Take three students. One is in her first term at Bath Spa University, living away from home for the first time. She comes from a family with a household income of just under ?25,000, so she is entitled to a loan to cover the cost of her fees, a government grant of ?2,835 and a maintenance loan of ?3,365 towards her living costs. She also receives a bursary of ?313 from her university.

Eighty miles away, another student is starting his first degree at the University of Oxford. Because his family income is also less than ?25,000, he is entitled to the same level of loan and grant as his Bath Spa peer, but his university offers him a bursary of ?3,150 a year - ten times the size of the one received by the Bath Spa student.

Meanwhile, at Lancaster University, an undergraduate from a well-off background is not entitled to a maintenance grant, but she has a loan to cover the cost of her fees, and one of ?3,470 for living costs. And because she exceeded the BBB offer her university made to her, gaining AAB in her A levels, she has been awarded a scholarship worth ?1,000.

It is now two years since the introduction of top-up fees, and despite the Government's plan for a fees market, Greenwich and Leeds Metropolitan are the only universities in England charging less than the full ?3,145.

ÖйúAƬ

But new research by Claire Callender, professor of ÖйúAƬ policy at Birkbeck, University of London, suggests that a diverse and competitive market has opened up elsewhere - in scholarships and bursaries. Her findings cast doubt on whether bursaries are always being used to widen participation and help students from poor backgrounds into ÖйúAƬ.

Callender's research shows:

a complex system, with some 303 separate bursary and scholarship schemes in operation at 117 ÖйúAƬ institutions in England;

ÖйúAƬ

wide disparities in the size of bursaries offered to the poorest students, with those for students in receipt of a full government maintenance grant varying from ?300 to ?3,500 a year depending on which university they attend;

40 per cent of bursaries and scholarships are not means-tested;

one quarter of all bursaries are based exclusively on merit.

Universities that charge more than ?2,835 must offer a statutory minimum bursary of ?310 for low-income students. In 2006-07, 12 per cent of universities offered the minimum. However, the vast majority - 88 per cent - chose to offer students additional money through their bursary schemes.

In 2006-07, three in ten ÖйúAƬ institutions had just one award scheme, but the average for the rest was more than three.

Significantly, while 60 per cent of bursaries included some form of means-testing, 40 per cent of bursaries did not take students' financial background into account.

For example, a student could qualify solely on the basis of academic or sporting prowess, or because he or she had studied locally at a partner college. At other universities, bursaries are automatically offered to all students at a flat rate, whatever their background.

Callender, who carried out the research last year while a Fulbright scholar at Harvard University, found that bursary schemes were most likely to be needs-based at the most selective universities and at those with the lowest proportions of state-school students.

ÖйúAƬ

Among those in the top ten of The Times Good University Guide 2007, 89 per cent of bursaries offered were needs-based, while among the Russell Group of research-intensive universities the figure was 77 per cent. On average, these institutions also offered bigger bursaries to the poorest students.

Among those in the top ten of The Times league table, the average bursary for students in receipt of a full maintenance grant was ?1,794, and at Russell Group institutions ?1,791. Meanwhile, the smallest bursaries - averaging ?642 a year - were offered by universities ranked lower in the league tables. Among the Million+ group of new universities, the figure was ?680. On average, the poorest students at the most selective universities received nearly three times more institutional aid than peers at the least selective.

Callender found that bursaries that were not necessarily linked to financial need were more common at ÖйúAƬ colleges (where 63 per cent of bursaries were non-means-tested), among the Million+ group (55 per cent), at universities with lower entry qualification scores (57 per cent), and where students were most likely to come from poor backgrounds (49 per cent).

Fourteen per cent of bursaries aimed to encourage students to complete their degrees by increasing the amount of money they received in each year of study. London South Bank University, for example, gives all students ?500 in their first year, ?750 in their second and third plus a ?250 "graduation bonus".

One third of bursaries and scholarships in England were linked to merit. These included those scholarships for students who had achieved good grades at A level or in their first year of university, or in a particular subject.

Although a minority of merit-based scholarships took financial need into account, the majority (78 per cent) did not, meaning that a quarter of all bursary and scholarship schemes were based solely on merit. These were most likely to be found in universities in the middle to upper ranks of The Times league table (50 per cent) and in old universities (45 per cent). Nearly 15 per cent of bursaries were targeted at local students.

Callender's research raises fundamental points about the bursary system, calling into question what bursaries are for. Are they about giving financial support to students from poor backgrounds during their time at university, or about universities influencing certain types of student to apply to their institution so that it can gain a market advantage? Are institutions acting in their own interest when they devise bursary schemes? Are they responding to government objectives? Is the system fair, or are some students losing out?

According to the National Union of Students, it is unjust that a student from a poor background at one university can get a bursary thousands of pounds bigger than a student at another.

Wes Streeting, president of the NUS, says: "Student support should be based on what students need, not where they study."

However, new universities argue that because they have a higher proportion of poor students, they cannot afford to offer more.

Official figures show that in 2006-07, the 1994 Group spent 18.3 per cent of their combined additional fee income on support for lower-income students. For the Russell Group, the figure was 21.2 per cent, while Million+ spent 21.8 per cent.

"I have likened the inequity of the system to the size of your pension being decided by the number of pensioners at the post office where you collect it. It is completely illogical," says Les Ebdon, chair of the Million+ group.

The most selective universities can "afford to be more magnanimous", as Callender puts it, because they have far fewer low-income students to support. And selective institutions, such as those in the Russell Group, argue that they need to be able to offer bigger bursaries because they have bigger barriers to overcome if they are to meet the government objective of getting more students from poor backgrounds into their institutions.

Wendy Piatt, director-general of the Russell Group, says there are good reasons for the elite institutions acting as they do. "We need to make sure that students aren't being deterred by the preconceptions and misconceptions about Russell Group universities being for 'a certain set' of students. We need to use any lever we have to persuade students to look at our universities and to say 'look, our universities are for you'."

But the NUS and Million+ say bigger bursaries do not help selective universities broaden their intake, and a recent report from the ÖйúAƬ Policy Institute appears to support their view.

The report, Financial Support in English Universities: The Case for a National Bursary Scheme, says "it is far from evident" that spending on bursaries is an effective way to increase institutions' socio-economic diversity.

"Between 2005 and 2006, the Russell Group and 1994 Group's 'market share' of undergraduate entrants from lower socio-economic groups remained at around 26 per cent," it says, although it acknowledges that bursaries may have a more significant impact as students' awareness of variations in the market increases.

Piatt says it is too early to tell, but she cites a study from the University of Nottingham that found that 80 per cent of lower-income entrants in 2006 were influenced in their decision to choose Nottingham by its bursary scheme. The proportion of low-income students at Nottingham rose from 17.9 per cent in 2005 to 22.4 per cent in 2006.

While selective universities appear to be focusing their bursary schemes predominantly on widening access, Callender's research suggests that, in contrast, institutions with high proportions of low-income students used their bursaries for other purposes.

"Often the institutional aid was used as part of a general recruitment strategy, as a means of attracting more students to their HEI (ÖйúAƬ institution) to maintain student numbers and to fill potentially empty places," says her paper, which is to be published in the Oxford Review of Education. She points to universities that offer bursaries to all new entrants as an example.

In these cases, institutional aid was used "as a competitive recruitment tool and as a means of marketing their institution", she says.

Ebdon says universities focused on widening participation offer a greater range of bursaries than those that are purely needs-based because they so easily "pass the Offa (Office for Fair Access) test".

It is easier to convince some students to apply for a scholarship linked to academic or sporting achievement than for a bursary linked to financial need, he says. "We are a proud people and nobody wants to put their hand up and say, 'I'm particularly poor.'"

In an effort to widen participation, many universities use bursaries to attract students from their local areas and regions.

"Bursary provision was meant to broaden students' choice of HEI. However, such provision locks students into their local institution, potentially narrowing and limiting their choice and even their HE experience," Callender's paper argues.

Ebdon, however, strongly defends locally targeted bursaries, arguing that those that are awarded to students from local partner colleges are helping many to progress.

"The choice for students from many poorer homes is their local university or no university," he says.

"Students who have childcare responsibilities, who may be carers, who have no car and can't afford public transport, can't afford to go across the country to university. Participation rates are closely correlated with access to a local 'access' university."

The most contentious bursaries and scholarships are those based on merit. Callender's paper suggests that their predominance in middle-ranking universities indicates that they are an attempt by these institutions to scale the league tables, which use high entry grades as an indicator of quality.

"In these HEIs, financial incentives were used to attract the brightest and best students to improve the quality of student intake. They aimed to enhance the reputation of the university and improve league table rankings by attracting excellent students," she says.

ÖйúAƬ

Research in the US shows that the key beneficiaries of merit-based scholarships are affluent students from high- and middle-income families.

Callender argues that these awards "represent the antithesis of government thinking behind bursaries".

Streeting agrees, describing them as "the most reprehensible bursaries and scholarships" on offer. "I can't stand them. They are designed to do nothing more than chase clever students around the system, and I think it is an outrage that taxpayers' money and student fee income is being spent in this way."

Callender worries that universities are "turning net price to their advantage in the competitive struggle for students" - and that this is not necessarily benefiting students.

"In some cases, bursaries and scholarships have perpetuated the inequities across HEIs that they were intended to alleviate; in some cases they may actually have exacerbated those inequities," she writes in her paper.

Hepi states that the differences mean that students at universities with high numbers of poor students are likely to be "significantly more reliant" on term-time work to help finance basic living costs than their peers at other, less socially inclusive, universities.

"Given what is known about the impact of term-time working on academic success, this inequity in levels of support may intensify the already unequal chances of successful outcomes for students with different levels of prior academic attainment," it warns.

When bursaries were introduced, the idea was that finance should not be a barrier to participation and students should be free to choose where and what to study, says Callender.

"The fact that ÖйúAƬ institutions are using their bursaries as a competitive mechanism for price-discounting turns that principle on its head."

Streeting believes that while there is a place for competition and markets, it is not in student support and student funding.

"Markets create winners and losers, and ultimately students from poor backgrounds who struggle to access the appropriate information and guidance will lose out."

Among those who argue that the current system is unfair, many favour a national bursary scheme under which a proportion of universities' additional fee income would be put into a central pot.

"The money could be shelled out according to need, and we would get rid of all this jiggery-pokery and inconsistency," says Ebdon.

But the Russell and 1994 groups are against the idea, with the Russell Group branding it "profoundly misguided". Both claim that forcing universities to pool income from fees would amount to a "tax" on students, and that institutions need flexibility in bursaries to allow them to target specific access problems.

Piatt says: "I could understand the idea of a national bursary scheme if we didn't have a really robust national student-support system but we do. As long as the fundamental foundation of student support is fair, I think it is only right that we allow for a bit of flexibility on top of that. We should not perversely penalise institutions for wanting to give a bit more in bursaries and scholarships than they have to."

She also says such a scheme could affect alumni donations targeted at student support.

But the Hepi report counters these arguments. It says that while there is little evidence that big bursaries are helping to widen access, under a national bursary scheme institutions would still be free to offer extra money on top of the national bursary.

"A national bursary scheme need not prevent institutions from offering additional financial aid in pursuit of their own strategic priorities or to more accurately reflect the likely cost faced by their own students," it says.

As such, it "would not eliminate inequity between institutions, since universities with fewer low-income students, and especially those with considerable resources other than fee income, would still be better placed to offer supplementary means-test bursaries. But the level of inequity would ... be reduced." There is no suggestion that donations from alumni would be affected, as they would remain targeted at a particular university and its students, the report adds.

According to Sir Martin Harris, director of Offa, which was set up to monitor institutions' efforts to safeguard access from low-income groups, much of the debate on a national bursary scheme misses the point.

The great majority of student support comes from the Treasury and is standard across England, Sir Martin emphasises. When top-up fees were introduced, legislation ensured that fees and bursaries were an area in which universities would be autonomous.

"That was separate from the publicly funded student-support arrangements, which are very properly politically determined and England-wide," he says.

He believes that maintaining this distinction is crucial: "If there is a minimum sum of money that poorer students should get come what may, then that is an England-wide issue and needs to be addressed by the England-wide student-support mechanisms," he says.

In other words, if students are not getting the right basic level of support, that is not an issue to be addressed through bursaries.

He believes that if a national bursary scheme had been introduced two years ago, the total sum that institutions would have put into bursaries would have been smaller.

"I would argue very strongly that because it was left to individual institutions the quantum of money put into bursaries is significantly higher than it would have been had ministers of the day determined that there should be a fixed percentage," he says.

"The minimum figure the then Secretary of State (for Education), Charles Clarke, had in mind at the time was 10 per cent of the additional fee income, with perhaps rather more for institutions that are less socially diverse - whereas what has actually happened is something like 25 per cent has been put into bursaries. It is my view that no government could have legislated at that time for anything approaching 25 per cent.

"My job as director of Offa is to worry about whether there is enough for the poorest - and I am satisfied, because I'm satisfied each year by their access agreements. If institutions choose to spend money over and above this, that is up to them."

But Hepi warns that if the status quo is maintained, the market in bursaries may result in an "arms race" for academically talented low-income students rather than a significant increase in the socio-economic diversity of the most selective universities.

"If these trends continue ... the highest-achieving students are likely to benefit disproportionately - in terms of both enhanced choice and financial support - from the financial aid offered by universities, but with little coterminous progress being made towards the Government's stated political goal of fair access for underrepresented groups," its report says.

Callender points to the US for comparison. It operates a different system, but institutional aid packages are becoming increasingly important as the richest universities compete to outdo each other on the size of their bursaries. And as the wealthy institutions race to offer more and more, the gap between the aid offered by the elite institutions and those that are less well off is widening.

But she argues that more research is needed in England to look at the extent to which different types of students are influenced by bursaries, and exactly which students are benefiting from the 40 per cent of bursaries that are not means-tested.

Callender has been commissioned by Offa to work on a new study with David Wilkinson from the National Institute for Economic and Social Research to look at awareness and take-up of bursaries, exploring the strategies and techniques used by universities in England to promote institutional bursaries to students, parents and academic advisers.

It will ask whether bursaries influence students' decisions about where and what to study, and the results will inform a good-practice guide for universities.

"I do think careful consideration should be given to a national bursary scheme - but we need more information on how it would work in practice and what its consequences might be," Callender says.

Law of unintended consequences

Claire Callender argues that the Government did not set out a clear rationale for the introduction of bursaries.

They were a concession to Labour backbenchers, she says, and as such "one of the unanticipated consequences of the 2004 ÖйúAƬ Act".

In an April 2003 government document, Widening Participation in ÖйúAƬ, bursaries are described as a way "to ensure that no additional financial barriers" to ÖйúAƬ were created by the introduction of top-up fees.

The following year, a letter from Charles Clarke, the Education Secretary at the time, to Sir Martin Harris, director of the Office for Fair Access, set out an expectation that universities would use some of their top-up fee funds to establish bursaries.

It said: "Institutions that decide to raise their fees above the current standard level should plan how they will safeguard and promote access ... there is an expectation that they will plough some of their extra income back into bursaries and other financial support for students, and outreach work. That is a general expectation for all institutions. However, I would expect that you would expect the most, in terms of outreach and financial support, from institutions whose records suggest that they have furthest to go in securing a diverse student body."

ÖйúAƬ