The Sutton Trust's latest report

The Sutton Trust's latest report, marked out by the claim that five schools in the UK sent as many people to Oxbridge over the past five years as the lowest 2,000 did, has been a fairly big news story this week. Broadly speaking, I have a lot of respect for the Sutton Trust, and there are definitely inequalities in terms of access to the top universities, but this study isn't the best one I've ever seen - in particular, the headline claim is overstated by a bit of statistical trickery.

The basic problem is that it deals in absolute numbers, without accounting for the size of the school. Four of the five schools that they cite have a larger-than-average sixth form size. Specifically, Eton has 257 pupils leaving the sixth form each year, Westminster has 188, St Paul's has 168, St Paul's Girl's School has 70, and Hills Road a full 873: by contrast, the average for the 2,343 schools they have data for is 136, and it's just 83 for the smallest 2,000. Just by dint of size alone, Hills Road will send many more pupils to top universities than a school like this will, which makes the statistic misleading.

This doesn't explain all (or even most) of the difference, though: but the underlying assumption here, that all schools ought to send an equal proportion of their pupils to Oxford, is questionable. All five of these schools are selective: rightly or wrongly, they're able to ensure that their pupils are better at tests and interviews than those in a comprehensive, by definition. It's not overwhelmingly surprising that a few years later, they go on to do better at tests and interviews and thus get into the top universities. Indeed, the fact that Hills Road - which isn't a fee-paying school - is in the top five is heartening, because that suggests it is a largely meritocratic process, not a result of money or private school connections.

Are there genuine inequalities here? Yes, but the problem isn't necessarily the existence of a cabal of 'elite schools' who have an unfair advantage. The problem is more a difference in the way success is measured, which means that 'non-elite' schools can't devote as much time to Oxbridge preparation. The standard metrics by which schools are judged (percentage of students getting 5 A*-C grades, average number of UCAS points) don't line up with what Russell Group universities want to see. Things like GNVQs, focusing on D/C boundary students and General Studies A-levels are encouraged to get the numbers up, whereas Oxbridge admissions need them to engage in the profoundly unprofitable practice of taking the A and A* students and stretching them further - which won't produce any increase in exam statistics. 'Elite schools', on the other hand, because they have a historically high Oxbridge admissions rate, can ‘advertise’ to potential students on the basis that they are so good at getting students into top universities – and so they both attract the type of students likely to go to Oxbridge, and they have a better incentive to really stretch the students. It’s not a genuinely unfair advantage – it’s just that they’re measured on a different, unofficial metric, which rewards them for devoting time to the top students..

On this logic, one of the solutions to the ‘problem’ would be to make this metric a lot broader – to judge all schools on how many of their pupils go to top universities, as has been suggested. The danger, though, is that a different metric will do more harm than good. Schools have scarce resources, and if we change the system to encourage them to funnel more resources towards the top students, that will mean that they aren’t being used to help the weaker students – and the numbers of illiterate and innumerate students are perhaps more worrying than any statistics on access to higher education. I’m by no means convinced that – if we have, say, a given amount of high-quality mathematics teaching to offer – that it’s best used to ensure that Mary gets into Cambridge rather than Anglia Ruskin, instead of overcoming Sally’s more basic difficulties in order to get her a C in GCSE Maths and a level of numeracy needed to be a functioning adult.

Essentially, there are good reasons for having schools (as a default position) work to get as many people as possible up to a baseline level of argument, writing, numeracy, scientific literacy and civic participation, which is what our current metrics measure. It’s dangerous to change that – to encourage schools instead to work on making the best students better – and it may well be worth accepting that some schools can have a niche in the market, and work hard to stretch the best pupils, in order to allow the majority of schools to use the teaching methods and resources which most benefit the majority.

More cherry-picking

After my mention of the risk of banks 'cherry-picking' the most profitable student loans, leaving the government to fund the rest, I thought it worth mentioning this openDemocracy post, which talks about something very similar happening with the Forensic Science Service.

Banks and student loans: why it's wrong

David Willets' latest suggestion for changing the university system, by encouraging banks to offer student loans instead of the government, is deeply worrying. There is, deep down, a reasonable justification for this idea: because student loans are only repaid at a relatively slow rate (9% of salary above £21,000) and written off after 30 years, the government will never recoup all the cost - somebody studying in London for three years after the fee rises come in, and taking out the full £7,675 London maintenance loan, would have to earn £39,500 on average every year for the first thirty years of his working life just to repay the capital, never mind the interest. Ultimately, this means that the government will be paying more to the universities than it receives back from these 'loans', it makes sense to try and reduce the cost of them by having banks pick up the slack.

Even so, this attempt to balance the government's books will be harmful to students. If banks begin to offer student loans, then - unsurprisingly - they'll be on terms much more similar to ordinary loans than they are now. Student loans, at present, are designed not to be a burden - they don't appear on a credit report, and are tied to income, so if you are unemployed, or taking a career break (to raise children, for example) there's no payments to make. In other words, they're a lot more similar to a tax than a credit card debt, precisely so that students don't have the stress that comes with ordinary debt. Turning them into a bank loan - so that students who struggle to get jobs have to pay their loans back while still unemployed, and so that it affects your mortgage applications later in life - would be a retrograde step.

There's also the problem that banks will be more discriminating about who they lend to. Specifically, they won't have the obligation that the government has to lend to everyone, and they'll be aiming to make a profit, which means they'll target universities like Imperial, the LSE, Oxbridge and Durham, and courses like Economics, Chemical Engineering and Law, which have the highest-paid graduates. This isn't just speculation - as the head of Universities UK, who's been in talks with the banks, says in the article, they're interested 'only if the university takes the risk, the government takes the risk, or they do it only for a select few institutions'. This maximises the bank's chance of being paid back - but it means that the government is no longer lending to the top graduates who will repay their student loans, only to those from less prestigious universities doing less lucrative degrees. This could end up as the worst sort of privatisation, where the banks make loans in cases where profits are available, and the government funds those who'll end up as a net loss - almost the same situation as with Royal Mail, where 'postal liberalisation' meant that private companies could take all the profitable business mail work, yet Royal Mail was still stuck with the costly requirement to deliver to every UK address, however remote, for a fixed price.

As the article notes, Santander seems to be heavily involved in these plans, which is even more worrying. Santander provide some of the worst customer service in the UK: they've increased profits by cutting staff, which has meant they can't deal with complaints and problems effectively, to the extent that their chief executive was summoned to explain it to Parliament's Treasury committee. We've already seen what happens when the student loans system breaks down; if we make it less accountable by having it run by banks, especially banks which already have abysmal customer service records, then it can only get worse.

But what alternatives are there? One solution is simply for the government to reverse its policy, and commit to spending more money on universities to cover this gap between what graduates officially owe and what they'll actually pay, but that's not something I advocate. If the government has extra money, then there are other areas it could increase spending on - social housing, primary and secondary education, disability benefits, pensions - which would make Britain better and fairer than increased university funding would.

The better solution is to increase what graduates end up paying. The 30-year payment limit seems arbitrary, and means that graduates who reach high-paid senior roles late in life stop contributing at that point; continuing repayments until retirement would make more sense. Similarly, the £21,000 limit, below which student loans aren't repaid, is new - for those, like me, who went to university before the £9,000 fees were introduced, the limit is £15,000. Although this was an important concession at the time, which made the increased fees less of a burden on students, it isn't serving its purpose: if it means that the government can't afford to fund the loans, and that students have to be funded by banks of questionable quality, on terms which are unfavourable to those who are lower-paid or have time out of work, then they're better off without it.

With these two reforms, rather than having to earn an average of £39,500 for thirty years for the government to get its money back, our average London student would only need to earn £27,000 annually, averaged over his whole career, which now includes any promotions he gets in his fifties. That's a lot more realistic; the mean UK salary, including non-graduates, was £26,510 in 2010, whereas a salary of £39,500 would put someone well within the top 20% of earners (source: Annual Survey of Hours and Earnings 2010, Table 1.7a).


Calculations: Information on loan amounts is taken from direct.gov.uk: £9,000 tuition fee plus £7,675 London maintenance loan for three years totals £50,025, which would need an average £1,667.50 payment per year for thirty years; £1,667.50 is 9% of £18,527.78, for a total salary of £39,527.78 once the £21,000 is added. If that £50,025 was repaid over 47 years (from graduation at 21 to retirement at 68), it would only be a £1064.36 annual payment; that's 9% of £11,826.22, for a total salary of £26,826.22 once the £15,000 is added. All these calculations exclude interest (of up to 3%) and the figures are in real terms (ignoring inflation changes).

MEPs and the death penalty

One of the unfortunate side effects of European Parliament elections is that, since you can only vote for parties and not for individuals, people with rather extreme views can get elected. In particular, one of the East Midlands' Conservative MEPs seems to be a rabid leave-the-EU bring-back-hanging the-Guardian-is-ruining-Britain type. In particular, in the wake of the Levi Bellfield case, he's calling for a return to hanging, attacking "bien pensant opinion — à la BBC and Guardian", and lamenting that the EU is banning us from hanging people*.


The comments on his post call him out on his flawed logic and his appeal to mob rule, so I won't deal with those here. I haven't, though, noticed many responses to his claim that

The arguments against simply don’t stand up.  What about the risk of a miscarriage of justice, and an innocent man being executed?  Of course capital punishment should only be applied where the guilt is clear, as in this case.  Of course the judge should be able to commute to life imprisonment if he has any mental reservation about the jury’s verdict.  That’s what judges get paid for — to use their judgement.

I'm not quite sure why he insists on the criterion of 'only where the guilt is clear'. I would hope - and I don't think this is idealistic of me - that nobody is convicted, regardless of sentence, where their guilt isn't clear. That's what it means to be 'beyond reasonable doubt'. I'm not really sure that I'm cheered by the reassurance that 'we'll only execute you if your guilt is clear', if the implication is 'we'll just lock you up for life if your guilt isn't clear'. Of course, the clearest way to show guilt is when the criminal confesses - but if criminals know that a guilty plea means death, whereas the doubt of a not guilty plea only means life imprisonment, that's a massive incentive for them to argue their innocence and encourages a costly trial.

Further, I'm not convinced that the judge's 'mental reservation' is an appropriate safeguard. The very reason we have juries is, in part, because we recognise that judges don't represent the general population: being a judge requires being a lawyer first (except in magistrates' courts, which don't try murder cases), which means that they are, almost without exception, university graduates who either got hired by a top law firm or were rich enough to fund their own training. They'll be as biased as juries, if not more, which is not what you want  when sentencing people to death. This is a perennial problem with American justice, as the ACLU notes:

The DOJ [Department of Justice] study also revealed the influence that the race of the victim has in determining potential capital cases. U.S. Attorneys recommended the death penalty in 36 % of the cases with black defendants and non-black victims, but only recommended the death penalty in 20 % of the cases with black defendants and black victims.

But more generally, I actually suspect that the Levi Bellfield case shows us why the death penalty would actually be a bad thing. The trial seems largely to have been about a sense of closure; Bellfield received the UK's harshest sentence, a whole life tariff, in 2008 for bludgeoning two students to death, so his conviction this week does not affect his status at all. If he had received the death penalty in 2008 (and, as a multiple murderer with nine previous convictions, he seems like a prime candidate),. then today, the Milly Dowler murder would still be officially unsolved, and the closure which the trial brought would be denied to the family. As they themselves put it three years ago:

Imagine not knowing how your daughter died, or where or when and by whose hand, and imagine how we as a family live.

One last worrying thing about the Bellfield trial - not related to Helmer this time, but to the general media response - is the criticism of the cross-examination of her parents and the call for more victim's rights. Why is this worrying? Because British law is not based on defence lawyers admitting that their client is a bad egg and giving in quietly. It's based on their attempting to show grounds - any grounds - for reasonable doubt, and if it's even plausible that a child could have run away from home rather than being abducted, then the defence need to be able to pursue that line of questioning. This isn't about allowing murderers to torment their victim's family in court (and, indeed, judges already have the power to step in when barristers go too far); it's about letting someone on trial, who may or may not be a murderer, have every opportunity to defend himself. Distressed parents are never a welcome sight, but in the long run, it's better to accept that rather than to say that certain lines of questioning are taboo, and thus risk wrongful convictions.

* Although the death penalty is forbidden by Protocol 13 of the ECHR and the Human Rights Act, neither of which are directly related to the EU.