This article is more than 9 years old

Who’s afraid of student loans?

A new study by the Institute of Fiscal Studies (IFS) on student loan costs published today will kick off a new round of suggestions on what to do next for higher education funding. Emran Mian maintains that we should be much more sanguine about student loan costs and writes that we need to tackle some even bigger problems. This piece attempts to explain why.
This article is more than 9 years old

Emran was formerly Director of the Social Market Foundation, and a member of Wonkhe's Editorial Group.

A new study by the Institute of Fiscal Studies (IFS) on student loan costs published today will kick off a new round of suggestions on what to do next for higher education funding. I maintain that we should be much more sanguine about student loan costs and this blog is an attempt to explain why.

To start with the IFS report, it forecasts that the cost to the Government of making student loans will be 43.3% of the amount that is lent. It is a remarkable figure, though it is not the 43 that caught my attention, it is the .3.

When forecasting over a 30 year repayment period, over a cohort of hundreds of thousands of borrowers, with about a dozen variables, it would be remarkable to be so precisely correct even if you were looking backwards in time never mind casting your model into the future when the economy, job types and career choices will change in ways that we don’t know about yet.

But let’s take the pretence of certainty out of the number and say that the cost of making student loans looks like it will be somewhere around 40%. Even this is based – as the IFS readily admits – on assumptions that are contestable.

For example, the report assumes that no one will repay their loans early, yet current information is that there are several hundred million pounds of early repayment. The IFS also assume that graduate earnings will rise by only a small margin above inflation. They may be right on the whole, but we used to assume in the Government’s version of the student loan calculation that, as well as what you might call cost of living increases in salaries, many graduates would from time to time benefit from a step change in their earnings, following promotion or winning a new job in a different company. I understand that the Government’s modelling has recently adopted much more pessimistic assumptions about the number of such step changes in earnings. Given the IFS forecast of student loan costs is so similar to the Government’s latest one, I wonder if they are making a similar judgement. They may be right, or the periods of economic growth we’re likely to see over the next 30 years might mean that lots more graduates than they predict do enjoy such step changes in their earnings. If they earn more, they pay back more and the Government’s loan costs decline.

But all I’m doing here is playing counter-assumptions against assumptions. Now I’m going to change the game, it’s time to play two important facts against the assumptions made by the IFS (and indeed the Government).

The first is on interest rates. The IFS and Government models assume that the Government’s cost of borrowing is 2.2% in real terms, i.e. it costs the Government as much as that to raise the money that it lends to students. Whenever the interest rate charged on student loans is less than the rate paid by the Government, then the Government is making a loss and this adds to the cost of the loans.

The problem though is that the Government’s cost of borrowing is nowhere near as high as the model assumes. The IFS report includes the calculation that the cost to Government of index-linked debt since the turn of the century has been more like 1.1%, i.e. it isn’t losing as much money on making loans as we might think, it’s just that the model assumes it does. Change the assumption in line with the facts and the cost of making student loans drops by a massive 13 percentage points to somewhere like 30%.

Now you might say that this is a temporary effect and that the costs of Government borrowing might rise in the future. Luckily, Government is able to sell index-linked debt with a term (50 years) that is even longer than that of student loans and so it can fix its borrowing costs at a low level. By the way, it’s already doing that, this isn’t hypothetical.

But this seems too good to be true. Can student loan costs really be more like 30%?

I’ve got another one. The costs forecast by the IFS assume that the repayment threshold for student loans is £21,000 and that it will rise every year in line with earnings. In a way they have to use that assumption, and so does the Government, because it is the declared policy. It is a good policy too, as it maintains the value of the repayment threshold, protecting lower-earning graduates long into the future.

But the problem is this: thresholds of this sort – e.g. those for income tax – have a habit of not rising in line with earnings. Until 2008, when first the Labour Government and then the Coalition began increasing the threshold for lower rate income tax, its value had fallen by something like 20% relative to wages. The threshold for the higher rate of income tax has fallen relative to wages by even more.

Depending on your style of political expression, this phenomenon is either known as ‘fiscal drag’ or ‘stealth tax’. In other words, it is quite easy not to uprate thresholds like these over time and to win the gains for the Exchequer. The same thing happened to the repayment threshold under the pre-2012 system of student loans. It is almost certain that it will happen again. And again the effect of changing an assumption in the model to reflect a fact – this time a political one – is non-trivial. The IFS calculates that, if the threshold was to rise by 2% per year rather than in line with earnings, that alone would knock something like 12 percentage points off student loan costs.

So there it is, this is why I can’t get excited about what some people describe as escalating student loan costs. The costs are de-escalated as soon as we change the assumption in the model about Government borrowing costs to reflect what Government is actually paying to borrow. And they drop even further if we assume that the repayment threshold will be subject to the same erosion over time as every other threshold in the tax system. Add some optimism about future graduate earnings and the cost of the student loan book may be more like 20% or even fall into the teens as compared to the 43.3% forecast by the IFS.

But while I’m sanguine about the cost of student loans there are other issues that we do need to think more about in funding higher education. The numbers of people studying part-time – i.e. those taking a different route into higher education than the traditional one assumed by policy makers or indeed model makers – is falling. There are big questions about whether we are likely to have enough people going on to postgraduate study in the future and whether the differences in participation rates between rich and poor that we are beginning to redress in undergraduate education will reappear later on. And equally isn’t it a problem in the end that politicians are already thinking about unwinding a funding system which was introduced barely two years ago and no one who entered on the new terms has yet graduated?

This is why I’m participating in the new panel on student funding launched by Universities UK today. Back when the Browne Review was launched in 2009, we had cross-party support for the task and the definition of the challenge we had to solve. It would be great to get back to that place. As I see it, it’s time to stop fussing about student loan costs and think about something more difficult instead.

Leave a Reply