repayment patterns of men and women. Even with zero real earnings growth, men repay 90 per cent of total borrowing under the NIC scheme at a zero real interest rate. With three per cent earnings growth and a zero real interest rate, men repay 94 per cent of their total borrowing. Given the high proportion of men who repay in full, the effect of earnings growth is more to speed repayment than to increase the number of men repaying; the average loan duration falls from 28 years with no earnings growth to 21 years with 3 per cent growth.

73. With men the effect of earnings growth was mainly to speed repayment. With women, in contrast, earnings growth dramatically increases total repayments (column 3). At a zero interest rate, women’s repayment rise from about 65 per cent of total borrowing with zero earnings growth to over 80 per cent with 3 per cent earnings growth, compared with 67 per cent under the government scheme.

74. Extension to further education. Income-contingent loans are also highly effective for students in non-advanced further education (Barr and Falkingham, 1993, Tables 5A, 5B and 5C, and Figure 4). Though such students have more heterogenous earnings profiles than graduates, they study for shorter periods and hence take out smaller loans. In the benchmark case, men repay nearly 96 per cent of their total borrowing, women 78 per cent, and men and women together, over 85 per cent. Earnings growth leads to further improvement: with three per cent earnings growth, men could pay a real interest rate of 3 per cent and still repay 92 per cent of total borrowing. There are two reasons for this excellent performance: students in further education generally spend longer in the labour force than graduates; and income-contingent arrangements are able to collect at least some repayments from relatively low earners.

75. Sensitivity analysis. The fraction of total lending which is repaid is the variable of greatest interest to financial markets. Table 2 therefore repeats the simulations in the first column of Table 1 for add-ons to NICs of 1 penny in the pound, 1.5p, 2p and 2.5p (more detailed results are given in Barr and Falkingham 1996, Tables 2A, B and C).

76. In principle, raising repayment rates has two effects: it speeds up repayment; and it may raise the fraction of total borrowing which is repaid. The relative importance of the two effects is an empirical matter. A higher repayment rate will always speed up repayment; it does not necessarily increase the fraction of total borrowing which is repaid. To illustrate, assume that there are two sort of graduate: those who go straight into high-earning, stable employment, and those who never find a job of any sort. If that is the case, higher repayment rates do nothing to increase total repayments; they merely lead to a faster repayment stream.

77. The benchmark case (zero earnings growth and zero real interest rate). As it turns out, increasing the repayment rate has a dramatic effect on the fraction of student borrowing which is repaid. The first line of Table 2 shows the benchmark case for men and women together. An add-on of 1 penny in the pound to NICs leads to repayment of 79 per cent of total borrowing. This rises to 89 per cent at 1½p, to 93 per cent at 2p, and to nearly 95 per cent at 2½p.Table 2 about here

78. Looking at men and women separately (Barr and Falkingham, 1996, Tables 2B and 2C) the fraction of total lending repaid by men in the benchmark case increases significantly, from 90 per cent with a 1p add-on to 95 per cent at 1½ p to 97 per cent at 2½p. Higher repayment rates for women have an even more dramatic effect, increasing the fraction of total lending which is repaid from 65 per cent for a repayment rate of 1p, to 81 per cent at 1½p and to 91 per cent at 2½p.

79. Earnings growth, as expected, sharply improves results. With 1½ per cent earnings growth and a 2 per cent real interest rate, the fraction of total lending repaid by men and women together (the middle section of Table 2) rises from 68 per cent at 1p to 82 per cent at 1½p, to over 90 per cent at 2½p. Similarly, with 3 per cent earnings growth, where graduates pay a 3 per cent real interest rate on their loans, 67 per cent of total borrowing is repaid with a 1p repayment rate, rising to 80 per cent at 1½p, to 86 per cent at 2p, and to 90 per cent at 2½p.

80. Differences between men and women. Though not shown in Table 2 (see Barr and Falkingham, 1996, Tables 2B and 2C), the different pattern of results for men and women is revealing. Looking at the example in the previous paragraph of 3 per cent real earnings growth and a 3 per cent real interest rate, the fraction of total lending repaid by men rises from 85 per cent of total lending for a 1p add-on to 92 per cent at 1.5p. With positive earnings growth, increasing repayment rates above 1½p speeds repayment but, in the case of men, does not greatly increase the fraction of total borrowing which is repaid. For women, in contrast, the fraction of total lending which is repaid rises steadily with the repayment rate, from 49 per cent with 3 per cent earnings growth, a 3 per cent real interest rate and a repayment rate of 1p, to 68 per cent with a repayment rate of 1½p, to 78 per cent at 2p, and to 83 per cent with a repayment rate of 2½p.

81. Interpreting the results. What do these numbers imply for the design of a loan system?

82. The fraction of total borrowing which will be repaid. Higher add-ons bring in more repayments more quickly. However, they also increase the risk of work disincentives. The discussion which follows is therefore based on two assumptions: that the repayment rate is 1½p per unit of borrowing; and that the government, as in other areas, rightly insists that men and women are treated as a single pool.

TABLE 2: PROPORTION OF LENDING REPAID WITH DIFFERENT REPAYMENT RATES^a

SOURCE: Barr and Falkingham (1996, Table 2A)

Note: ^a Simulations are for men and women full-time students in advanced further and higher education.

83. Thus the relevant figures are those in the second column of Table 2: if the real earnings of graduates rise by 3 per cent per year and students pay a 3 per cent real interest rate, an add-on to NICs of 1½p will generate a repayment stream which ends up repaying 80 per cent of total lending to students, i.e. £1 billion of student debt could be sold for £800 million. Repayment would be similar with an increase in real earnings of only 1½ per cent and a real interest rate of 2 per cent.

84. These results, through fairly robust, should not be invested with spurious precision. As described at the start of the section, LIFEMOD, from which the numbers derive, is based on 1985 data. Though marriage and fertility patterns have not changed significantly since then, labour market conditions have. Thus the results just described do not automatically apply today. Detailed, up-to-date estimates of repayment rates are needed.

85. The intuition of income-contingent repayments requires discussion because it differs significantly from that of mortgage-type loans. In 1985 terms, a 1½p add-on repays a total loan of £3000. Anticipating the discussion in section 4.3, translation into today’s terms means that an add-on of 2.3 pence can repay a loan of £10,000 in 1995 prices. At first sight this repayment rate appears too low to repay a loan of this size. Consider a new graduate earning £16,000 per year (broadly the national average) and paying an interest rate of 6 per cent (3 per cent superimposed on an inflation rate of 3 per cent). The first year’s interest charge is 6 per cent of £10,000, i.e. £600; but her loan repayment, 2.3 per cent of £16,000, is only £368. Thus her repayment does not cover her interest liability, and her nominal debt increases.

86. This result is entirely characteristic of income-contingent repayments. With mortgage-type loans, the annual repayment is fixed. As earnings rise, repayments therefore fall as a fraction of earnings — mortgage repayments are front-loaded. With income-contingent loans, in contrast, what is fixed is not the amount of annual repayment, but the fraction of income which is repaid. As real earnings rise so, therefore, do repayments. Income-contingent repayments are end-loaded, with heavy repayments in later years. A typical pattern, therefore, is for nominal debt to rise in the early years of repayment, then to start to fall, and to fall very rapidly in the final years of the loan.

87. Figure 1 (based on the data in Table A1) illustrates the time-path of the two methods of repayment assuming an initial loan of £1000, income-contingent repayments of 0.23 pence per pound of earnings, a starting salary of £16,000, inflation of 3 per cent per year and real earnings growth of 3 per cent. In the early years of the loan, income-contingent repayments (the dotted line) fail to cover interest charges and nominal debt rises. As real earnings rise, however, so do real repayments; the steep slope of the dotted line shows the speed with which the loan is extinguished in the later years of repayment. The loan is repaid in year 28. For comparison, the darker line shows the time path of mortgage repayments designed to repay the same loan over the same period. Being front-loaded, repayments are larger in the early years and smaller in the later years.Figure 1 about here

FIGURE 1: THE TRAJECTORY OF LOAN REPAYMENTS: INCOME-CONTINGENT AND MORTGAGE-TYPE LOANS

- - - - Income-contingent repayments

—— Mortgage repayments

88. A more radical option. The heavy repayments in later years under income-contingent arrangements opens up the possibility of a more radical option. Discussion throughout the paper is based on an estimate that 80 per cent of lending to students will be repaid. That figure could be increased to 90 per cent or even higher by relaxing the assumption that no graduate repays more than he/she has borrowed. Since income-contingent repayments are at their highest in the final years of the loan, a few extra years of repayment after the loan has been repaid have a powerful effect.

89. Suppose repayments continue for an extra 2-3 years after the individual graduate has repaid his loan. The data in the first column of Table A1 show that by year 31, this would result in additional nominal repayments of £471 or, adjusting for inflation of 3 per cent annually, of £188 in real terms. Such additional repayments, discounted to the present and averaged across all graduates, is equivalent to 6 per cent of the original debt, raising total repayments by graduates from 80 per cent to 86 per cent of their total borrowing. If repayment continues for between 4 and 5 years after the loan has been extinguished, total repayments rise to 92 per cent of the total amount borrowed, with major implications for the attractiveness of the debt to private markets, hence to the price at which it can be sold, and hence to the public expenditure savings engendered by the NIC scheme.

90. It can be argued that such an arrangement does not violate the integrity of the scheme as a genuine loan. A person who borrows to buy a house repays only what she has borrowed plus interest; however, many lenders make mortgage protection insurance obligatory. Similarly, student loans can be thought of as having two components: loan repayments (the first n years of repayment) plus an insurance premium (the extra years at the end). The effect of such premiums would be to ensure that the cohort as a whole repaid close to its entire debt.

4.2 The Impact on Public Spending

91. Under present arrangements, the maintenance grant and loan are both entirely publicly funded, expenditure currently running at about £1.7 billion per year. The scheme set out above would reduce this figure very substantially.

92. If loans took over the entire maintenance package and were privatised at an 80 per cent securitisation rate, there would be an instant saving of over £1.3 billion per year.

93. The SLC currently has an accumulated stock of about £2 billion of loans made under the current system. If these were securitised there would be an additional — albeit once-and-for-all — saving of about £1.6 billion (i.e. 80 per cent of £2 billion). Saving on this scale, however, depends on graduates transferring to the new repayment mechanism. This is plausible. Most graduates would actively prefer to repay through the NIC mechanism, so that holders of loans issued under the current system could readily be persuaded to switch to repayment through NICs. If incentives are needed, there are good ways of persuading holders of old loans to transfer: there could be a small rebate to those who converted; there could be an amnesty to defaulters who agreed to repay their old loan through the NIC mechanism; and, in the case of students with an old loan who become students again, a condition for a new loan could be to convert the old loan to the new repayment mechanism.

94. The scale of securitisation could be higher. If all maintenance grants plus parental contributions were replaced by loans (i.e. if students were allowed to borrow an amount equal to grant plus parental contribution plus loan), expenditure on student support would rise from £1.7 billion to around £2.5 billion. Securitisation would bring in some £2 billion; the Treasury would pay the remaining £500 million, either as direct public spending in the year in question (securitisation method 1 in section 3.2) or up to that amount at some future date (securitisation method 2). Thus the parental contribution could be abolished, and there would still be a net annual saving in public spending of the order of £1.2 billion.

95. The real value of the grant fell by about 25 per cent between the early 1960s and the mid-1980s. If the student support package were increased to allow students to borrow an amount whose real value approximated the early 1960s grant plus parental contributions, the total package would increase from £2.5 billion to around £3.1 billion. With an 80 per cent securitisation rate, the Treasury share would be around £625 million. Students could, if they wished, increase their living standards to those of students in the 1960s without resort to compulsory parental contributions, at a saving in public spending, compared with present arrangements, of over £1 billion per year.

96. Further expansion is possible. It may be desirable to extend loans to larger numbers of borrowers — to any or all of part-time students, postgraduate students and students doing vocational training, or to cover any further increase in student numbers. Further expansion would be possible to cover innovative educational ideas.

4.3 Repayments from the Perspective of Students

97. This section discusses the add-on necessary to support loans of different sizes, based on Tables 1 and 2. The wide variation in the repayment experiences of individual students is discussed in Barr and Falkingham (1993, section 3.2).

98. Method. The discussion in section 4.1 used a benchmark loan in 1985 of £1000 per year of study (roughly half of the 1985 grant) repaid by an add-on to NICs of one penny in the pound. Table 3 translates these figures into today’s terms. Since repayments are directly related to a person’s earnings, the relevant adjustment factor is the increase in nominal earnings. The earnings index for nonmanual workers rose from 100 to 217.3 between 1985 and 1995. Thus in today’s terms, a loan of £2183 per year, or just over £6500 for a three-year degree, could be repaid by:

• an add-on to NICs of one penny in the pound using the data on Table 1; in this case, each £1000 (in today’s prices) borrowed could be repaid by an add-on of just over 0.15 pence in the pound; or
• an add-on of 1½p using the data in the second column of Table 2; in this case, each £1000 borrowed could be repaid by an add-on of just below 0.23 pence in the pound.

99. Undergraduate degrees. Table 3 gives some examples, based on the latter figure, i.e. that students could repay their loan with an add-on of 0.23 pence per £1000 of borrowing.Table 3 about here

• A loan of £2000 per year (roughly half the current London maintenance grant), hence £6000 over three years, paying a 3 per cent real interest rate, can be repaid through an add-on to NICs of about 1.4 pence in the pound (line 1).
• A loan of £4000, roughly the full London maintenance grant, leading to a total loan of £12,000, could be repaid through an add-on of 2¾ pence in the pound (line 2).
• If the support package were increased by 25 per cent to restore the erosion since the early 1960s, students would have about £5000 per year to live on, repaid by an add-on of about 3½ pence (line 3).
• If, in addition, students paid fees of £1000 per year, leading to an annual loan of £6000, the add-on would be just over 4 pence (line 4).

• A student doing a four-year degree outside London, borrowing £4000 per year for maintenance and £500 per year for fees would take out a total loan of £18,000, repaid by an add-on of just over 4 pence (line 4).
• A student doing a three-year degree at an internationally competitive institution in London might borrow £5000 per year for maintenance plus £4000 per year for fees. The total loan of £27,000 could be repaid by an add-on of 6.2 pence (line 5)

TABLE 3: REPAYMENT RATES FOR LOANS OF DIFFERENT SIZES

Note: ^a In comparison with an add-on of 1 penny per pound of earnings, an add-on of 1½ pence has the twin effects of accelerating repayments and increasing the fraction of total borrowing which is repaid. See the discussion in section 4.1.

100. The extension to graduate studies. Loans could also be made available on the same terms to graduate students.

• If students doing a one-year masters degree paid fees of £7000 and borrowed £5000 for maintenance, the add-on would be 2¾ pence (line 6).
• A student doing a three-year degree outside London, borrowing £4000 per year for maintenance and then doing a one-year masters degree would borrow a total of £24,000, repaid by an add-on of 5½ pence (lines 2+6).
• An identical student, doing a degree in London and paying undergraduate fees of £1000 per year and then proceeding to a masters degree, could repay with an add-on of 6.9 pence (lines 4+6).

101. Thus an add-on to NICs of 4 pence in the pound would cover all but the most expensive undergraduate degrees. Those borrowing less would repay faster. To allow flexibility, an alternative approach is through an add-on of 3 pence in the pound for students whose total borrowing was below (say) £12,000 and 6 pence in the pound where total borrowing was between £12,000 and £24,000 with, possibly, a 7½p add-on for larger loans. Since loans pay a market interest rate and repayments cease once the loan has been repaid, the larger add-on has no major equity implications. In particular, if graduates’ earnings grew faster than expected and/or graduate unemployment fell sharply, loans would be repaid more quickly, but nobody would repay more than he or she had borrowed, plus interest. Note also that graduates have the option at any time to accelerate repayments, including the option to repay in full.

102. Other examples. The figures in Table 3 can be applied to other options. It has been suggested, for example, that all students should receive two years of free tuition (i.e there should be no additional fees for the first two years), but that thereafter they should receive no additional public support.

103. Consider a student outside London, who takes out a loan to cover maintenance of £4000 per year for three years plus an additional £5000 to cover fees in her third year. Her total loan of £17,000 could be repaid by an add-on of 3.9 pence.

104. An identical student doing a degree in London would take out a loan of £5000 per year for three years plus an additional, say, £7000 to study at an internationally competitive institution would borrow a total of £22,000, which could be repaid by an add-on of just over 5 pence.

4.4 Potential Effects on University Income: the Great Unknown

105. The basic arithmetic is simple. If there are 2 million students, an average additional fee of £1000 per student per year generates a gross increase in university income of £2 billion.

106. Simplicity, however, ends there, since increases in fee income from students may, to a greater or lesser extent, be offset by reductions in public funding for universities. At its worst, taxpayer support could be reduced by £1 million for each £1 million in fee income. Thus the effect on university income of introducing student fees is unknown and unknowable.

107. What can be said, however, is that many, if not all, universities have no choice. Public funding will never cover the costs of the internationally competitive institutions. Realistically, it is unlikely that it will cover the costs of universities with above-average quality and above-average costs. These institutions, if no others, will be compelled to charge additional fees.

108. More generally, the price of not charging fees is to continue to be underfunded, with no possibility of restoring quality. British universities would become like many continental universities. The realistic choice is not between adequate public funding and the uncertainties of the market place. With a mass system, the only choice is between inadequate public funding on the one hand and a chance to supplement public funding with private earning on the other.

5 The Way Ahead

109. Areas for more detailed work. Though the policy design is fairly complete, a number of details need to be established, for example the maximum amount a student is allowed to borrow (a) in any one year, and (b) in total.

110. The discussion of mechanics in section 3 — much of it based on contact with administrative and financial practitioners — was intended to establish the administrative and financial feasibility of the policy outlined in section 2. The next step is detailed work in both areas.

111. On the administrative side, the SLC, which has already given thought to future possibilities, and the Contributions Agency need to discuss the most cost-effective way of administering the scheme, basing cost estimates on the marginal cost of additional activities.

112. Making the scheme operational on the financial side involves the right steps in the right order. Since financial institutions will have to buy the asset, they must be involved in discussion of all relevant aspects of the scheme prior to legislation. Particularly relevant are institutions such as life insurance companies and pension funds whose major interest is in long-term assets.

• It is necessary, first, to establish the criteria required by international statistical conventions on the divide between public and private expenditure. For that reason the Central Statistical Office must be involved, and also the Treasury.
• The detailed mechanics of the way in which student debt should be securitised (two different ways were discussed in section 3.2) require discussion by all interested parties.
• Work is needed on the level of the interest rate paid by students, and on whether it would be fixed in nominal terms (like government bonds), fixed in real terms (like indexed gilts) or variable (like mortgages)
• Additional estimates are needed of the fraction of total student borrowing which is likely to be repaid.

113. Political aspects. One of the reasons why the Committee of Inquiry was established was the political sensitivity of higher education finance. Yet, if the nettle is grasped, the politics of student loans need not be the source of political terror they have become. If loans are properly structured, there will be very few people who experience loss but no gain.

114. Students. There is only one group of unambiguous losers: those students who would have found a place at university even in the old, élite days. Instead of receiving generous tax funding, they now have to take out loans.

115. For other students, the balance of gains and losses is more complex.

• They lose relative to perceptions of a long-gone grant system.
• They also lose in that they generally have to take out a larger loan than at present. There are, however, offsetting factors: repayment is exactly tailored to the individual graduate’s earnings; repayment is spread over more years than under the current scheme; and even the highest-earning graduates receive protection, since the upper earnings limit for NICs imposes a ceiling on annual repayments.
• The great bulk of students gain because they can get a place at university in the first place. Under the old system, though in principle eligible for generous grants, they would mostly not have found a university place. If we were to return to the participation rates of the early 1960s, nearly five-sixths of current university students would have to be asked to leave.
• Students also benefit because they would gain access to their future earnings and would therefore no longer be compelled, via the parental contribution, to depend on family support. Parents who wished to help their children could do so; and children who were prepared to accept parental assistance could take out a smaller loan.
• Students also gain, if the scheme is properly structured, because they can have a higher living standard than currently.
• Gainers also include students just above the 85 per cent threshold, whose payments, being exactly related to their income, are lower under the NIC scheme than under present arrangements.

116. Parents gain unambiguously. They are no longer subject to a means test; and the abolition of parental contributions, coupled with a non-penal loan system, leaves a neutral choice for families about whether, how much, and in what form, parents assisted their children. They retain the freedom to do so as much as at present, but are no longer under moral pressure to do so, greatly reducing parental guilt trips.

117. The taxpayer. Provided student debt is privatised, the taxpayer benefits. Taxes can be cut or the savings can be used for other purposes, for example in addressing critical problems in primary and secondary education.

118. Universities. A central part of the politics of universities is the pent-up demand for pay increases, estimated by the Committee of Vice-Chancellors and Principals to be £1.2 billion in 1995/96. Alongside the need for internationally competitive institutions to pay internationally competitive salaries is the more general need to maintain overall quality in the university sector. A co-equal source of discontent is the concern of university staff about protecting and enhancing the quality of higher education. Reform of the sort described above is the only way to free the resources to deal with these problems.

119. Concluding thoughts. In principle there are three choices for higher education:

• to return to an élite system;
• to continue to be underfunded, becoming more and more like continental universities; or
• to introduce an effective system of loans which allows universities to charge additional fees.

120. In practice, however, there is no choice, because the first two options are not viable. By 1989, higher education funding in Britain was already stretched with a 14 per cent participation rate — the lowest in the entire developed world. Global competition requires an increasing fraction of the labour force to be increasingly well-trained in broad, flexible skills. Labour market trends increasingly show the bleak employment prospects of those with little education. A return to an élite system is both economically unthinkable and politically impossible.

121. Continued underfunding of a mass system, similarly, puts national economic competitiveness at risk. Furthermore, with the large number of overseas students studying in Britain, higher education is one of the country’s most successful export industries. Underfunding undermines both the direct exports and the subsequent more wide-ranging exports which result indirectly. The Asian ‘tigers’, it should be noted, do not aspire to the Continental model and, not least because of their geographical proximity, the Australians were among the first to recognise the need to put more resources into their system of higher education.

122. The third option is the only one left. There really is no choice.

TABLE A1:   REPAYMENT STREAMS UNDER DIFFERENT ASSUMPTIONS

REFERENCES

Barden, Laing, Nicholas Barr and Gordon Higginson (1991), An Analysis of Student Loan Options, CVCP/CDP Occasional Paper, London: Committee of Vice-Chancellors and Principals and Committee of Directors of Polytechnics.

Barnes, John and Nicholas Barr (1988), Strategies for Higher Education: The Alternative White Paper, Aberdeen University Press for the David Hume Institute, Edinburgh, and the Suntory-Toyota International Centre for Economics and Related Disciplines, London School of Economics.

Barr, Nicholas (1989), Student Loans: the Next Steps, Aberdeen University Press for the David Hume Institute, Edinburgh, and the Suntory-Toyota International Centre for Economics and Related Disciplines, London School of Economics.

Barr, Nicholas (1993), ‘Alternative Funding Resources for Higher Education’, Economic Journal, Vol. 103, No. 418, pp. 718-28, May.

Barr, Nicholas, and Iain Crawford (1996), Student Loans: Where Are We Now? London School of Economics, Welfare State Programme, Discussion Paper WSP/127.

Barr, Nicholas and Jane Falkingham (1993), Paying for Learning, London School of Economics, Welfare State Programme, Discussion Paper WSP/94.

Barr, Nicholas, and Jane Falkingham (1996), Repayment Rates for Student Loans: Some Sensitivity Tests, London School of Economics, Welfare State Programme, Discussion Paper WSP/127.

Barr, Nicholas, and William Low (1988), Student Grants and Student Poverty, Welfare State Programme, Discussion Paper No. WSP/28, London: London School of Economics.

Falkingham, Jane and Carli Lessof (1991) LIFEMOD - The Formative Years, Welfare State Programme Research Note No. 24. London: London School of Economics.

Falkingham, Jane and Carli Lessof (1992), ‘Playing God - the construction of a dynamic microsimulation model’ in Hancock, Ruth and Sutherland, Holly (eds) Microsimulation Models for Public Policy Analysis: New Frontiers, London: London School of Economics.

Glennerster, Howard, Jane Falkingham, and Nicholas Barr (1995), ‘Education Funding, Equity and the Life Cycle’ in Falkingham, Jane and Hills, John (eds), The Dynamic of Welfare: The Welfare State and the Life Cycle, Hemel Hempstead: Prentice-Hall/Harvester Wheatsheaf, pp. 137-149.

Harding, Ann (1993), Lifetime Repayment Patterns for HECS and Austudy Loans, National Centre for Social and Economic Modelling, University of Canberra, Discussion Paper No. 1.

National Education Commission (1993), Learning to Succeed, Paul Hamlyn Foundation National Commission on Education, London: Heinemann.

[Return to Top] [Return to Index]