It’s difficult not to notice the outrage of graduates stuck with Plan 2 student loans at the moment.
The main gripe is that their loans are getting bigger year on year despite making repayments, thanks to punitive interest rates, and they’re realising they’re stuck paying an extra 9 per cent “tax” for 30 years.
This shouldn’t come as a surprise, though, as the Student Loans Company’s (SLC) Official Statistics predicted in 2022 that only 20 per cent were likely to repay in full.
It’s likely worse than this, as the government now expects to get back significantly less of the money it has lent to students than it thought previously.
And this is consistent with my own research, where I used some simple salary profile data to show that only straight-A grade top-earning graduates – around the top 10–15 per cent – are likely to earn enough to pay back in full.
What the IFS says
Against this backdrop, the Institute for Fiscal Studies (IFS) has recently published data that contradicts the Official Statistics and appears to go against the reality of what Plan 2 graduates are seeing in their annual statements each year.
“Overall, we expect just under half of the latest entry cohort repaying under this system not to clear their loans by the end of the repayment period.
The IFS may not have access to the same data as the Department for Education (DfE), of course all statistical forecasts of this nature are sensitive to small differences in assumptions and behavioural treatments, and its data does take into account the freeze in the repayment threshold – but that doesn’t explain its fundamentally different prediction that over 50 per cent will repay in full.
It can only stack up if the IFS has far higher predictions of graduate earnings than the DfE. The IFS is clearly entitled to use whatever data and modelling it sees fit, as this was independently commissioned research, but apparent contradictions of this nature can fuel distrust.
Enter the graduate premium
The debate about the perceived unfairness of student loans has inevitably spread over into whether the degrees that the loans pay for are worth the money.
There are many unofficial views and reports that attempt to work out the supposed graduate premium – often used as the justification of the cost of the degree and the debt – but in fact there is only one report that falls within the definition of Official Statistics, and that was the now defunct Graduate labour market statistics, which the DfE produced.
This is due to be replaced this year, as a result of my work, with a new method of comparing graduate pay to non-graduate pay based on the Longitudinal Educational Outcomes (LEO) database.
The IFS does, though, have an “official” role with regards to comparing graduate pay and non-graduate pay, as its report The impact of undergraduate degrees on lifetime earnings from February 2020 – though not part of Official Statistics – was commissioned by the DfE. And when this report was published, the government announced it on GOV.UK under the excited headline:
“Graduates enjoy £100k earnings bonus over lifetime. New research shows majority of graduates are better off for going to university and empowers students to make informed career choices.
£100k better off?
Since its publication, this report – largely due to the public’s high trust rating of the IFS – has been arguably more influential than the Official Statistics in feeding the public perception that going to university remains a good idea financially, regardless of the already high percentage of school leavers going into HE.
This is in no small part due to the IFS’s own press release, which starts with the very positive statement:
Going to university is a very good investment for most students. Over their working lives, men will be £130,000 better off on average by going to university after taxes, student loan repayments and foregone earnings are taken into account. For women, this figure is £100,000.
And even though the report goes on to say that 20 per cent were getting a negative financial outcome, the impression given overall is that this is largely down to poor course choices – so if you just avoid these, then everything is fine.
Given that the DfE has decided to allow an outside contractor to produce this lifetime earnings report, it’s imperative that it maintains appropriate checks and balances with regards to accuracy and transparency of data, and ensures a certain amount of methodological alignment with Official Statistics.
The DfE also needs to be mindful that all think tanks, research, and policy groups will inevitably have an agenda and inherent bias that will be reflected in the work they produce, and it’s dangerous to treat any organisation in a reverential manner as if normal healthy scepticism doesn’t apply.
With regards to education and in particular the percentage going into HE, this is an area that is prone to being ideology-driven, while at the same time there is undoubted commercial interest from the HE sector in ensuring that participation remains high.
Should this report exist?
With regards to the IFS lifetime earnings report of 2020, I carried out my own appraisal as far as I’m able, and identified a number of fundamental concerns. The first category questions whether the report should be published at all.
It’s clear, even from the name of the report, that it’s based on an assumption of causation – that the “impact” of a degree is the main and essential cause of employment pay. Yet no proof is given that the degree is impactful or to what extent, nor that it is essential.
Assuming a high level of causation isn’t statistically sound given that the majority of graduates end up in jobs that have no direct connection to their degree subject, and even those with a connection haven’t been shown to have needed the degree rather than learning on the job as a trainee, had the employer given them the opportunity.
While it’s true that all graduate outcome statistics suffer from this problem of assumed yet unproven causation, for this report to assume unproven causation over a whole lifetime’s earnings is problematic in the extreme.
The report is also inevitably too highly speculative. Attempting to work out the earnings of graduates versus non-graduates over a lifetime is far from an exact science, and it’s reckless to issue the report when it’s bound to be interpreted as “exact figures” to be wholly relied upon rather than speculative projections. Even though the level of speculation is clearly stated in the report, this doesn’t matter if the public are bound to ignore this and treat it as a definitive statement of fact.
It’s also troublesome that the level of complication about the parameters and modelling assumptions means it may simply not be possible to assess differing assumptions or even unconscious bias within the model – which is why it’s questionable to give this to an outside policy research group, as illustrated by the differing graduate salary assumptions leading to the IFS predicting totally different levels of loan repayments as explained above.
Already out of date
The 2020 report was 15 years out of date when it was published, as it was based on graduates from the mid-2000s, yet the report is bound to be interpreted as if it’s relevant for today’s prospective students.
This is a particular problem given the change in conditions for graduates in the job market that occur over the years and the different make-up of the graduate population, with an increased number from lower prior academic attainment given the rise in HE participation over this period.
Turning to the methodology of the report, the cost of going to university is grossly understated – the report claims that the cost for the average graduate is £nil for men and even around £10k positive for women, even though the total cost of going to university for a three-year live-away-from-home degree is actually around £90k now according to some research.
Some of this £90k will be made up of parental contributions, as the full maintenance loan isn’t sufficient to cover living costs – and there’s an argument that parental contributions should still be included in the figures as a real cost, albeit to the graduate’s family rather than the graduate themselves.
An element of this £90k will be subsidised by the government in the form of student debt write-offs, but under Plan 5 loans this is now only relatively modest at around £15k per graduate, so £75k is missing from the costs.
Mean or median?
It’s accepted practice that the median is the best measure in graduate statistics to avoid the mean average being distorted by higher earners. If the median were used, then the overall average lifetime extra earning would be far lower at £70k instead of £130k or £100k.
There is no good justifiable reason not to use the much lower median, particularly given that this report is going to be highly relevant to ordinary members of the public.
If the IFS were to correct both of the above – the understated costs and the use of the mean – then the revised overall average would in fact be close to £nil instead of £100k or £130k, which would of course fundamentally affect how the results were perceived.
Publishing the overall average when used in graduate statistics is now discredited, as the Office for Statistics Regulation (OSR) has ruled with regard to the misleading overall average of the graduate premium in the now discontinued Graduate labour market statistics.
Instead, any report attempting to work out the difference between graduate and non-graduate pay should only publish in bands of prior academic attainment – which is the intention with the new LEO-based Graduate and postgraduate labour market outcomes report due out next year to replace the Graduate Labour Market Statistics (GLMS).
The IFS report should use the same bandings as this report and should also not publish the misleading overall average at all anywhere in the report.
Doubling down
These concerns have been raised with the IFS and the DfE over the last 18 months of my work in this area. However, instead of considering withdrawing or amending the report – or perhaps even because of my concerns – the DfE has decided to double down and has issued a £75k contract to the IFS to update this report and re-issue it in summer 2026.
This is at the same time as the new LEO-based Graduate and postgraduate labour market outcomes that will replace the Graduate labour market statistics, which contained the defunct graduate premium data.
So it’s destined to be published in direct confrontation with the Official Statistics, and is likely to get more attention and will diminish the impact of what should be highly interesting and impactful official data.
This isn’t an ideal scenario, as the new LEO data deserves space to be considered by the public in its own right without being drowned out by the impact of the IFS report’s findings.
All these concerns with the IFS report have been raised with the OSR, but to date it is refusing to look into the issues on the technicality – which I dispute – that though the report is commissioned by the DfE, because it doesn’t actually produce it, the report doesn’t fall within Official Statistics and as such isn’t in its remit.
And this is the crux of the matter – by using this loophole to subcontract this report, the DfE avoids the check and balance of the OSR, which is designed as a safeguard mechanism to scrutinise and oversee government statistics.
Too much at stake
There are hundreds of billions of pounds of government taxpayer money at stake that hang on the perception created by graduate statistics, and half the country’s school leavers are having their personal finances blighted by student loans – so it’s difficult to think of more consequential statistics than graduate statistics at the moment.
If this lifetime earnings report is to be produced at all – or any other subcontracted report for that matter – it’s vital it should be produced under tighter controls that involve legislative OSR scrutiny to eliminate the possibility of error, bias, and inconsistent modelling.
It’s also important that the OSR investigates concerns raised like mine in relation to subcontracted reports – rather than disregard them.
Thanks Paul – as ever you set out the depressing reality that mass HE is a poor ‘buy’ for the taxpayer in terms of future student/graduate debt to be written off (and compared to the same taxpayer spending on, say, FE or on Primary schools) and a very poor ‘buy’ for far too many of the students/families incurring fees & living-costs loans (but at least the recent significant rise in commuter-students living at home means some are wisely avoiding high rents). Keep up the good work!
There are various issues with this analysis. I won’t go into them all, but to pick on one – the point on median vs mean. Means can be an appropriate way of looking at lifetime earnings comparisons because:
1. policy analysis often cares about total economic return.
2. large positive returns for some individuals are economically meaningful.
To say there is no justifiable reason is not correct.
This comment is also very telling:
“The DfE also needs to be mindful that all think tanks, research, and policy groups will inevitably have an agenda and inherent bias that will be reflected in the work they produce”.
Indeed. And Paul’s wider agenda is that far fewer people should be going to university, however, his narrow viewpoint entirely misses the bigger picture, picked up by John Burn-Murdoch’s analysis here:
https://www.ft.com/content/649d3c64-b8e5-4979-9f0c-9aebd43642e2
To summarise, many other nations (the US,Spain, Canada, France, to name a few) have seen similar expansions in Higher Education participation over the last 30 years or so, yet only the UK has seen a fall in the graduate earnings premium. Perhaps we need to be looking at wider issues to do with the under-performance of the labour market?
Finally, what I’ve not seen in any of these articles, is what we would do with the students who would no longer have HE as an option. They would be disproportionately from less well-off backgrounds (because a key driver of educational outcomes is still, unfortunately, income/wealth) – where would they go?
None of this is to say that there aren’t issues with the system. Some of the recent stories on graduate loan debt turn my stomach, and I was interested to see that even the USA has systems in place to mitigate against some of the outcomes we are seeing here. But the narrow minded response of cutting access to HE is not, in my opinion, the solution.
I simply don’t believe that a further 3 years of academic education (more often than not in an unrelated subject) is necessary to be useful in the vast majority of jobs that society needs doing, and that this is a myth that the HE sector has created in order to drive up customer demand. I just don’t buy the concept that something ‘magical’ happens in HE that turns people into fully operational members of the workplace, and I think we have completely forgotten that learning on the job with perhaps targeted training can be a far more effective way of developing a career. So I think that we need to stop treating 18 year old non graduates as if they aren’t yet fit for the workplace and encourage employers to employ then direct from school as opposed to the discriminatory practice of employing only graduates for even the most basic trainee positions. Thus forcing ‘everybody’ to go through an HE paywall to enter the workplace. I can’t comment on other countries data, but the LEO data clearly shows that graduates with lower prior academic attainment (below about 3 B’s at A-level) are not achieving pay much above minimum / average wage. So why on earth are we content to allow them to be buying into an HE sector that is only giving them a one-way ticket to a lifetime of being in the student loan debt trap and a huge write off for the tax payer. This isn’t an opportunity for them , it is exploitation. And yes, this is absolutely my opinion / ideology / bias ; there is no problem with having one for commentators like me & think tanks ; its just that the DfE needs to be careful that they themselves don’t let ideology affect their output, and that utilising a think tank like the IFS to produce ‘official’ data is problematic.
Thanks Paul. A lot of fair points here, but also some magical thinking. The UK is terrible at investing in workplace training – not looked extensively, but seems like Germany and the Netherlands spend around 1/3rd more than us, for example.
I also think your comment on the ‘myth’ that HE has created around degrees being needed for employment needs some evidence – employers are asking for them, yes, but I don’t think that’s because HEIs have made them do it. In fact, I think it’s to do with the first point – employers in the UK don’t want to invest in skills. By asking for a degree they can then make the argument that universities don’t prepare young people for work and feel like they have off-loaded the problem. Also, if this myth is real, then it’s global – and pretty much every other country is making a better go of it than we are.
The system has some big issues, but the overall fix needs to be broader than HE. Comes back to the question, if not university, then what are we going to ask our young people to do? 50% in the trades? Something else?
Without a decent plan, this feels like Brexit to me. Let’s add a massive shock to a system and not worry about the consequences. Well, the consequences will be institutions failing (some argue this is fine, but they tend to ignore the fact that universities employ a lot of people and tend to pump a lot of money into the economy, locally and nationally) and a lot of young people, mostly from less well-off backgrounds, with nothing to do.
And the link between the income/wealth of parents and kids educational outcomes is real. It’s not a reflection on their academic ability or intelligence, unless you believe the two are correlated, but I can assure you that they are not.
Thanks for your comments. I won’t attempt to answer all of your points and will just answer one. You ask ‘If not University, then what are our young people to do’. Well if the majority of graduates are ending up in jobs that are nothing to do with the degree, then the answer is quite simple – just do exactly the same job, but start at 18 instead of 21. We just need employers to be willing to employ them as trainees and not only want to employ graduates – even though the course subject was nothing or very little to do with the job they are recruiting for. It is age discrimination I would argue which is actually illegal. So we need positive political messaging about 18 year old hard working tax payers who go to work instead of University as opposed to Bridget Phillipson saying that you are a write off if you don’t enter HE. And we need to incentivise employers to employ 18 year olds through wage and training cost subsidy. And this is for both academic and non-academic school leavers. Because all student loans are a bad thing for a graduate’s finances , no matter what the terms, so we need pathways to work that don’t involve getting into debt and having to pay into the HE paywall where we can, so as to not make it compulsory to get into debt just to be considered for a job.
With regards to the specific point about using the Mean (£130k/£100k) instead of the Median (£70k) , then I concede that there will be different opinions. The point I was making is that the LEO data and the HESA outcomes survey data uses the Median , so why use the Mean for lifetime earnings ? Particularly when it is clear that it will be distorted by those few with significantly higher earnings? And that the judgment of the IFS in using the far higher mean as the headline figure should be able to be scrutinised by the OSR so that they can decide and settle the issue. But by the DfE using a subcontractor , they are putting the report outside the remit of the OSR , unless the OSR evokes its special discretionary powers to investigate reports that fall outside the strict definition of Official Statistics.