Memorandum from Dr Nicholas Barr (SS 09)
1. This paper puts forward a strategy for
achieving two objectives in higher educationimproved access
and increased qualityabout which there is unanimous agreement.
2. Diagnosis. The introduction of income-contingent
repayments in 1998 was a genuine and enormous advance. However,
two strategic problems remain. First, income-contingency is little
understood, causing unnecessary fear of debt (solutions are discussed
in section 4.2). Second, all the funding problems of the current
system go backdirectly or indirectlyto the subsidised
interest rate on student loans. Australia and New Zealand face
identical problems for identical reasons.
3. Interest subsidies create three problems.
They are badly-targeted, mainly benefiting high-earning graduates
in mid career. They are expensive (a recently-developed model
estimates conservatively that out of next year's lending to students
of £2,500 million about £700 million will never come
back because of interest subsidies). Third, because loans are
so expensive, the Treasury rations them. Thus interest subsidies,
like most subsidies harm the people they are meant to help. There
was an experiment with subsidies called Communism. It did not
work. The result is that loans are too small, leading to student
poverty and extensive use of credit card debt; and loans are means-tested:
thus parental contributions and upfront costs continue.
4. Prescription. If graduates pay an interest
rate equal to the government's cost of borrowing (not the bank
overdraft rate), repayments increase from about 50 per cent of
total borrowing to about 85 per cent (the remaining 15 per cent
shortfall being mainly due to low lifetime earnings), largely
eliminating the fiscal impediment to expanding loans. The move
is politically less difficult than it sounds. Interest rates are
currently low, so that a move to the government's cost of borrowing
involves only a small increase to the rate that graduates pay.
Second, a graduate's monthly repayments depend only on her income;
thus an increase in interest rates has no effect on monthly repayments,
instead affecting the duration of the loanmaking it clear
that repayments are simply a form of targeted income tax.
5. Policies. Removing interest subsidies
is the single essential key to solving current funding problems.
The considerable resources thereby released underwrite the strategy
for quality and access in section 4. The strategy has three mutually
reinforcing elements: flexible fees, a wide-ranging loan system
and active measures to promote access.
6. Flexible fees are necessary to reflect
diversity, to arrest quality decline and to assist some redistribution
of teaching budget towards institutions with more remedial teaching.
Specifically, fees should be increased initially to £2,000,
but with institutions free to charge less. All fees should be
fully covered by a loan entitlement.
7. A wide-ranging loan system.
Loans should be adequate to cover
living costs and tuition fees, making higher education free at
the point of use, thus addressing student poverty and freeing
students from high-cost borrowing such as overdrafts and credit
Loan entitlement should become universal,
eliminating the unpopular and complex income test and, at a stroke,
getting rid of parental contributions.
The combined effect of these twin elements is
equivalent to bringing in universal grants in combination with
an income-related graduate contribution (section 4.2). Additional
options include extending loans to students in further education
and to postgraduates.
8. Active measures to promote access. There
are two impediments to accessfinancial poverty and information
poverty. The strategy outlined in section 4.3 aims to address
Grants and scholarships for students
from poor backgrounds.
Extra personal and academic support
when students from poor backgrounds reach university.
Raising the aspirations of schoolchildren.
More resources earlier in the system,
including financial support for 16-19 year olds.
1. This paper sets out a strategy for promoting
access and strengthening quality. Though explicitly about higher
education, the arguments apply equally to the tertiary sector
as a whole. Successive sections discuss:
The many things we all agree about.
Lessons from economic theory.
Problems with current arrangements,
and key elements of solutions, including an indication of the
scale of the prize to be won.
1. WHAT WE
2. Since the finance of higher education
is controversial, it is useful to start by setting out some large
areas of unanimous agreement.
3. The problem. There is agreement, first,
about two core problems:
Students are poor because the system
of support does not give them enough to live on. Two results follow:
students have to turn to expensive overdraft and credit-card debt
and/or to extensive part-time work; and the parsimony of support
is an impediment to access for people contemplating university.
Universities are poor, creating worries
about quality. The UK would have to spend an extra £3.5 billion
per year to reach the EU average.
4. Objectives. There is also agreementstrong
and universal agreementabout two central objectives.
Improved access. The socioeconomic
mix in higher education has barely changed in 40 years. Everyone
supports widening participation in the interests of social justice,
and also for reasons of national economic performance.
Improved quality. Again, there is
no disagreement: the quality and diversity of higher education
is important for its own sake, and for national competitiveness.
5. Four propositions. Resources are clearly
key to achieving these objectives. To that end, the discussion
of resources throughout the paper is based on four propositions.
UK higher education needs substantial
additional funding for reasons of national economic performance
and because higher education is an important export industry.
Funding on the necessary scale will
not come from the taxpayer, given an ageing population, rising
health expenditure, competition from other parts of the education
sector and competitive global pressures.
Reform will therefore be ineffective
unless it can deliver an immediate and sustained injection of
private funding. The way to achieve this is through a student
loan scheme which can draw in private finance on fiscally attractive
Phasing out the interest subsidy
on the current loan scheme is essential to that end. Interest
subsidies are costly (about one-third of total lending never comes
back because of their cost), distortionary and badly-targeted.
Instead of paying an interest rate equal to the rate of inflation
(as currently), graduates should pay a rate equal to the government's
cost of borrowing (not the interest rate on bank overdrafts or
credit cards). The considerable savings would be much better used
to expand the loan system and for the explicit, targeted measures
to promote access set out in section 4.3.
2. LESSONS FROM
6. Economic theory offers three strong sets
of results, summarised here briefly.
7. The days of central planning have gone,
both for students and for higher education institutions. The system
should empower the individual choices of students and potential
students. The key theoretical question is whether students are
well-informed or can become well-informed. My answer is yes. The
role of government is not to plan student choices, but to make
sure that students have easy access to timely, accurate and relevant
information andparticularly for students from poorer backgroundsalso
8. The supply side should also be liberalised.
Forty years ago, with an elite system, it was possible, as a polite
myth, to assume that all universities were equally good and hence
could be funded broadly equally. Today we have a mass system,
meaning more higher education institutions, more students, and
much greater diversity of subject matterall changes which
are warmly to be welcomed. As a result, however, the characteristics
and the costs of different degrees at different institutions vary
widely. Thus universities need to be funded differentially. In
principle this could be done by an all-knowing central planner.
In practice, the problem is too complex for that to be the sole
mechanism. A mass system in an increasingly complex world needs
a funding mechanism which allows institutions to charge differential
prices to reflect their differential characteristics.
9. Supply-side liberalisation is not only
necessary; it is also desirable. Increased competition between
institutions will make them more responsive to student preferences.
Some students will wish to study full-time but on an accelerated
basis, for example studying for four terms per year rather than
three; others, in contrast, will wish to study part-time, for
example through evening courses. A system which can offer students
and prospective students a wider range of choice is efficient;
and the added option of part-time study while continuing to work
also assists access.
10. Graduates should contribute to the costs
of their degrees. A second strong result from economic theory
is that higher education should not be freeits costs should
be shared between the taxpayer and the graduate. There are two
mutually reinforcing arguments.
11. We cannot afford free higher education.
The argument is simple. Forty years ago, with a 5 per cent participation
rate it was fiscally feasible to rely mainly on public funding
to support a high-quality higher education system. The welcome
expansion to a 35 per cent participation rate, with aspirations
to a 50 per cent rate, however, mean that public funding has to
be supplemented on a significant scale by private funding. This
is all the more the case because:
12. We should not have free higher education.
It is well-known that graduates on average have significantly
higher earnings than non-graduates. The Dearing Report (National
Committee of Inquiry into Higher Education, 1997, paragraph 18.13)
suggests that "compared to those without higher education
qualifications who were qualified to enter higher education, those
with higher education qualifications:
have higher employment rates;
enjoy an average private rate of
return of some 11 to 14 per cent".
13. Since higher education creates social
benefits it is right that there should always be a taxpayer contribution.
But given the robust evidence on private rates of return, excessive
reliance on public funding is inefficient. It is also regressive,
and hence unjust, since the major beneficiaries of free higher
education are the predominantly middle-class participants. A government
committed to improving access should not spray scarce taxpayer
pounds indiscriminately across the entire student body but should
instead target those resources on people for whom access is most
14. A well-designed student loan scheme
has core features. The third set of conclusions from economic
theory sheds light on the design of student loans. Four features
stand out, summarised here only briefly (for fuller discussion,
see Barr, 2001, Ch 12).
Income-contingent repayments, ie
loans with repayments calculated as x per cent of the graduate's
subsequent earnings until she has repaid her loan, are fundamental.
The arguments are now well-understood. Income-contingent repayments
instantly and automatically respond to changes in earnings: people
with low earnings make low repayments; and people with low lifetime
earnings do not repay in full. The effect is to protect borrowers
against excessive risk, with gains both in efficiency and in terms
of access (see also Barr and Crawford, 1997, evidence to this
Large enough to cover all living
costs and all tuition fees. This feature makes higher education
free at the point of usethe important advance made by the
Cubie arrangements in Scotland. Students are no longer pushed
towards expensive credit-card debt; and parental contributions
can be abolished, a liberation both for students and their parents.
An unsubsidised interest rate, as
explained in section 3.3, is essential for fiscal reasons, for
efficiency reasons, and in the interests of access.
A capacity to bring in private funds.
Student loans bring in private funding through students' subsequent
repayments. However, there is a net saving to the taxpayer only
when the scheme is mature, ie when the inflow of repayments from
earlier cohorts of students matches or exceeds the outgoings to
this year's borrowers. That process takes 15-20 years. If extra
resources are needed immediately, it is desirable to have a loan
scheme which brings in private money upfront, creating an immediate
injection of private finance.
3. PROBLEMS WITH
3.1 The good news
15. Income-contingent loans, with repayments
collected alongside income tax, were introduced for UK students
starting their degrees in or after 1998. This move represents
unambiguous progress and deserves loud applause.
3.2 The bad news
16. That, however, exhausts the good news.
The problems described below were both predictable and predicted,
eg Barr and Crawford (1997) in evidence to this Committee.
17. The system of student support impedes
access in several ways.
18. Deficient loan design. The current scheme
conforms with only one of the four criteriaincome-contingent
repaymentsin paragraph 14. It fails the remaining three
The loan is too small to cover living
costs; it is income tested, so that not all students are entitled
to a full loan; and there is no loan to cover tuition fees. Thus
the system incorporates upfront charges, students remain poor,
and parental contributions continue. All these features impede
The loan incorporates an interest
subsidy. The resulting problems are discussed in detail in section
The scheme is capable of bringing
in private finance but, because of the interest subsidy, only
on fiscally unattractive terms, again, discussed further below.
19. Continued reliance on parental contributions.
The problems of parental contributions merit additional discussion.
As a philosophical matter, is it
right to force young adults to depend on their parents?
Student poverty: the scale and volume
of unpaid contributions is well known.
Impediments to access: unpaid contributions
cause some students to drop out, and the threat of unpaid contributions
deters an unknown number of others from applying in the first
Distorted choices: in other cases,
parents pay the contribution, but with conditions attached: "we
will pay, but only if you do a sensible subject."
The previous three problems all have
troubling gender and ethnic aspects, and the point is, if anything,
even stronger in respect of spouse contributions.
The income test necessary to assess
parental contributions is intrusive and has high compliance and
20. Its gets worse! Assessment of family
income has to take account of whether a student is making maintenance
payments (a deduction from his assessable income) or is the recipient
of maintenance payments (which may be an addition to his assessable
income) (Department of Education and Skills, 2001, Ch 6, paragraphs
54-58). And the relevance of a spouse's income raises the vexed
issue of cohabitation: a woman whose husband has a high salary
is not entitled to a full loan; nor is one whose partner has a
high incomebut that requires finding out whether or not
a student is cohabiting. Paragraphs B117-8 of the guidance notes
just cited are titled "Advice on identifying a cohabiting
couple". Such factorswhich should lie wholly outside
the system of student supportare an inescapable concomitant
of an income test.
21. None of this is an attack on family
support: where families wish to help, such support should be applauded.
The attack is twofold. Policy should not be based on an assumption
that parents will support their children. Such an assumption may,
at a stretch, have been valid for an elite system of higher education,
regarded as a luxury good for middle-class families; it is invalid
for mass higher education as an investment good, and totally inapplicable
to expanding access. The policy is bad also because it forces
students into dependence on parental contributions or spouse contributions,
since there is no option to take out a larger loan in place of
22. Complexity. Annex 1 gives a very simplified
explanation of the current system.
But student support in practice is so complex that nobody fully
understands the system. Someone from a poor background pays no
tuition fee and is entitled to a full loan. The assessment of
a student's financial position is based on parental income for
a younger student, or on his or her spouse's or partner's income.
Parental or spouse income has two effects: as income rises, the
tuition fee rises; once the fee has reached its maximum (£1,075
in 2001-02), the effect of additional parental income is to reduce
the size of the loan to which the student is entitled. All students,
however rich their parents or spouse, are entitled to a loan equal
to about 75 per cent of the maximum loan except that scholarship
and similar income, if high enough, can reduce loan entitlement
below that 75 per cent minimum.
Such complexity has major ill-effects: students, prospective students,
and their parents cannot understand the system; it is a nightmare
to administer; and complexity, per se, impedes access.
23. Inadequate dissemination of information.
Perhaps the greatest impediment to access is the fact that the
wider public totally fails to understand income-contingent repayments.
The scale of this ignorance cannot be exaggerated. Most people
are completely unaware that loan repayments are de facto a form
of income taxbut paid only by graduates and switched off
once the graduate has repaid what he or she borrowed. In this
respect, the Government is deeply culpable over its negligence
in vigorously and repeatedly explaining this point. The resulting
ignorance unnecessarily aggravates debt aversion and is a further
impediment to access. The topic is taken up in detail in paragraphs
24. The post-Dearing arrangements are also
bad news on the supply side.
25. Continued central planning. The strong
theoretical case against central planning of higher education
was alluded to earlier. Though nobody quarrels with the need for
universities to be publicly and transparently accountable, there
are few defenders of the particular mechanisms, of which the QAA
and RAE are only the tip of the iceberg. In addition, there has
been central control of the number of students at each university
and of tuition feesin other words, both price and quantity
were determined by the central plannera situation only
partly eased by the proposed lifting of the numbers cap.
26. Such planning impedes quality. Alsoand
entirely unintendedit impedes access to UK students to
the best universities. Again, this was predicted to this Committee:
"A flat fee will continue the erosion of
quality at the best universities, which face the biggest shortfalls.
If this policy continues, the result will be to deprive British
students of the chance of an internationally cutting-edge undergraduate
degree in one of two ways. The quality of the best institutions
might fall; British students could still get places, but the quality
of the degree would be less. Alternatively, the best institutions
will largely stop teaching British undergraduates (for whom they
receive on average £4,000 per year) and will use the fees
from foreign undergraduates (around £8,000 per year) to preserve
their excellence. The government is considering trying to prevent
British universities from charging additional fees to UK/EU students.
Again, this is done for equity reasons; again, it ends up harming
the very people it is aimed at helping, in this case by creating
a situation where British students will find it harder and harder
to get places at the best universities" (Barr and Crawford,
1997, paragraph 57).
27. Inadequate university income. The immediate
post-Dearing arrangements brought universities not an extra penny,
for the reasons explained in Barr and Crawford (1997), with worrying
effects on quality. The story in 1997, in a nutshell, was as follows:
(a) Public spending on higher education would
not go up (the budget said so).
(b) Parental contributions (ie private spending)
would not go up (the Secretary of State said so).
(c) Loans to students (the other potential
source of private spending) counted in their entirety as public
28. There has been some improvement since
1997. Public spending under (a) has increased; debt sales have
brought in some private money under (c); and the move from cash-flow
to resource accounting has further assisted under (c). These developments
are all genuinely welcome; but they do not change the reality
that, at its core, the system continues to be publicly funded;
and given its greater political salience, student support has
crowded out university income. The story in Australia is exactly
the same (see Annex 3, paragraph 95), and for exactly the same
underlying reasonthe interest subsidy on student loans.
3.3 The worst news: interest subsidies
29. It is important to understand the scale
of the problems that interest subsidies cause.
30. What interest rate? First, it is important
to be clear what I am not saying:
At present graduates pay an interest
rate equal to the rate of inflation. Press discussion of "market
interest rates" evokes worries about high interest rates
associated with credit cards and overdrafts. That is not what
is meant. The interest rate which graduates should pay on their
loans is the government's cost of borrowing, ie broadly the interest
rate the Monetary Policy Committee announces.
The attack is on blanket interest
subsidies. A strong case can be made for targeted interest subsidiesfor
example someone who is unemployed or caring for young children
or other dependantsto make sure that their debt does not
spiral upwards. Mechanisms for such targeted assistance are discussed
in section 4.2.
31. Why are interest subsidies such a problem?
Interest subsidies are:
Regressive. As explained in paragraphs
42 and 43, interest subsidies do not benefit students (who do
not make repayments), nor low-earning graduates, but better-off
graduates in mid-career.
Expensive. Interest subsidies are
enormously costly. Evidence from debt sales suggests that of all
the money lent to students, about one-third never comes back because
of the cost of interest subsidies. The scale of the resulting
losses is discussed shortly. The high fiscal costs of loans create
two further sets of ill-effects.
Inimical to quality. Expensive student
support crowds out university income; thus interest subsidies
conflict directly with improved quality.
Distortionary. Because loans are
so expensive, the Treasury rations them. Thus interest subsidies,
like most subsidies, create shortageslike rent control,
they end up harming the very people they were meant to help. There
was an experiment with price subsidies called Communism. It did
32. In the case of loans, shortages manifest
themselves in the following ways:
The full loan is too small to cover
a student's living costs, leading to student poverty.
Loans are means-tested: as a result
parental/spouse/partner contributions continue, and higher education
involves upfront costs and charges.
Loans are restricted, for example
are not available to students in further education (impeding access),
nor to postgraduates (putting national competitiveness at risk).
33. In short, interest subsidies create
a fiscal black hole which aggravates problems both of access and
A move to an unsubsidised rate is not just a technicalityit
is the single essential key to solving current problems of funding
tertiary education. The considerable resources thereby released
underwrite the array of major policy advances set out in section
34. How much extra money? Suppose graduates
pay an interest rate equal to the government's cost of borrowing
rather than, as now, the inflation rate.
Do the resulting savings make the move worthwhile?
35. Barr and Falkingham (1993, 1996), using
LIFEMOD, a microsimulation model, found that for every 100 the
government lends, only about 50 is repaid. Of the missing 50,
20 is not repaid because of fraud, early death, and emigration
(all of which have a relatively small effect), and mainly because
some graduates have low lifetime earnings and so never repay their
loan in full, and 30 is not repaid because of the interest subsidy.
In other words, the interest subsidy converts nearly one-third
of the loan into a grant.
36. Previous sales of student debt offer
independent evidence. The debt was sold for about 50 pence per
pound of its face value. Official estimates suggest that of the
missing 50 pence about 15 pence was because of low lifetime income,
etc, and 35 pence because of the interest subsidy. The evidence
on the interest subsidy is compelling. The government did not
use LIFEMOD; thus the official estimates and the simulation results
reinforce each other.
37. A recent modelling exercise
offers further confirmation. The model is conservative in at least
two ways. First, it assumes that the subsidised interest rate
is 2.5 per cent, the unsubsidised rate 5 per cent, thus implicitly
assuming that the real rate will stay as low as 2½ per cent
throughout the lifetime of the loan. Second, the accounting of
debt forgiveness is on an accruals basis, even though forgiveness
normally occurs only at the end of the loan period.
The model takes as its starting point projected loan outgoings
for the next academic year of £2,500 million.
With debt forgiveness of 15 per cent (ie assuming that 15 per
cent of student borrowing will never be repaid because of low
lifetime earnings, etc),
a move to an interest rate equal to the government's cost of borrowing
would release sustainable savings of £700 million per year.
That sum could be used to finance additional grants and scholarships
of up to £700 million. Alternatively, it could be used to
underwrite an expanded loan system; with 15 per cent debt forgiveness,
it would be possible nearly to triple the total amount of lending.
Or the sum could be used for a combination of the two policies.
38. When will the money be available? It
is necessary at this stage to distinguish the cash-flow costs
of loans (ie the money which is required now, but which will eventually
be repaid), from the fiscal costs (ie borrowing which will never
be repaid). With the present loan scheme, for each 100 that students
borrow (the cash-flow cost), roughly 50 will never be repaid (the
fiscal cost). The argument above is that eliminating interest
subsidies reduces the fiscal cost of loans from about 50 per cent
to about 15 per cent. Though the Treasury will, of course, take
cognisance of cash-flow costs, the abolition of interest subsidies
makes it possible to have a loan system that is larger, but at
the same time has smaller fiscal costs.
39. Seen through the eyes of the Treasury,
£700 million is the present value of the annual saving, and
is thus a significant and sustainable long-run resource. Seen
through the eyes of the Department for Education and Skills, however,
the initial savings in cash-flow terms are small.
But higher education needs to benefit immediately from the long-run
savings. That will require a deal between the Treasury and the
Department for Education and Skills for an early injection of
additional resources. One way to finance such a deal is by selling
a further tranche of student debt, which could yield up to £2
billion. For the reasons explained in the previous paragraph,
such a deal makes sense in both educational and fiscal terms.
40. Ensuring low fiscal costs is, of course
essential. Once that is done, the Treasury can choose how to deal
with the cash flow costs of the loan system. It could do so out
of taxation: under resource accounting, the fiscal costs (ie lending
that is not expected to come back) appears as current education
spending, while expected repayments appear in the capital account
as a financial asset.
With the present loan scheme, about half of lending to students
is counted as spending out of the education budget. If there were
no interest subsidy, only 15 per cent, or so, of total lending
would appear as current education spending, and the 85 per cent
expected repayment would appear in the capital account.
41. Alternatively, the Treasury could deal
with the upfront cash flow costs by bringing in private money
to finance the scheme. This can be done in various ways. The debt-sale
approach has been extensively discussed (Barr and Falkingham 1993,
1996). There has been less exploration of front-end funding, which
has two variants. With retail lending, individual students borrow
from private lenders; thus student borrowing is individualized.
With wholesale lending, the loans administration borrows private
money in tranches of (say) £2 billion, which it then lends
42. Grasping the nettle. The scale of the
prize is clearly enormous. But political worries about raising
interest rates persist. These, however, should not be exaggerated.
There is already growing support
for removing interest subsidies (eg Piatt and Robinson, 2001).
Interest rates are currently low;
thus the interest rate would have rise by no more than 2½
per cent from the rate in the current loan scheme.
A person's monthly repayments depend
only on her income; thus interest rates have no effect on monthly
repayments, but only on the duration of repayment. Once such a
scheme has been introduced, this feature will become obvious to
The wider public should understand
who benefits from interest subsidies. Interest subsidies do not
help students (it is not students who make loan repayments, but
graduates); they do not help low-earning graduates, since unpaid
debt is forgiven after 25 years; they do not help higher-earning
graduates early in their careers (since monthly repayments are
a fraction of earnings, and so are not affected by the interest
rate); the only people they help are higher-earning graduates
in mid career, whose loan repayments are switched off earlier
because of the interest subsidies than would be the case without
43. In sum.
Interest subsidies are targeted with
exquisite accuracy: they benefit successful professionals in mid
career. Thus the NUS position, defending interest subsidies, is
arguing for continued subsidies for those who need it least at
a time when they need it least, and the hell with today's struggling
Removing the interest subsidy finances
policies to promote access and quality, thus helping students
rather than graduates. The proposal is not to eliminate subsidies,
but to replace blanket (ie untargeted) subsidies by targeted interventions.
Income-contingent loans are simply
a form of income-tax; and income-tax, together with a write-off
after 25 years, bases loan repayments on outcomes, and hence targets
subsidies accurately on graduates with low lifetime earnings.
Once students and their parents come to understand
these point, their worries and, with them, much of the political
hullabaloo, will recedean issue discussed in detail in
4. THE POLICY
44. This section sets out a strategy for
access and quality with three elements:
flexible fees, to address the quality
issue, and to begin to free universities from unnecessary central
a wide-ranging loan scheme to empower
choice for the generality of students;
wide-ranging but targeted measures
to promote access; these need to start early in the school system.
45. The three elements are a strategy, not
just a bunch of ad hoc policies: they are designed to achieve
explicit objectives, and the elements are mutually reinforcing.
This does not mean that the policies below must be swallowed whole,
but does mean that attempts indiscriminately to pick and mix will
fail to achieve the policy's objectives.
4.1 Flexible fees
46. A medium-term aim. Flexible fees are
both necessary and desirable for at least three reasons.
47. To address diversity. Historically,
with a small tertiary system and a limited range of subjects,
it was possible for central planners to determine funding levels
for different institutions. Today, however, the higher education
system is large, diverse and complex. As a result, (a) the necessary
variation in funding is much greater than formerly and (b) the
problem is now too complex for a central planner to have the sole
power of decision about how resources should be divided between
institutions. Thus institutions should have the freedom to set
their own fee levels (that freedom could be constrained).
48. To arrest quality decline. Without higher
fees, quality will continue to be eroded and, given flat fees,
eroded most at the best institutions.
49. To prevent crowding out. At present,
the best institutions tend to receive more funding for teaching
in an attempt to protect their quality. This risks crowding out
universities with a different mission. With higher fees, the best
institutions could paddle their own canoes to a greater extent
than currently, freeing resources for institutions which have
to do more remedial teaching. Thusan important and poorly-understood
pointflexible fees benefit all tertiary institutions.
50. Arguments that do not stand up are of
two sorts. It is sometimes argued that students should pay the
full costs of tuition. That argument overlooks the significant
(albeit hard-to-quantify) external benefits of higher education.
There is a strong case for continuing taxpayer subsidy for tuition
for all students. Other people argue the oppositethat tertiary
education should be entirely funded from taxation. There may be
a case for that policy for sub-university education, but, applied
to higher education, the argument is flawed. The argument that
university education should be free at the point of use (which
I support), does not mean that there should be no charges. Free
tuition is expensive, and its benefits accrue disproportionately
to people from better-off backgrounds, who go on to be among the
best-off. Thus free tuition is badly targeted; the money could
do much more for access if spent in a way that directed resources
specifically at those groups for whom access is most fragile and
those who do not benefit financially from their degrees. Lowering
tuition charges for higher education in other countries has not
improved access. Conversely, introducing the Higher Education
Contribution Scheme in Australia in 1989, with tuition charges
paid via an income-contingent loan, did not harm access (Chapman
51. How high should the fee be? Taxpayer
support plus tuition fees should cover the costs of teaching.
At research-intensive universities, fees should be higher to the
extent of the value-added in teaching by research-active people,
but should not cross-subsidise research.
52. Pitfalls to avoid. Though flexible fees
are desirable, there should be no "big bang" liberalisation.
As a first step, the flat fee should be increased to £2,000.
However, institutions should be free to charge less if they wish,
and the ceiling should be raised over time. As indicated below,
any increase in the ceiling should be accompanied by an equal
increase in the loan entitlement.
4.2 A wide-ranging income-contingent loan
53. The second element in the strategy is
a wide-ranging loan scheme. Discussion starts with policy options
and then turns to a much-neglected aspect of loanshow to
make sure that the public understands the nature of the beast.
54. As explained in section 2, a well-designed
loan scheme has four characteristics: income-contingent repayments;
sufficient to cover all living costs and all tuition fees; with
an interest rate equal to the government's cost of borrowing;
and capable of bringing in private funds. A loan scheme of this
sort opens up the following options.
55. Universal. If there is no interest subsidy
there is no need to ration loans, which can therefore be made
available as a universal entitlement. The administratively complex
and politically unpopular means test disappears; so does the relevance
of such factors as whether a student is married or cohabiting.
56. Adequate. For the same reason, the loan
entitlement can be large enough to cover realistic living costs
and all tuition fees.
A universal, adequate loan has major advantages.
It eliminates student poverty.
It makes higher education free at
the point of use.
It makes it possible to abolish parental
It frees students from forced reliance
on expensive credit card debt and/or the need for extensive part-time
work, and thus in substantial measure addresses the worries of
middle-class students and their parents, and also the (rightful)
complaints by the NUS about the amount of high-cost student debt.
It is simple for students and their
parents to understand.
It is vastly simpler to administer
than current arrangements, since the administratively cumbersome
and unpopular means test can be abolished.
There is no need to belabour the helpful political
resonances of all these features.
57. Extending loans to other groups of students.
The scheme could be extended to postgraduate students, starting
to address worries about research capacity and national productivity.
It could also become a universal entitlement throughout tertiary
education, buttressing existing sketchy student support for further
education and vocational training, thus contributing to access.
58. Protecting low earners is a clear priority.
Income-contingent repayments do so automatically, since low earners
make low repayments, and people with low lifetime earnings do
not repay in full. Nevertheless, many people are afraid of rising
debt, and interest subsidies were introduced to assuage those
fears. As argued earlier, however, interest subsidies are a costly,
non-transparent and ineffective way of promoting access, and benefit
exactly the wrong people. The policies below are more transparent
and better targeted ways of protecting low earners.
59. Scholarships, ie grants which do not
carry an obligation to make income-contingent repayments.
60. Stopping repayments after 25 years.
Scholarships help people at the start of the process. Complete
debt forgiveness after (say) 25 years helps them at the end of
61. Targeted interest subsidies based on
current income. It is also possible to help people during the
process. Under the simplest mechanism, anyone receiving a credit
for national insurance contributionssomeone who is unemployed
or looking after young childrenreceives an interest subsidy.
As an extension (as in New Zealand), anyone whose earnings are
so low that his or her income-contingent repayment fails to cover
the interest element, similarly receives an interest subsidy.
62. Conditional subsidies. The mechanism
in the previous paragraph protects low earners and people with
career breaks. However, some of the subsidies benefit people who
subsequently have high earnings. Thus there are advantages in
terms of fiscal cost and targeting to have a system which pays
interest subsidies based on current income, but with a facility
to claw back the subsidy element for people who end up with high
lifetime income. A workable such scheme has been designed, which
the Student Loans Company could administer without difficulty.
Such a scheme has major advantages: it gives assistance at the
time they need it to people with low income and/or with a career
break, facilitating access and preventing people from worrying
that their repayment obligation is spiralling upwards; but it
is cost free, since the clawback mechanism ensures that only the
lifetime poor keep the subsidy. It is also flexible: it would,
for example, be possible to give conditional subsidies to some
groups (eg someone who went travelling after graduation), but
unconditional subsidies to others (eg someone looking after young
children or elderly dependantsanalogous to current proposals
to assist teachers with their loan repayments).
63. Past failures adequately to explain
income-contingent loans has created unnecessary disquiet. Medical
practitioners sometimes talk about "the worried well"people
who are in good health, but whose life is made less happy by misplaced
worries that they are not well. Analogously, it can be argued,
"worried debtors" are concerned that student loans will
be a millstone. Students perceive a large debt, but miss two important
points of context. First, repayments are merely an addition to
their future income tax. Thus the risk they take is no different
from the risk we all face that at some time in the future the
basic rate of income tax might increase (I am old enough to remember
a 33 per cent basic rate). Second, they do not see how much they
will pay over the years in income tax or national insurance contributions,
nor what they will spend on food (or drink) over a 25-year period.
None of these items cause people to worry, despite the fact that
someone who starts on average earnings and remains on average
earnings all his working life, will pay nearly £300,000
in income tax and national insurance contributions over a 35-year
career, a topic discussed more fully in Annex 2.
64. Alongside the policy design in the previous
paragraphs, it is thus imperative to explain what is going on.
I am not at this stage digressing into political presentation,
but want instead to point out some analytical equivalences which
should inform explanation.
65. The following schemes are all analytically
66. Scheme 1: An income-contingent loan.
Earlier discussion described an income-contingent loan, for which
graduates make repayments of x per cent of their earnings until
they have repaid the loan at an interest rate equal to the government's
cost of borrowing. What has been missing thus far in government
action is a clear explanation that a person's loan repayments
are, in effect, a form of income tax: a person's repayments track
changes in his or her earnings instantly and automatically, and
thus nobody repays more than he or she can afford.
67. The point is fundamental. Suppose we
start from the argument (NUS, etc) that higher education should
be paid out of taxation. But that means that the costs of higher
education are paid by the generality of income recipients, including
low earners, non-graduates, pensioners, and the like. It is widely
acknowledged that this is unfair. To deal with that unfairness,
some people have argued that higher education should be financed
through a graduate tax, which can be thought of as additional
income tax, but paid only by the beneficiaries of higher education.
But a graduate tax has its own unfairness, since people with high
lifetime earnings repay considerably more than they have borrowed
(this is true of any successful professional, without having to
refer to extreme cases like Mick Jagger or Stelios Haji-Ioannou).
To address the latter problem, it is suggested that a graduate
tax should not have an indefinite duration, but should be capped,
ie "switched off" once a person has repaid an agreed
contribution towards the costs of his/her degree. An income-contingent
loan is exactly thatit is a graduate tax which is capped
at 100 per cent of the initial sum borrowed. Put another way,
the only difference between a graduate tax and an income-contingent
loan is the duration of repayment: with a graduate tax, repayment
lasts (say) 25 years, regardless of how much a person has borrowed
or repaid; with an income-contingent loan the duration of repayment
bears an exact relation to the initial amount borrowed.
68. Thus income-contingent repayments can
be described as repayment of debt. But with equal accuracy they
can be described as a form of targeted income taxtargeted
by being imposed only on graduates and by being "switched
off" once the loan has been repaid. This is a form of taxation
that is more efficient and fairer than funding via the entire
body of taxpayers.
69. Scheme 2: Free higher education
(CUBIE). If loans are universal and adequate to cover all living
costs and all tuition fees, higher education is free at the point
of use. Thus the system can be described as free higher education,
paid for by a targeted income tax or a graduate contribution.
To repeat, the only difference from tax funding is that (a) the
tax is paid only by those who have been to university and (b)
the additional tax is capped.
70. Scheme 3: Universal grants plus
a graduate contribution. Suppose that all students receive a grant
large enough to cover their living costs and tuition fees; and
suppose that the system is financed by an extra x pence in the
pound added to a person's income tax rate. The contribution does
not go on for ever: it is "switched off" once a person
has contributed an amount equal to the grant he/she received;
and nobody pays for more than (say) 25 years. Thus the scheme
is simple: there is a universal grant, financed by a graduate
contribution payable for a maximum of 25 years, but for some people
a shorter time.
71. Two additional points specify the scheme
The duration of the graduate contribution
bears an exact relation to the amount of grant a student has received.
Though administratively complex, this is precisely the task that
the Student Loans Company currently performs: higher earners repay
their grant more quickly; lower earners take longer and are further
protected in the ways outlined in paragraphs 58-62; and nobody
makes contributions for more than 25 years.
Students for whom access is fragile
are helped in the ways set out immediately below. The scholarships
and other forms of financial help they receive do not make them
liable for a graduate contribution.
72. Since the three schemes are equivalent,
any is an accurate description of the system of student support.
If this were better understoodparticularly the very close
relation of income-contingent repayments to tax fundingthe
problem of the "worried well" would be greatly diminished.
73. A point in conclusion: the equivalence
arguments in the previous paragraphs are analytical and should
be judged on the quality of their logic. They are not presentational
arguments based on logical thin ice.
4.3 Active measures to promote access
74. Income-contingent loans measure ability
to pay on the basis of where a person ends up, ie his or her subsequent
income. This is the best approach for students from better-off
backgrounds, who are generally well-informed about the benefits
of tertiary education. However, students from socially-excluded
backgrounds are typically badly informed. Precisely for that reason,
a wide range of additional measures to promote accessthe
third leg of the strategyis necessary.
75. Put another way, it is not only financial
poverty which impedes access, but also information poverty. Any
strategy for access therefore needs two elements:
those which involve money, and
those which involve information and
The following are no more than indications of
the scale of necessary actions.
76. Grants and scholarships in higher education.
Money measures include scholarships for students from poor backgrounds.
They could be based on parental income, but should also include
money for schools and universities to award to students from poor
backgrounds. There should be financial incentives to universities
to widen participation; and universities would, in any case, wish
to gather resources for scholarships to enable them to recruit
the brightest students, regardless of their financial background.
77. For some students the biggest hurdle
is to realise that they are good enough to do well in higher education:
for them, scholarships which make their first year entirely free
give them a risk-free opportunity to test their abilities and
to become well-informed about higher education. Once such a student
does well in the first year, he or she will be much more prepared
to take out at least a partial loan for the rest of the degree.
78. Extra personal and intellectual support
in higher education. A second ingredient in promoting access is
extra personal and intellectual support, at least in the early
days, for access students to make sure that, once a student starts
at university, he or she gets the necessary support to make the
transition. It is no good persuading someone to go to university
if he or she is then allowed to sink without trace.
79. Raising aspirations of school children.
Action is also needed much earlier. Information and raising aspirations
are critical. The saddest impediment to access is someone who
has never even thought of applying to university. The sorts of
schemes involved include Saturday Schools, which bring schoolchildren
from poor areas to university to study on Saturday mornings; summer
schools, which do something similar during the summer vacation;
visit days, when schoolchildren can visit a university; visits
by academics to schools to make the idea of higher education more
tangible; visits by current students, ideally from the same or
a similar school, to schools in deprived areas; and mentoring
of schoolchildren by current university students, preferably from
a similar background. Such activities needs to start early, eg
at age 12.
80. More resources earlier in the system.
Problems of access to higher education cannot be solved entirely
within the higher education sector. Thus more resources are needed
earlier in the education system, which is where the real barriers
to access occur.
This includes more resources for teaching. It also includes financial
support for 16-18 year olds.
81. At risk of sounding repetitive, the
root of all the funding problems of the present system go backdirectly
or indirectlyto the interest subsidy on student loans.
Interest subsidies aggravate the
shortage of resources in the sector, compete with university funding,
cause loans to be inadequate, crowd out expenditure to promote
access, and are deeply regressive.
Setting the interest rate that graduates
pay equal to the government's cost of borrowing is the key to
addressing all these problems and, on a conservative estimate,
yields some £700 million per year. That sum is important
for its own sake; it is even more important because a fiscally
cost-effective loan scheme can leverage a much greater volume
of private finance. Specifically, with no increase in long-run
fiscal costs, it would be possible nearly to triple total lending,
or to combine a smaller expansion of loans with additional grant
Grasping the nettle makes possible
the policies for access and quality set out in section 4.
The move is less difficult than it
soundsperhaps the political equivalent of the "worried
well". In Sweden and the Netherlands graduates pay an interest
rate broadly equal to the government's cost of borrowing, and
the matter is not regarded as in any way noteworthy; and Australia
and New Zealand are currently facing the same reality as the UK.
In contrast, if the interest subsidy
is not eliminated, we might as well all pack our bags and go home.
1 This paper draws on Barr (2001, Chs 10-14), and
in part on work while a Visiting Scholar at the Fiscal Affairs
Department at the IMF in Spring 2000. It also draws on collaboration
on policy design with Iain Crawford for more years than either
of us care to contemplate, on advice on factual matters and administrative
feasibility from Colin Ward and his team at the Student Loans
Company, and on recent work by the three of us on a project advising
the Hungarian government. An earlier version was presented at
a meeting of the Parliamentary Universities Group. Back
Department of Economics, London School of Economics and Political
Science, Houghton Street, London WC2A 2AE, UK. Back
For fuller discussion, see Barr (2001, Chs 11 and 12). Back
For assessment of systems of higher education finance in other
countries (the USA, Australia, New Zealand, the Netherlands and
Sweden), see Barr (2001, Ch 13). National Audit Office (2002)
reaches very similar conclusions about the problems of the present
system, in particular its failure to improve access. Back
Barr and Low (1988), using data for 1982-3, found that about
half of students entitled to parental contribution received less
than they were supposed to, and the shortfall was substantial:
students whose parents gave them less than the system supposed
received only £53 of every £100 of assessed parental
contribution. As a result, one student in thirteen remained below
the poverty line even when income from all sources was included.
Subsequent work based on 1992-3 data found that 37 per cent of
students received less than the assessed parental contribution
(Committee of Vice-Chancellors and Principals (1996, p 14), quoting
an official survey). Callender and Kemp (2000, p 3) report that
"By 1998/9, the proportion of students who failed to receive
their full assessed parental contribution had doubled to three
in ten students. The mean shortfall for these students (average
assessed contribution minus average actual parental contribution)
was £719". Back
Without wishing to seem frivolous, I challenge Committee members
to explain the operation of the income-test by which a student's
loan entitlement and tuition fee are assessed, as set out in the
guidance notes from the Department for Education and Skills (2001)
to the Local Education Authorities, who administer the income
Chapter 6). Back
Originally, a student's loan entitlement was reduced pound for
pound with any scholarship income in excess of £1,000 per
year. The disregard was subsequently increased; in 2001-2 it is
£4,000 (Department for Education and Skills, 2001, Table
"Today the Government announce a new deal for higher education,
involving new funding for universities and colleges, free higher
education for the less well-off, no parent having to pay more
than at present and a fair system of repayment linked to ability
to pay" (Hansard (Commons), 23 July 1997, col. 949)
(emphasis added). "Our response to Dearing ensures that fees
and maintenance together do not place an increased burden on middle-income
families" (ibid., col 950). Back
Other countries are reaching a similar conclusion. New Zealand,
having flirted with interest subsidies since 2000, are contemplating
reversing that short-lived, ill-advised experiment. A government
report published last November, concluded that: "Participation
goals should continue to be supported through a Student Loan Scheme
with income-contingent repayments as at present. The Commission
believes, however, that the current policy of writing off interest
on loans for full-time and low-income students while they are
studying is not an effective use of the government's resources.
While this policy has decreased the length of time taken to repay
loans after graduation, it has also led to an increase in the
number of students taking out loans and in the overall level of
student debt. To compound matters, the policy has made it possible
for learners to borrow money and invest it for private gain (arbitrage).
Consequently, the Commission believes that this policy should
be discontinued-or that, as a minimum, the incentives for arbitrage
should be removed. Any savings accruing to the government as a
result of modifying the current loan scheme should be reinvested
in the tertiary education system and be used for the benefit of
students" (New Zealand Tertiary Education Advisory Committee,
2001, p 14). Back
Charging a market interest rate on income-contingent loans raises
issues under the Consumer Credit Act. Solutions to the problem
are outlined in Annex 4. Back
The model was originally developed by the Chief Executive of
the Student Loans Company as part of a joint LSE/SLC project advising
the Hungarian Government, and has been adapted to simulate alternatives
to the UK system. Back
The calculations assume that a portion of the total debt forgiveness
occurs each year, when repayments are due. This gives a higher
assessment of the cost of forgiveness than one which reflects
the real life position, where the level of forgiveness is known
only at the end of the repayment period. The cost is higher because
discounting for the loss of purchasing power over time is greater
in the earlier years of repayment. The approach was adopted because
it represents the most punitive accounting approach that can be
taken for costs, depending on the resource accounting policy applied
by the Treasury. Back
This is the forecast figure for 2002-3 including income-contingent
loans, mortgage loans and hardship loans (the latter two being
small). The outturn for 2000-1 was £1,840 million, the forecast
for 2001-2, £2,282 million. Back
New Zealand, with longer experience of income-contingent repayments,
uses an official estimate of 10 per cent in its public accounts. Back
The gross annual saving from a move to an unsubsidised interest
rate is £X million, considerably larger than £700 million.
The figure of £700 million is the answer to the following
question: the government wants to finance grants by selling bonds
which are repaid after n years, where n is commensurate with the
maximum duration of the student loan; if repayments, including
interest, over n years is £X million, what is the maximum
face value of the bonds? The answer is £700 million-the amount
that can be spent on grants allowing for the cost of financing
those grants. Thus the figure is a very conservative one. Back
To oversimplify, if the interest rate students pay rises by 2.5
per cent, the extra interest in year 1 is 2.5 per cent x £2,500
million, ie £62.5 million. In year 2, the saving in cash-flow
terms are double that sum, in year 3 triple, etc. Back
Resource accounting has been introduced in the UK only recently.
For a description of the way student loans are treated in the
public accounts in New Zealand, which has had resource accounting
for longer, see Barr (1997), evidence to this Committee. Back
There has been extensive development work on front-end funding
(see Barr, 2001, Ch 14). The result is not just academic theory,
but has been extensively tested with the IMF, Eurostat, and financial
market actors. Back
See, similarly, Piatt and Robinson (2001) and Council for Industry
and Higher Education (2001). Back
Extending loans to cover tuition fees raises issues under EU
legislation. These are discussed in Annex 4. Back
In today's prices, assuming annual real earnings growth of 2Ö
per cent. Back
As explained in Annex 3, a graduate tax has problems in addition
to being unfair: it leads inescapably to continued reliance on
public funding; it is a hypothecated tax; and it raises difficult
boundary problems. Back
Educational exclusion has a clear connection with broader social
exclusion-for example, less than 25 per cent of children in care
in the UK end up with any formal educational qualification. Back