Skip to content

We’re not in 2014 anymore …

May 12, 2017

What’s the current estimated ‘loss’ to government on new student loans issued (to English-domiciled students and EU students studying at English HEIs)?

23 per cent according to the 2015/16 BIS departmental accounts (published last summer). £11.460billion new loans were issued and associated future repayments are expected to amount to £8.826bn over the next three decades or so. It’s those associated repayments, which you would lose by converting loans made to cover tuition fees into grants.

The Treasury has currently set a ‘target impairment’ (or target RAB) of 28% (see target impairent correction) and, as maintenance grants disappear and are replaced with higher maintenance loans, that non-repayment rate is likely to close in on the target. London Economics have an independent estimate of 28.6% for new loans issued in 2016/17.

If the target is missed a particular set of budgeting conventions kick in for the Department for Education, which has taken over responsibility for student loans from BIS. These will require the department to make savings from elsewhere in order to cover any surplus resulting from overshooting the target.

We are a long way from 2014 when the expected ‘loss’ on new loans issued had hit 45% after the government had budgeted for 30%. How has this recovery happened?

Firstly, the repayment threshold for those who have started since 2012 has been frozen at £21,000pa until 2021 – instead of rising in line with average earnings from 2017. A lower repayment threshold means more people in repayment and higher repayments – so more income is generated for the government. This was a retrospective price hike for borrowers.

Secondly, last year the government changed the discount rate on student loan repayments leading it to value future payments more highly. The government now discounts by RPI plus 0.7% rather than RPI plus 2.2%. With RPI at a constant 3%, that would make a future payment of £1000 in five years’ time worth £833 today rather than £774.  In contrast to the threshold freeze, the projected cash stream didn’t change: instead how that stream is valued improved.

Note that the value of £8.826bn repayments will be revised down in the next accounts due to recent upwards movement in RPI. As the discount rate moves up with RPI, the future value of repayments comes down – putting extra pressure on the DfE budget today (earlier this calendar year, DfE made a supplementary resource claim for an additional £11bn to cover movements in ‘the macroeconomic determinants of the student loan book’).



  1. Brian permalink

    “earlier this calendar year, DfE made a supplementary resource claim for an additional £11bn to cover movements in ‘the macroeconomic determinants of the student loan book’”

    Where did you find that out?

    • Supplementary Estimates are published by the Treasury every year in Feb. They show what departments have requested in addition to their main budget.
      You can find them here

      In this case, the £11bn may not all be used but it will be used to cover a ‘stock charge’ – a deterioration in the value of existing loans. This will be due in part to recent increases in RPI changing the discount rate and, for pre-2012 loans, the decision by the Bank of England to lower base rates to 0.25% in August. The latter lowered the interest rate on pre-2012 loans to 1.25%.

  2. Brian permalink

    Presumably higher RPI also reduces collections from pre-2012 loans due to it raising the repayment threshold at a faster rate?

    Can DfE just claim extra from HMT to cover extra resource charges on the student loan book? I thought they now had to find savings from elsewhere in their budgets and HMT were no longer sanctioning further budget increases to cover them blowing their budget on student loan policy?

  3. Yes – that’s also the case.

    DfE can request this additional resource – you could argue that RPI and bank base rates aren’t in the DfE’s control. But the new accounting treatment mentioned in the post only applies to loans ‘not earmarked for sale’. (ie pre-2012 loans don’t get treated in this way).

    Full text from the latest Consolidated Budgeting Guidance:
    (DEL = departmental expenditure limit)
    (R in RDEL & RAME – stands for ‘resource’, ie budget to spend)
    (AME – annually managed expenditure – for volatile items)

    Higher Education student loans not subject to a sale
    8.22 Any revaluations of the impairment that occur periodically because the original values were based on forecasts that have turned out to be incorrect, or because of updates made to the student loans model, and which go beyond the target impairment set by the Treasury, will be charged to DEL over a 30-year period (unless departments decide to cover the costs from their DEL over a shorter timeframe). One thirtieth of the total cost will be charged to non-ring-fenced RDEL each year for 30 years, with the residual amount each year RAME. The net effect of these entries in RDEL and RAME each year will equal the annual impairment charge due to these forecast changes. Revaluations of the impairment that occur for any other reason will be charged to RDEL in full and in-year, in the same way that the original impairment is charged to RDEL.

    • Dan permalink

      If we’re not in 2014 anymore then we’re certainly not in 2010 anymore when the Browne review used a discount rate of RPI + 2.2% to recommend a real interest rate be applied to learns which the Government of tge time then used to justify setting a rate taper of RPI to RPI + 3% (and postgraduate loans use RPI + 3% flat). So if the discount rate is now RPI + 0.7% surely these no longer justify such a high real rate of interest? Surely there should be some mechanism to review the various variables applied to loans from time to time as the economy changes? How can the Government continue to justify applying a rate that now far exceeds the Government cost of borrowing when the terms are flexible to allow them to review the parameters from time to time? They’re happy to flex terms to suit them (withdrawal of repayment holidays, freezing thresholds for longer than originally announced, setting interest at nil rather than negative when RPI falls below 0) but the only time I’ve seen them flex terms to the benefit of students is when they raised the threshold from £10,000 to £15,000 in 2005 and then by inflation from 2012 but even this was a worse deal than students were originally promised (in 2000 the Government said they intended to raise the threshold by average earnings, in 2005 they said they were instead raising the threshold to £15k and then again annually by inflation from 2010 but it still stood at £15k until 2012). They introduced a new threshold for new students from 2016 of £21k which flexed terms for new borrowers but surely the interest rate is overdue a review if the discount rate has been revised substantially lower as it changes the figure Browne based the recommendation on which led to the current formula.

      • This is an important point and one which would need longer treatment.

        A couple of quick points:

        The government did not accept the Browne recommendations on interest. Browne recommended RPI plus 2.2% for ‘high earners’ and RPI-only for those earning below £21 000, but the tapering would have worked very differently. Browne recommended an ‘interest rebate’ for those who made repayments that were less than the real interest accruing. That is, for Browne there would be no compounding of interest above RPI: for lower earners debts would not grow in real terms.

        From what I understand, the government looked to New Zealand for the RPI + 3 idea. There was also a fair amount of discussion at the time that the long-term average government cost of borrowing was in the region of 3%. Barr & Johnson’s 2010 paper on ‘Interest Rate Subsidies’ wrote:

        “The objective is to design a loan scheme in which subsidies are well-targeted. This suggests an interest rate that fully covers the government’s real cost of borrowing; as already discussed, we regard 3 per cent as the best estimate of that variable. A 2 per cent rate benefits better-off graduates with no change in the position of the least-well off, and is thus less well targeted than a 3 per cent rate coupled with 25-year forgiveness.
        “If 3 per cent were regarded as fractionally higher than the long-term real rate on government bonds, the model analysed here would introduce a small cohort risk premium, bringing an element of social insurance into the scheme. Saying the same thing a different way the scheme would incorporate a small amount of redistribution from richer to poorer graduates in a cohort.
        “Note that we are not proposing that graduates should be charged a 3 per cent real rate, but that they should pay an interest rate equal to the actual rate at which government borrows. The 3 per cent figure is intended to illustrate how the policy would work, rather than a prediction of what will actually happen.”

        Subsequently there were proposals to reduce the discount rate to RPI plus 1.1% by Leunig & Shephard – but, if I recall correctly, these proposals made no mention of altering the interest rate. Note that many think that the high interest rate incentivises those who can to make additional early repayments or to avoid borrowing in the first place (these analyses would underscore the ‘deadweight’ cost of subsidizing those who would have paid upfront and the reason RPI + 3 is applied to undergraduates while studying is to avoid the opportunistic wealthy taking out a loan to invest or spend elsewhere).

        I agree though that the broader issue now needs addressing. A complicating factor was revealed during my research into government accounting and budgeting – the discount rate isn’t really conceptaulised within the Treasury as the ‘cost of government borrowing’ and indeed the Treasury were looking into recasting discount rates in relation to GDP growth until recently.

        A further factor – the real rate of interest was probably explicitly set up at plus 3% in order to facilitate a future sale of student loans. Certainly back in 2011/12 this was the discount rate that the private sector told the government that they would use to price the new student loans.

        And I agree with your general complaint – which has been a theme on this site and my other work since it began – parliament granted the government administrative power to change terms on loan repayments. The 2011 Education Act contained the relevant clauses relating to interest rates on student loans. The government intends to use these powers to control the level of repayments and will ‘manage’ loans accordingly. This is the point of ‘flex’.

        There was a long discussion about the repayment threshold for pre-2012 loans and RPI under a post back in January. I think we concluded that since the relevant RPI in 2011 was negative, borrowers were spared a reduction in the threshold then. My reading though is that the new Coalition government did not reintroduce the annual RPI uprating because they wanted to keep a commitment an earlier government had made; instead some form of indexing was required in order to prepare the pre-2012 loans for sale (a change of plan – the intention stated in the 2011 white paper was to sell the new post-2012 loans and not to undertake a ‘retrospective’ sale – where the interest rate cap presents difficulties in achieving value for money).

  4. ps owing to illness last year, I was slow to realise that the government has a flat interest rate of RPi plus 3% for postgraduate loans. I see these loans as problematic and would always have preferred any extra cash to go into increasing the loan amounts available for undergraduate maintenance.

    pps owing to a legislative ‘pecularity’, interest is not calculated as you would expect. It is RPI plus 3 percentage points rather than a 3% increase on top of an RPI increase. That means e.g 1.031 + 0.03 rather than 1.031*1.03. Or 6.1% rather than 6.2%!

    • Brian permalink

      Very interesting post re: interest and flexing loan terms and one which I totally agree with. For borrowers like me with both pre and post-2012 loans the interest rate has a different effect – it keeps me in repayment longer if I’m a lower earner due to my write off date for being student debt free being far longer than 30 years away ‘overall’ and the decision to collect my joint loans above a lower repayment threshold (the uprating of which was delayed) serves as a double whammy – higher repayments for longer. So the interest on post-2012 for me is particularly relevant as its punitively high rate could result in me losing both of the main subsidies/benefits of pre-2012 loans – the lower rate (as it’s written off with the rest of the loan after 25 years and the write off as the higher interest on the post-2012 loan keeps me in repayment beyond the date my pre-2012 is gone).

      The post Andrew refers to regarding uprating of the pre-2012 repayment threshold is this one:

      To provide some more evidence of the RPI uprating of the pre-2012 threshold merely been a belated implementation of the outgoing Labour Government’s plans, Nick Hillman sent me an email after I extensively commented on his piece

      He was Special Adviser to David Willetts when the decisions were made so if anyone knows what happened he does. I copy his email below:

      Nick Hillman
      Wed 14/09/2016 08:57

      Thanks Brian,

      One of the things I learnt from your piece was that £15k was meant to rise form 2010. From memory, I think what happened was that our officials insisted it should increase from 2012 in line with inflation because that was what the previous Government had proposed (which your evidence suggests is only half-true at best). I suspect there was a political and an economic reason for not raising it to £21k – politically, politicians wanted to be able to say that the post-2012 students would repay less each month and economically, the Treasury would not have allowed us, I suspect, to raise the £15k to £21k. But the way it affects your loan balance is fascinating.

      I have written at length on other aspects of the Coalition’s HE reforms, which may be of interest – see


      There’s also several other things that evidence the Government being unwilling to alter the existing system too much (and the intention to sell was probably a contributing reason for this) which I pointed out on Nick’s piece:

      In Parliament in 2010, a Liberal Democrat MP asked: “If raising the repayment threshold is to benefit every single graduate, in the Minister’s [David Willetts’] words, can he confirm that current students—and indeed, current graduates—will see their repayment threshold raised also?”
      In reply, David Willetts, as universities minister said: “These are proposals for the future, which come in for 2012. They are not retrospective changes, and for current graduates the existing regime will not be changed. This is only for the future from 2012 onwards. I am grateful for this opportunity to make that clear.”

      David Willetts made the point of emphasising in a letter to NUS that the Government had made no changes to the pre-2012 lian terms other than for a “pre-planned” annual increase in the repayment threshold by RPI:

      Click to access Willetts.pdf

      As I posted elsewhere the following FOI releases (submissions to the ministers) give an excellent insight into the narrative of events:

      Click to access Annex%20A%20Repayment%20Threshold%20Redacted%20FOI.pdf

      Click to access Gibney%2020936%20Annex%20B.pdf

      • Dan permalink

        In Australia the HECs (Higher Education Contributions) repayment threshold is to be lowered to $42,000 from over $55,000 for all graduates. I’m not sure if this differs much from the New Zealand and UK schemes of income-contingent loans but it underlines how it isn’t the norm as in the UK to run separate repayment thresholds for different cohorts of borrowers. There’s usually one repayment threshold for the overall scheme that is amended like tax thresholds from time to time for everyone.

        The problem I have with the system in the UK is that there is inconsistency in who new proposals are applied to. There’ve been instances of applying new terms to all borrowers and also instances of applying new terms to only new borrowers. And with this lack of consistency there is little incentive for Government to apply more generous terms for all borrowers as the more loans there are to apply it to the more expensive it is. So it is all too easy for a Government to come along with this attractive policy of reducing student loan repayments by raising the repayment threshold so all those in work get excited only for them to announce that it only applies to future starters! This is a particular problem if the Government adopt a structure of periodic reviews of thresholds. It’s all too tempting – as the Government did in 2010 – to act on a recommendation to raise the threshold from £15k to £21k as it hadn’t been raised since 2005 and twist it into a comparison of “the new system will be better than the old one” leaving existing borrowers still with the lower threshold which prompted the recommendation to increase it in the first place! I can’t help but think had Labour still been in power when the Browne review was released that they’d have made different decisions and made changes to the existing system (like they did in 2005) making the threshold £21k for existing borrowers too, rather than flat out refusing to change the existing system. There certainly wouldn’t have been this political nonsense of “that is Labour’s system and Labour’s threshold and this is ours and ours is better”…

      • Brian permalink

        Just noticed that RPI for April has been announced today at 3.5% and although this rate doesn’t affect student loans (the 3.1% from March is used) it’s a sign of things to come I think and if March 2018’s RPI is 4%+ then we’ll be seeing the first 7%+ rate for UK student loans since the very first year of mortgage-style loans in 1990/91 when the rate was 9.8%!

        If that happens then I can see two things resulting:
        (a) the media will slate the interest rate even more than this year; and
        (b) it will raise the pre-2012 threshold from £18,330 in 2018 to above £19k in 2019 which means it is rapidly closing in on the £21k post-2012 threshold, making it much more likely that the thresholds will be merged into a single threshold in the early 2020s. I can see no desirable reason for running two thresholds that are barely different from each other…

  5. Yes – we also had the conversation about repayment thresholds and different terms for pre-201 and post-2012 loans back in January on the same post that Brian has linked to.
    Briefly, I believe the current government’s intention is to freeze the post-2012 threshold until the pre-2012 catches up and then link the post-2012 to RPI.

    On the other hand, the interest rates for pre-2012 are set in primary legislation to be the lower of bank base rates plus 1 pp or RPI. For post-2012, the coalition gave governments leeway to set anything up to the equivalent of commercial rates.

    The main reason why things are not equalised across the board is that graduating debt is not the same and the combination of parameters (threshold, interest, repayment rate above threshold, write-off period and graduating debt) have complex interactions. The main idea behind the post-2012 combination was to mimic a proportionate graduate tax – see the graph from IFS in the 12 May post on the cost of abolishing tuition fees.

    On the subject of RPI – we do have a central bank that’s meant to target 2% inflation … pre- and post-2012 loans were not designed with the current macro state of affairs in mind (but for the latter on the assumption that we would return to the trends of the ‘great moderation’ much more quickly)!

    I don’t think the threshold freeze was justified for existing borrowers but I do think a revision of interest should be considered for September (when March’s RPI is due to determine the rate applied for AY 2017/18).

  6. Dan permalink

    Two new FOI releases show just how far out macroeconomic forecasts have been so far, with knock on effects for budget calculations and policy decisions:

    An annual RPI forecast of 2.8% was used for 2010 onwards:

    An annual average earnings forecast of 4.7% was used from 2010 onwards:

    For most of the last 7 years RPI has been higher than earnings growth, and that looks set to continue. As has been correctly noted previously, the £15,000 threshold was not uprated by 2.8% annually from 2010 (which the budget calculations assumed would happen); this RPI-uprating wasn’t eventually implemented until 2012 so the threshold is still lower than was assumed it would be.

    The forecast compared with the actual threshold amounts were (annual growth of 2.8% rounded up to the nearest £5, as per the regulations):-
    April 2010: £15,420 (actual: £15,000)
    April 2011: £15,855 (actual: £15,000)
    April 2012: £16,300 (actual: £15,795)
    April 2013: £16,760 (actual: £16,365)
    April 2014: £17,230 (actual: £16,910)
    April 2015: £17,715 (actual: £17,335)
    April 2016: £18,215 (actual: £17,495)
    April 2017: £18,730 (actual: £17,775)
    April 2018: £19,255 (actual: £18,330)

    So borrowers are currently repaying £85 per year more (9% x (£18,730 – £17,775)) than was assumed based on threshold levels. Obviously average earnings have grown far less than 4.7% per year since 2010 so taking earnings levels into account, borrowers will be paying far less overall than was originally budgeted for both pre-2012 and post-2012 loans. The decision to freeze the post-2012 threshold relaxed the pressure on that budget; the delayed implementing of the £15,000 threshold uprating had already offset the pressure on that budget.

  7. Brian permalink

    After the below submission to ministers in March 2009 cancelled the expected increase by RPI to the student loan repayment threshold for April 2010 and replaced it with an intention to review for April 2011 with an expectation that this would uprate by RPI from April 2011 instead

    Click to access Annex%20A%20Repayment%20Threshold%20Redacted%20FOI.pdf

    the HE minister at the time, David Lammy, responded to a parliamentary question on the decision to postpone the increase in the student loan repayment threshold for 12 months from April 2010 by confirming that the Labour Government would be considering the threshold for April 2011 early in 2010

    It was clearly the intention of the Government at that time (notwithstanding the forthcoming recommendations in the Browne Review which would probably not have recommended a one-off increase to £21,000 had the threshold already been shown to be rising, as that was the reason given for the significant increase – the fact it had remained at £15,000 since 2005) to start uprating by RPI in April 2011.

    Using actual RPI figures (rather than forecast in the post above) the threshold history would be:
    April 2010: £15,000
    April 2011: £15,660
    April 2012: £16,490
    April 2013: £17,085
    April 2014: £17,650
    April 2015: £18,095
    April 2016: £18,260
    April 2017: £18,555
    April 2018: £19,135

    So borrowers are currently repaying £70 per year more (9% x (£18,555 – £17,775)) that if RPI-uprating had started in April 2011, rather than April 2012.

    I’ve also been following developments on the interest rate with interest (ignore the pun!):

    The IFS report “Higher Education funding in England: past, present and options for the future” (, published July 2017, highlights “…the impact of the positive rate above RPI, and the use of RPI itself to index interest rates.”

    It acknowledges that while there is a benefit to the system of applying a positive real rate of interest there are 2 problems with the current formula of RPI plus up to 3 percentage points.

    (1) The Browne Review “Securing a sustainable future for higher education: an independent review of higher education funding & student finance” ( contained the recommendation (page 35) to set a positive real rate of interest in line with the Government’s cost of borrowing (which was then – in line with the discount rate – assumed to be RPI + 2.2%):

    “Students with higher earnings after graduation will pay a real interest rate on the outstanding balance for the costs of learning and living. The interest rate will be equal to the Government’s cost of borrowing (inflation plus 2.2%).”

    It is this recommendation that led to the current variable formula of RPI + 0-3% for post-2012 loans and subsequently RPI + 3% on postgraduate loans.

    The real rate of up to 3% came from the real rate of 2.2% which was the then Government cost of borrowing.

    That 2.2% figure has now been replaced with 0.7%, as of November 2015, thereby making the 3% real rate far in excess of the Government’s cost of borrowing the original formula was meant to reflect based on the recommendation.

    (2) The IFS report (page 24) highlights that “RPI has been shown to systematically overstate the rate of inflation by around 1 percentage point.”

    The Browne Review recommendation, if made today, would read:

    “Students with higher earnings after graduation will pay a real interest rate on the outstanding balance for the costs of learning and living. The interest rate will be equal to the Government’s cost of borrowing (inflation plus 0.7%).”

    The IFS report notes that “…if the government were to try to extract more money from graduates, there are numerous parameters in the student loan system [such as reducing the repayment threshold in real terms etc.] that it can draw from. Different parameters will have different impacts on graduates, but it is important to recognise that the contributions of the highest earners are already considerably higher than the amount they borrow.”

    Jo Johnson has stated in an article in the Guardian on 4th July 2017 ( that “we will continue to look at the details of the student finance regime to ensure it remains fair and effective.”

    Given that the interest rate formula of RPI + 3% is excessive in meeting the “inflation plus 0.7%” Browne recommendation on two counts – the overstating impact of RPI compared with CPI and the 3% real rate on top, surely the interest rate formula should be calibrated.

    With RPI spiking above 3% in March 2017, and possibly above 4% in March 2018, there have been many calls for the interest rate to be reviewed in the short-term with the high interest rate becoming almost as toxic as tuition fees for the Government – and needlessly so, as they gain little from setting it so high. A real rate is fair enough (after all, Browne recommended it) but an “unfair” real rate is not, and the current formula is unfair on two grounds, as I’ve pointed out.

    In the short-term to make the interest rate “fairer” the 3% real rate could be reduced to closer to the Government’s real cost of borrowing of 0.7%.

    A longer-term reform of the interest rate is to replace use of the outdated and discredited RPI with CPI. Put together this meets the policy objective based on Browne of “inflation + 0.7%” and would also contribute to a more sustainable system as the repayment threshold (for pre-2012 loans) is currently set to an annual RPI increase which would then change to an annual CPI increase to be consistent with the use of CPI in the interest rate formula. This reform could take place following the review of the £21,000 post-2012 threshold in 2021 where it could be decided to maintain the post-2012 threshold at £21,000 and then adopt the pre-2012 threshold for all loans when its real terms value reaches £21,000, at which point it would increase annually by CPI. This has the effect of extracting more repayments (as more graduates are repaying each year due to the threshold increasing by less than it would with RPI even which would more than offset a small number of higher earners accruing and repaying slightly less due to CPI rather than RPI being used to set a “fairer” interest rate – as IFS pointed out “…it is important to recognise that the contributions of the highest earners are already considerably higher than the amount they borrow” and this would remain the case after ditching RPI for CPI).

    Clearly, this approach would be a win-win for the Government
    ( as it would recoup more in repayments while being perceived to be fairer and more sustainable as:
    (a) more borrowers would repay in full, thereby reducing the scope for politically toxic headlines like three-quarters of borrowers not being able to repay in full makes the system unsustainable ( and the interest rate is “eye-watering” (
    (b) the interest rate is “fairer” (not only being based on a credible inflation index but more accurately following the cost of borrowing to Government); and
    (c) the repayment threshold is still seen to be rising (although over time, not by as much as it would if RPI was used throughout the system, meaning more repayments come in as there are a vast pool of lower to middle earning graduates who would repay a greater proportion of their loan which more than offsets the extra (most would argue unfair and unreasonably high) interest repayments relinquished from a smaller pool of higher earners, overall improving both presentation and sustainability.

    The original interest rate formula was set in a different context in 2010: it was before RPI was discredited by ONS due to its statistical calculation failings following changes made to its coverage and RPI was only used in any case to maintain consistency with the use of RPI for existing loans; and before the Government’s cost of borrowing changed from 2.2% to 0.7%. The exorbitant interest rate is becoming politically toxic. Tuition fees I agree with; an unreasonably high interest rate I don’t. The Government have flexible powers under section 76 of the Education Act 2011 to vary the interest rates applied to post-2012 (and postgraduate) student loans ( It’s time to start using those powers. Jo Johnson’s words that “we will continue to look at the details of the student finance regime to ensure it remains fair and effective” need to turn into actions before the Conservatives will be getting my vote in the next election.

    • Dan permalink

      Swapping RPI for CPI (or CPIH now ONS seems to prioritise that measure in reporting) is a great option – kills 2 birds with 1 stone:

      Interest rate is lower and more palatable which is far better presentationally (more graduates repay in full improving perceptions of sustainability) and also the repayment threshold rises at a slower pace (earnings growth should exceed CPI in normal years, gradually increasing the amount repaid over time which is the cornerstone of a sustainable system) but crucially is still seen to be rising with the net result of the two being a net saving due to the extra repayments from the bulk of borrowers outweighing the loss of the unreasonably high negative RAB currently levied on high earners which could be lost anyway if high earners opt to (re)pay early.

      Jo Johnson take note – this is an easy gain presentationally:

      Keep post-2012 threshold at £21k until pre-2012 threshold catches up (likely by April 2022 or 2023) then amend the RPI inflation measure to replace it with CPI/CPIH and continue annual increases following CPI.

      In meantime lower max rate on post-2012 and postgrad loans to RPI + 2% to better reflect Government cost of borrowing of RPI + 0.7%.

      Result is you end up with a better system all round by the early 2020s:
      One repayment threshold uprated by CPI (significantly simplifying admin and complexity for employers/HMRC/SLC/borrowers);
      and a max interest rate of CPI + 2% which is equovalent to around RPI + 1%, approximating the Government cost of borrowing.

  8. As I’ve said before, the Browne review recommendations on interest were not accepted (nor was its recommendation to have no cap). Browne recommended a real interest rate, but one where unpaid monthly interest was not added to the account (ie the real element was not compounded – only RPI was compounded).

    I also don’t agree with the IFS that the reporting discount rate for student loans is a ‘cost of borrowing’. It’s much more complicated than that: the interest rate taper is primarily about mimicking a progressive graduate tax, not about tracking the discount rate. That the discount rate was reduced to RPI +0.7% was mainly a piece of political opportunism – it undercut parliamentary discussion about the unsustainability of the system. The main discount rate for assets other than student loans is still RPI + 2.2% and the social time preference rate that will be used to judge whether a student loan sale achieves value for money is RPI + 2.5% + another undisclosed element to reflect risk. Bottom line – the RPI +0.7% discount rate is not the cost of borrowing – that’s even lower at around about 1.3% (ie well under RPI)! (Again, don’t trust the Browne review on these matters – most of what it suggested was rejected.)

    I agree that it is problematic that RPI is used in these calculations. RPI is not an official statistic anymore. But indexing the threshold to RPI instead of CPI benefits current borrowers.

    You both seem keen on extracting extra repayments from borrowers – for presentational reasons? But the measures suggested are regressive, either only benefiting high earners (who could anyway make extra repayments without penalty) or taking more repayments from middle earners.

    What I think the IFS has got wrong – see the latest blog post – is that the government wants to encourage early repayment because it changes the cashflow trajectory and thereby helps to bring down public sector net debt in the short-term. The Treasury worries about this more than whether the scheme loses the equivalent of 30p or 35p in the £ over 35 years.

    I agree though that the media coverage of the interest rate will be considered very seriously by the government, because it has become politically toxic, but as the IFS makes clear it benefits the highest earners most in concrete terms.

  9. Brian permalink

    What I’m saying is that if RPI is to be axed (and why shouldn’t it be frankly as it’s a flawed statistic?) then the perfect time to do it for the government presentationally (which is obviously important politically) would be when they review the post-2012 threshold.

    At the minute the ‘fuss’ on the threshold is about it staying frozen rather than it rising each year with “inflation”. The initial proposal of it rising with “earnings” (as has been acknowledged in previous reviews of the pre-2012 threshold: was enormously expensive given that earnings rise – over the long term – faster than prices, but if the threshold was seen to be going up by any measure of price inflation I don’t think anyone would bat an eyelid that it wasn’t going up with earnings as “rising with inflation” is what the people talk about anyway – I’ve seen practically anyone who talks about the threshold staying frozen talking in terms of it should go up with “inflation” – only a very few wonks talk about it in terms of the initial proposal of it “rising with earnings”, yet there’s a massive difference economically so it’s an easy gain presentationally to link it to inflation rather than earnings.

    It has also long been criticised that RPI is used to set interest for student loans when it is the highest – and now least credible – of the possible choices of inflation index, and when the government use CPI to uprate practically everything else that they pay out:

    “Mr Nick Raynsford (Greenwich and Woolwich) (Lab): The Minister referred to the interest rate that will be reclaimed as 3% plus inflation without being precise about which measure of inflation he was adopting. I think I detected from a subsequent answer that he was proposing to use the retail prices index. Given that the Government propose to substitute the consumer prices index for the RPI in relation to a huge range of benefit payments that involve the Government paying out, why is he proposing to adopt the RPI in this regard, particularly considering that he has talked about trying to minimise repayment costs for students?

    Mr Willetts: We are, of course, following the precedent set by the previous Labour Government.”

    “…the Government use the measure of RPI, when the rest of the time they tell us that the consumer prices index is the correct measure of inflation. They seem to hold two contradictory positions: when they are paying money out to their citizens—in the form of benefits, for example—they say that CPI is the correct measure of inflation, but when they are collecting money, they say that RPI is the correct measure. Holding two contradictory positions at the same time is what George Orwell called “doublethink”. The Government seem to believe in both.” (Helen Jones in the student loans debate).

    It is not “regressive” when you consider the system holistically. The overall repayment profile would remain progressive and it would also be more common sense and politically and economically sustainable when you look across the system as a whole and consider ALL types of borrowers, not just the “post-2012 undergraduate first degree” type of borrower.

    For example, the RPI to CPI change would benefit all those lower and middle earning borrowers with smaller loan balances such as those who have postgraduate loans, and those who borrowed under the pre-2012 loans for their first degree and subsequently for a short course such as PGCE teaching under the post-2012 loans ( as it would allow more people to repay in full before 30 years are up. It is shortsighted to just treat all lower and middle earners in all circumstances as repaying 9% above the threshold for 30 years when the system isn’t as simple as that and when the interest rate can effect lower and middle earners by increasing the length of time spent in repayment. It’s common sense and fair to give more people the opportunity to pay off their loan in full and more quickly and I think this policy change strikes the right balance.

    You can maintain an incentive for early repayments (and even with CPI + real interest, the interest rate would still be on the high side, just more reasonable) without catching the above types of borrowers in the same net with the interest rate. For example, if the Treasury wants to incentivise early repayments then there are other ways to do that – more joined up thinking would make use of the new Lifetime ISA:

    The Dearing Review (paragraph 21.60, “…examined two approaches to incentivising savings for higher education:
    • a savings scheme with tax relief, as now available for pension plans, such as a modified Personal Equity Plan (PEP) or a Tax Exempt Special Savings Account (TESSA);
    • a long term savings bond, with the incentive of a Government contribution when the bond is cashed in, provided it is used to support learning.”

    The new Lifetime ISA fits both of these criteria, in that it is a tax-exempt savings account (i.e. ISA, the successor to TESSA) and receives a Government contribution of 25% on deposits. The Department for Education could suggest to the Treasury that funding higher education is worthy of being one of the ‘life events’ that are exempt from withdrawal charges on the Lifetime ISA, thereby incentivising early repayment if the incentivise to clear the loan balance was there (which it would be for people if the interest rate wasn’t so high) and indeed contributions towards higher education from more than just the borrower.

    Withdrawals from the LISA would be required to be sent to the Student Loans Company to repay a student loan; in this way the amount of charge-free withdrawal is proportionate to the support a student was assessed by Student Finance as in need of receiving, whilst also allowing compliance to be policed. Early repayments keep a lid on the RAB charge, and in turn support a strong foothold on the sustainability of the system. Allowing people (including family) to save up for their education in a LISA by designating student loan repayments as a ‘charge-free’ event for withdrawals allows private funding and Government support to be further combined to pay for education (Government support comes in the form of LISA bonuses or write-offs, and private contributions to LISA used to repay student loans would exceed the bonus element of the fund), exactly as Dearing envisaged. This revives an idea on page 17 of the 2010 Conservative Party Manifesto ( which pledged to “provide 10,000 extra university places this year, paid for by giving graduates incentives to pay back their student loans early on an entirely voluntary basis.” Now the student numbers cap is off completely, it is paramount that the system is sustainable and allowing LISA savings to pay for education through contributing to repaying student loans is giving graduates an early repayment incentive (the LISA bonuses effectively act as a discount) which is exactly what was pledged in the 2010 Manifesto.

    If employer contributions were also allowed into a LISA, this would also allow for an implicit employer contribution to supporting the costs of a student’s higher education which was part of the tripartite recommendation in the Dearing Review; paragraph 18.26 stated:

    ‘the general view on funding coming through the submissions is that the responsibility for funding the system should be shared between the government, students and employers; that the funding arrangements should make possible such a balance; and that the funding system should help increase the flexibility of the system.’

    Since ICR were first introduced in 1998, so far no way has been found to enable/require employers to meet their component of contribution, alongside students and taxpayers as one of the beneficiaries of higher education.

    For this joined-up policy making to work, the incentive must be there to clear the loan balance, and reducing the interest rate would create a greater pool of borrowers who are motivated to clear their loan balance – the importance of this must not be understated. Vroom’s expectancy theory looks at what makes us motivated and stated that Motivating force = Valence x Expectancy. By ‘valence’ he meant how much you want the outcome that could be achieved by a certain course of action. The ‘expectancy’ is your mental expectation, your assessment of a probability that you’ll actually achieve that result, and he said both of these are taken into account in the motivational process.

    In the case of student loan repayments, applying Vroom’s theory translates into a motivation to repay depending on the desire to repay in full and the realistic likelihood that a borrower feels they have of doing so. As David Willetts also pointed out on page 39 of his policy pamphlet
    (, “…we have learned that slow payback is not particularly popular. ‘By a margin of almost two to one, undergraduates and parents would rather a student loan is paid back quicker, with higher monthly repayment, than longer, with smaller monthly repayments.’ […] Parents were more concerned about the size of the loan (64 per cent) than terms of repayment (29 per cent) and so strongly preferred the lower repayment threshold (44 per cent versus 36 per cent).”

    It is clear that many borrowers and parents actually WANT to have the chance of repaying in full but only if they are given a fair crack of the whip and feel they realistically have a chance of doing so. Sadly with the continuing use of RPI in an interest rate formula set at up to RPI+3%, for many, if not most borrowers, this immediately kills any motivation to repay in full as the interest applied is just too high to even be offset by statutory minimum repayments for a large group of borrowers (this is the case with 9% deductions for post-2012 borrowers and may even be the case with lower 6% deductions for those with only a master’s loan as everyone faces RPI+3%, not just those earning
    above a higher interest threshold, and certainly will for those with a doctoral loan and those with a combination). A small but very significant psychological change from RPI (which even without 3% added on top is certainly high in the current environment compared to Bank of England base lending rates and those available commercially in the market – the ‘better than commercial rates or on better terms’ requirement of section 22 of the Teaching and Higher Education Act 1998 (brought in by section 76 of the Education Act 2011) which stemmed from the Consumer Credit Directive is based on credit card levels as pointed out by Intergenerational Foundation here ( to CPI would be the one policy change that would have a lasting change on motivational behaviour in aiming to pay off the loan in full (for example through extra voluntarily payments from time to time and not participating in salary sacrifice schemes and other repayment avoiding techniques such as having 2 jobs where the threshold is legally applied in full to both, to reduce or stop student loan repayments; or indeed evade repayments by leaving the country). We constantly see – e.g. as recently as last night’s Question Time – how demoralised it is to see you’ve made repayments that haven’t even reduced the loan balance.

    Page 328 of the Dearing Review considered the problems surrounding a real rate of interest on student loans: “we found that moving to a rate of interest which is close to a commercial rate of interest (the current [mortgage-style] scheme has an interest rate equivalent only to the Retail Prices Index) created certain problems. The protection of income contingent arrangements for the low paid means that a significant minority of graduates would be making repayments which did not even cover the interest on their loan, let alone repay the debt. Their debts would, therefore, continue to grow through life becoming, in some cases, very large before write-off. Even though individuals would be protected from unreasonable debt-servicing burdens, we felt that the possibility of an ever-rising debt would be a deterrent to participation in higher education. Although the same difficulty can arise with any real rate of interest, it is severe with commercial interest rates. We were told that the Australians had considered real interest rates for their Higher Education Contributions scheme, but had rejected them for the same reasons.”

    Now obviously policy and evidence has evolved since the Dearing Review, so while there is little evidence of ever-rising debt being a deterrent to participation in higher education where there are the safeguards of a 30 year write-off in place and repayments are income-contingent above a high threshold, the level of the real interest does impact borrower behaviour where their loan balances are low enough to be in ‘repayable range’ given an average salary path. Most postgraduate loans (but certainly for master’s only) would be in repayable range with a slightly less extreme and also fairer interest rate more in line with Government’s cost of borrowing, and the same is true of the maximum rate for post-2012 balances where borrowers have taken a short course of study or withdrawn.

    While the Browne Review recommended a real rate of interest (except for lower earners), it only recommended “the interest rate on the loans is the low rate that Government itself pays on borrowing money” as I said – and although the Government didn’t accept this proposal its its exact form, it was clearly used to inform the formulation of a real interest taper.

    A switch away from RPI would therefore be fairer for borrowers and the overall sustainability of the system; the benefit is threefold: an interest rate more closely aligned with Browne’s recommendation with an inflation index that meets international standards, more borrowers feeling they can repay their loan in full (both increasing perceptions of fairness politically and economically) and a more slowly rising repayment threshold (the latter two potentially reducing the RAB charge due to the increase in repayments from both borrowers earning over the repayment threshold and those who make extra payments in the desire to clear their loan balance which they otherwise would’ve left to be written-off at significant cost to Government if they felt they wouldn’t clear the balance due to the extortionate interest).

    Implementing the switch from RPI to CPI in the early 2020s upon review of the post-2012 threshold would also be soon around the point that the last of the old mortgage-style student loans would be due to expire (ordinarily 25 years from the last loan agreement) as these loans ceased being issued for a small number of starters in 1998 (under transitional arrangements) and the change from RPI could not apply to mortgage-style loans due to this index being explicitly stated in some of the later credit agreements ( and regulations applying to these loans for earlier credit agreements ( It can’t really be justified to continue using RPI in ICR loans for the sake of a few outstanding mortgage-style loans. The Government could continue to set the RPI figure for these old loans separately alongside the more palatable inflation figure for ICR loans until the last of these loans expires.

    RPI has been relatively low recently, resulting in a maximum interest rate of RPI+3% that has been on the high side, but still begrudgingly accepted by borrowers who, after all, haven’t been in repayment and so hadn’t had the demoralising and demotivating experience of seeing the interest far exceed their repayments. However, as inflation rises and borrowers see 7%+ interest getting applied to their loan balances, it is increasingly indefensible (and potentially unsustainable in that it reduces the number of borrowers who will ever clear their loan balances) to continue with the policy of imposing a maximum rate of RPI+3% on borrowers’ loan balances, as if nothing else, it is now well above the Government’s long-term cost of borrowing and results in lower repayments from the majority of borrowers (due to a more rapidly rising repayment threshold) and less borrowers fully repaying (or having any chance whatsoever of doing so). For a switch that increases sustainability and value for money, it would be a very popular decision among students.

    The only potential stumbling block I can see for a switch from RPI to CPI is that it could (although highly unlikely) be questioned whether the Government could legally do that for transferred loans (loans that have been sold), as although the maximum interest rate would likely be lower, crucially the repayment threshold would also likely be lower over time than before the amendment (crucial for sustainability), and Section 5 subsection (6) of the Sale of Student Loans Act 2008 ( states:

    “But in amending loan regulations the Secretary of State shall aim to ensure that no borrower whose loan is transferred is in a worse position, as the result of the amendment, than would have been the case had the loan not been transferred.”

    Of course, it is far from clear-cut whether early pre-2012 borrowers (which would form the initial tranche of loans to be sold) would be worse off anyway if the regulations were amended to use CPI in place of RPI, as their loan balances are not cancelled until age 65, which means many repay in full, and nearly all of these loans do not consist of tuition fee loans (only smallish maintenance loans) which weren’t available until 2006; therefore a switch to a lower inflation index results in
    potentially lower interest applying (notwithstanding the operation of the low interest cap) and therefore lower total repayments, even though it would mean a more slowly rising repayment threshold, and so some would be in a better position as a result of such an amendment to the inflation index used as they repay quicker and therefore incur less interest. The language of the legislation is pretty loose in any case – “shall aim to ensure” – and of course the amendment would have nothing to do with the sale of the loans as it equally applies to unsold loans and so the subsection isn’t breached as they’re in no worse position than had the loan not been sold.

    The extra repayments brought in from the threshold rising more slowly with CPI would reduce the long-term costs of the system, reduce the politically toxic headlines by lowering the interest rate to more sensible levels, finally remove all the flaws of criticisms of using RPI and encourage more borrowers to repay in full without hurting those lower and middle earning borrowers in some parts of the system who have smaller loan balances are kept in repayment longer due the “very high” interest.

    Politically, the presentation of the policy switch from RPI to CPI could be presented could be presented as favourable to borrowers if it is announced at the time of the review of the post-2012 threshold as it would end the freeze on the threshold, increasing it annually by CPI (which meets most people’s expectations of “rising with inflation”) and also reduce the interest rate. In sum, win-win!

    • I’m srry to have to remind you, but the threshold for pre-2012 was fixed to an RPI index to facilitate a loan sale (this is why the RPI index is now in place, whether it tallied with a pre-2010 Labour promise or not). RPI is being used because potential private purchasers will be pricing student loans using RPI. Mortgage-style loans have had their threshold calculations set in sale agreements and so are not subject to change.

      And only pre-2006 ICR loans are written off at 65, other pre-2012 ICR loans last 25 years.

      For real interest rates, it is Nicholas Barr’s work that was influential, not Browne. See “Interest Rate Subsidies: a better class of drain”.

      I think you are missing the point – borrowers have the option of making additional repayments should they desire (though it’s not often in their interest to do so), increasing mandatory repayments removes that flexibility and makes lower and middle earners pay more. If borrowers would prefer to make higher payments they can now – I’d suggest the concrete evidence is whether they do or not.
      The solution to the psychological problem is not to make people pay more, which just lowers disposable income.

      From the government perspective, ‘fully repaying’ means repayments exceed the equivalent of what was originally loaned using the RPI plus 0.7% discount rate. It does not mean that the account need be cleared. The sustainability of the loan scheme does not depend on balances being cleared in that sense. The sustainability of the scheme is a function of repayments generated compared to what I lent – it’s not determined primarily by what percentage of accounts are still open when loans are written off. The scheme is not (financially) unsustainable because 75% of accounts will still be open at write-off. It all depends on what repayments have been generated (and when) and what “loss” is budgeted for.

      It is not desirable to equalise the threshold and interest rates for pre- and post-2012 loans, because they starting debts were so different. As IFS point out – the progressive gains from the 2012 system were lost by freezing the threshold.

      Postgraduate loans are a different matter – and are basically a bad policy.

  10. Brian permalink

    I don’t disagree with you on the economics but on the political side I do. A lot of this is about perceptions and it’s no good having a system where there is little goodwill from stakeholders and constant bad headlines.

    Yes, borrowers can repay early, but at the moment there is little incentive too. Lowering the interest rate and threshold for that matter would create more of an incentive and rejig the system to something more akin to what it’s called: a loan repayment. Yes, for many borrowers with post-2012 loans it ACTS as a graduate tax, but for others (and not just high earners) it is a loan that will eventually be repaid in full.

    Personally, for me, it is in my interest to pay off both of my post-2012 loan balance (which despite getting some maintenance grant is already £16,000 due to sky-high interest) and my postgraduate loan balance which is already £10,460). Therefore even though I’m only likely to be a middle-earner, it is in my interest to have less interest applied so my repayments are switched off sooner than 30 years. And so I am happy to pay more in repayments so I can clear my balance, but I’d rather everyone had to pay a bit more back given that many have received much more maintenance support than I did (due to the support increasing under the post-2012 system).

    It’s all about striking the right balance: some borrowers may want to retrain and do second degrees. At the minute access to further loan borrowing is severely restricted due to the high cost of writing off loans. If there needs to be a trade-off between lowering the interest rate and reducing costs of the system then I’d rather the repayment threshold rose more slowly too. And I think it makes no sense to maintain a lower repayment threshold for pre-2012 borrowers. It’s no longer about a comparison of systems: everyone with an income-contingent student loan who is a “low earner” should not be making repayments. If someone earning £20,000 is a “low earner” than why should they currently be making repayments towards their pre-2012 loan when they don’t have to for a post-2012 loan?

    Post-2012 borrowers will have had a higher threshold than pre-2012 borrowers for a decade between 2012 and 2022 and the lower threshold affects lower earners most, not least because it will keep some of them who go on to borrow for further courses in repayment until 65 and therefore subject to a 9% tax above a lower threshold until 65; and others not getting any benefit from a 25 year write-off as their new post-2012 loan lasts 30 years and without policy changes won’t have a chance for these lower to middle earners of being repaid. It’s also undesirable for HMRC/SLC/Business to administrate and terribly confusing for past borrowers who see the “student loan threshold” of £21,000 scattered all over the media and then say “why am I getting fleeced with higher deductions?” It’s would be ridiculous to run two thresholds so close together for different cohorts. In 2030 it’d be “students start repaying if they earn £22,535 if they started in 2012 or later and £20,465 if they started before 2012”. One threshold is enough complexity, quite frankly.

    On the pre-2012 threshold uprating, I don’t disagree that the RPI uprating was extended indefinitely to facilitate a loan sale as the explanatory memorandum that amended the regulations to remove the time limit on the uprating says so:, but RPI indexing was used in the first place (initially regulated between 2012-2015) as this is what the previous Labour Government had planned: the minister’s letter to NUS ( confirms that “there have been no changes to the pre-2012 ICR loan repayment terms since 2010, other than for a pre-planned annual increase in the repayment threshold in line with inflation (RPI)”, and so does the previous special adviser to this minister (Nick Hillman)’s email.

  11. Pre-2012 loans are currently accruing interest at 1.25% (bank base rate + 1pp). They also had lower commencing debt. You cannot compare the two schemes by simply looking at the repayment threshold – you will only see short-term cash flow implications, not the long-term.

    ICR loans are complex because of this interaction of features. If you are going to worsen post-2012 loans to bring them into line, then they need some compensating in other respects.

    The whole fee-loan approach mitigates against retraining – that’s one of the biggest problems. The complexity is another. I’m in favour of a wholesale rethink – my quibbling here is mainly: I don’t think the tweaks you propose resolve the fundamental problems.

    More generally, I am confused as to why you seem so keen to repay more, but don’t then use the option to make additional repayments. Do we agree on the following difference: lowering the interest rate benefits higher earners most, but lowering the repayment threshold means almost everyone (bar those who never earn above the new threshold) pays more?

    I accept that the psychological benefit of not having the nominal balance race away is significant but the two mechanisms (interest rate and threshold) are very different in their effects.

    I would rather more debate was about the level of likely repayments rather than whether balances are cleared. Since having more people clear their balance can be achieved by making higher education more costly (having people repay more) or less costly (lowering initial debt & interest).

  12. Brian permalink

    “I am confused as to why you seem so keen to repay more, but don’t then use the option to make additional repayments.”

    I’m not keen to repay more in total – that’s the point. No-one, including me, can be certain this far out whether they will repay the full balance and the amount of time spent in repayment directly impacts the total amount ultimately repaid.

    As I’ve said before I have a 25 year write-off of my pre-2012 loan and a write-off of my post-2012 loan scheduled for 8 years later than this. So if my post-2012 balance isn’t cleared by the 25 years, the supposed benefit of the write-off is useless as it only writes off part of my overall loan (the favourable interest rate on pre-2012 loans would also prove useless to me) and I continue 9% repayments for up to a further 8 years (making 33 years in total). This increases the total amount I would repay RELATIVE (yes, I’m comparing fairness here) to someone who borrowed much more than me but only makes repayments for 30 years above a HIGHER threshold, meaning despite being a lower to middle earner, I repay more in total and have less disposable income year to year than the latter borrower with my income would. Some with a pre-2006 loan who takes out further borrowing under the post-2012 loans would be even worse off as they’d potentially face deductions above the LOWER threshold up to 65 as the post-2012 loan increases their smaller debt on a massive scale.

    This is what I mean when I talk about equalising thresholds. With the wide ranging scenarios it is surely fairest to to set repayments that are defined as “affordable” for all borrowers regardless of total debt? The only way to set this annual repayment amount is through the threshold and the repayment rate. That is the what those two features are designed to do – make ongoing repayments affordable on a monthly basis.

    You say “if you are going to worsen post-2012 loans to bring them into line, then they need some compensating in other respects.” I don’t think they do – because they were already more generous relative to pre-2012 loans with maintenance support available in the first place. Pre-2012 borrowers may or may not benefit from a write off at 65 or 25 years depending on whether they clear their balance. The amount they repay in total relative to someone with a post-2012 loan could be more or less. The various scenarios are complex and that is why I think it is fairest (because you can’t set different parameters for each individual permutation) for a single threshold to operate which has the sole purpose of defining what the Government of the time deem is “affordable” at the time.

    Making additional repayments now transfers the risk that I wouldn’t repay the lot even after 33 years (or take out even more borrowing in the future under this or any other system or variant that is brought in) wholly onto me. Given its a finely balanced risk I’d rather be in the position of having the Government take that decision to increase my ongoing repayments but reduce the interest rate so I don’t feel like I’ve chucked money away needlessly in early repayments and so it increases both the likelihood of me finishing repayments after 25 years and is “good” for the system overall, looking at it from a taxpayer perspective.

    “Do we agree on the following difference: lowering the interest rate benefits higher earners most, but lowering the repayment threshold means almost everyone (bar those who never earn above the new threshold) pays more?”

    Looking at the borrower who only has borrowed only on the post-2012 system with circa £50k of debt, the interest rate reduction reduces repayments from a smaller pool of high earners, and the repayment threshold reduction over time increases repayments from lower to middle earners (i.e. the bulk). You could view that in isolation as “regressive” but you could equally view it as “reasonable” if that’s how the system was already set up as overall repayments remain “progressive”. High earners are already returning a surplus of repayments if they don’t pay early so in my view, they are paying enough – i.e. what it is fair and reasonable – after the reform I propose, and still more than everyone else.

    If you want high earners to pay through the roof then you can introduce a full-on graduate tax, but this would just encourage these graduates to emigrate or try and avoid/evade the tax – neither is desirable.

    But the “regressive” impacts of the reform only hold when you are looking at the borrower who only has borrowed only on the post-2012 system with circa £50k of debt, not in all cases. For other types of borrower, like me, on lower to middle earnings, it is “progressive” in bringing forward the date of my write-off subsidy. Overall, considering all situations, I think it’s reasonable and is why I say across all economic and political considerations, it strikes the right balance.

  13. But the value of money is not stable. A £ today is not the same as a £ in 20 years’ time. By thinking in terms of cash totals, instead of present value, you are misjudging the value of repayments. You keep talking about the extra years of payment at the end rather than the value of repayments overall – earlier repayments are worth more than those of the same future cash because they are nearer to today.

    You can fix the high earners repaying more than they’ve borrowed (in NPV terms) by lowering the interest rate. That’s very different from what you’re proposing around the threshold.

    I don’t think we can fix the complexities of your situation by penalising new starters.

  14. Brian permalink

    I know the value of money isn’t stable – I can see that when a load of RPI interest + 3% more is bunged onto my post-2012 and postgrad loans!

    Reducing the repayment threshold over time could increase the total amount I repay; it could also decrease it. But you must admit it’s hardly fair to just dismiss borrowers in all but a ‘standard’ position as an inconvenience to policy making. And a best fit in my view is having a single threshold. Fine – uprate that threshold by RPI so there is no change from the status quo, but have a single threshold so that all borrowers on a yearly basis repay the same “affordable” amount – this is to avoid the real possibility of creating particularly unfair circumstances of having a situation where lower earnings are paying a 9% tax above a lower threshold for years and years (one of the points of increasing the threshold was to mitigate the impact of increased debt so it is only fair that if any borrower has increased debt and an extended repayment period, they should benefit from the same annual repayments).

    Yes, some people without pre-2012 loans will make repayments in years where pre-2012 borrowers might not have a loan balance, but equally post-2012 borrowers could give themselves a 25 year repayment period by spending 5 years below the threshold or out of work during the latter repayment years/early retirement etc. and yet still have benefited from the increased borrowing available to them in the first place. On the other hand pre-2012 students had less maintenance borrowing available to them so if they clear their loan balance within 25 years then it’s fair enough as they borrowed less maintenance to start with.

    I’d be quite happy for the interest rate to be reduced to e.g. RPI + 1% to at least give people a chance of eating into their loan balance on a yearly basis which is an important consideration and frequent criticism.

    The reason I prefer ditching RPI is a matter of principle as much as anything. It overstates inflation and is exceeding earnings growth at the moment so reduces the NPV of repayments on both counts. It’s a flawed statistic.

    You could uprate the threshold with earnings (as long as it also covers pre-2012 loans) while ditching RPI in the interest rate, uprate the threshold with CPI for the medium term to embed the system as sustainable and then swap to an earnings based uprating or one-off increase etc. in the long-term.

    When the pre-2012 threshold was reviewed in 2009, it was clearly acknowledged that uprating by RPI may not be politically sustainable in the long-term and envisaged later uprating with earnings.

    Click to access Annex%20A%20Repayment%20Threshold%20Redacted%20FOI.pdf

    It is not my intention to set the threshold at a CPI/RPI increase for evermore but by keeping separate thresholds, this is essentially condemning the pre-2012 threshold to be fixed for evermore as all the focus then would be on managing the post-2012 system, when that is not the intention of income-contingent loans – they clearly aren’t meant to be indefinitely fixed in that way.

  15. I don’t dismiss it at all – but the villain may be in policy governing concurrent loans rather than post-2012 loans itself. I would revisit the decision to slice repayments between your pre-2012 and post-2012 loan. And PGT loans need a complete rethink.

    You should have your own discount rate, which you use to determine what you think a £1 next year is worth to you today – you shouldn’t be using the government’s.

    I don’t think fairness can be cashed out anyway than in terms of the NPV value of all estimated repayments generated by the loans. It’s not a tax because of differing interest, graduating debt & write-off. You suggestion works if it were a tax.

    In the end, balances are a secondary matter – it’s the repayments generated that counts. Obsessing about clearing the balance loses sight of this.

    I’m afraid to say the focus now will always be on the current loans and if the abolition of fees is on the cards, then all existing loan holders may end up feeling hard done by. Even if there is an attempt to abolish post-2012 tuition fee debt – I doubt it would be extended back beyond then.

  16. Dan permalink

    “RPI is being used because potential private purchasers will be pricing student loans using RPI. Mortgage-style loans have had their threshold calculations set in sale agreements and so are not subject to change.”

    This may be so, but there are no plans to ‘fix’ pre-2012 loan terms in a sale agreement – in line with the Sale of Student Loans Act 2008, the regulations that apply to ‘transferred’ (sold) loans will continue to be subject to amendment from time to time, as with all ICR loans, as confirmed by Jo Johnson:

    Click to access 161007_Jo_Johnson_to_Lord_Hollick_response.pdf

    • You need to read to the end of the relevant paragraph. The terms will be stipulated in the sale contract and any government that used its powers to change terms in future would have to compensate purchasers. In practice this means they are fixed. For a precedent see the setting of the threshold for mortgage-style loans. The planned sale is the reason why pre-2012 are indexed to rpi for the remainder of their terms.

      • Purchasers have no power to change terms.

      • Dan permalink

        They’d only have to be compensated if the change was to the loan purchaser’s detriment. In practice I think changes to the terms (amendments to regulations) are likely to be operational in nature for sold loans relating to loan collection and admin rather than the core repayment parameters but that said, changes to the borrower’s detriment are still possible on the grounds that they would have been made if the loan hadn’t been sold – having a single threshold (i.e. a crossover of terms) with post-2012 loans lessens the legal risk for Government as (once all pre-2012 loans have been sold) they can point to recent unsold loans to which the negative amendment would equally apply in the case of lowering the threshold or repayment rate, and not be in breach of the so called ‘protections’ to borrowers (which are loosey goosey) in the Sale of Student Loans Act 2008.

        Talk at the time of the decision to continue the RPI uprating for the lifetime of the loans in 2014 when NUS raised concerns about changes to terms from David Willetts was that a sale makes it less likely, not more, that terms would be changed. And while this may be true, it depends on the sale agreement: just as there could be a compensation condition to the purchaser for making terms more favourable to borrowers, there could be a rebate to Government condition if the terms are made less favourable to borrowers. In this way the risk to loan purchasers is taken away from policy changes and is all about macroeconomic changes.

      • HMRC and SLC will continue to collect and administer pre-2012 loans after a sale. In fact, the government is now explicitly pursuing a securitisation, which means that purchasers are buying a right to a share of an income stream.

        I’m afraid I don’t follow the rest of your first paragraph.

        As far as possible the government will remove the ‘democratic risk’ from any sale agreement. And will fix parameters in advance – viz indexing RPI rather than reviewing it periodically. For an example of a complex contractual idea to mitigate the risk associated with low bank base rates, you can look at ‘synthetic hedges’:

        Also see what happened with mortgage-style loans recently. The law does not specify a way of calculating the repayment threshold, but the sale contract does. So an obvious anomaly to the detriment of borrowers was allowed to stand.

      • Dan permalink

        What I was explaining in my first paragraph is that a loan purchaser would only need to be compensated if the change reduced the amount of repayments they’d receive. If the change increased them, then the government may want to the loan purchaser to compensate the government and such a mechanism could also be built into the sale agreement. That would take the risk to both sides away from policy changes and all about the macroeconomic variables.

        What I was also pointing out is that changes to the borrower’s detriment of sold loans are prohibited by the Sale of Student Loans Act 2008 but only if they wouldn’t have been made had the loans remained unsold. Once all pre-2012 loans are sold, a future government could for example, decide to increase the repayment rate on all income-contingent loans to 10% from 9% which would have to capture all those pre-2012 loans for it to be workable, or indeed reduce the repayment threshold which would have knock-on consequences for pre-2012 loan repayments. As the change is also applying to unsold loans, it can’t really be argued that the change has been made because the loans have been sold, so it doesn’t breach the Sale of Student Loans Act 2008 as although the borrower is in a worse position, it’s not because their loan has been sold.

        The terms and conditions of student loans are made up of those in the loan request form (which remain fixed) and those in regulations made under the Teaching and Higher Education Act 1998 which can be amended, and indeed are amended regularly, as they set out not just the ‘core repayment parameters’ but also all the administrative and operational terms and procedures which change from time to time.

        A letter from a former higher education minister confirms this set up:

        “I should add that borrowers enter into an agreement based on the terms and conditions of their income contingent student loan contract. The declaration by the student requesting a loan says that the borrower has read and understood the guidance booklet and that they agree that any loans made to them will be under the terms set out in the Loan Request Form, and within Regulations which are made under Section 22 of the Teaching and Higher Education Act 1998, as amended from time to time. The Education (Student Loans) (Repayment) Regulations 2009 also form part of the terms and conditions of the contract.”

        The below post links to all the amendments to those regulations (and therefore terms of the loan) that have happened over the years:

        In the report stage of the Higher Education and Research Bill in the House of Lords, the government made clear that the tabled amendment aimed at preventing future retrospective changes prevented both favourable changes for borrowers and also would prevent the necessary administrative amendments to the terms and conditions to ensure that the loans can continue to be collected efficiently. An example of this was the repayment regulations having to be amended in 2012 to accommodate HMRC moving to an electronic system to collect PAYE income tax through employers. Not being able to make this ​type of technical change to the regulations would eventually affect our ability to collect repayments through the tax system.

        Lord Young of Cookham

        Perversely, the noble Lord’s amendment would prevent the Government making any changes to the loan agreement that would favour the borrower. In other words, one of the effects of the amendment would be that we would not be able to alter the terms to the advantage of the borrower if the situation changed.

        Lord Watson of Invergowrie

        As I said earlier, that is what the amendment is designed to do. The point is, when you reach an agreement you stick by it; you do not vary it either way. I am certainly not advocating that it should be varied the other way. My question was whether the noble Lord and his Government would be prepared to vary it the other way, had earnings risen by more than had been anticipated.

        Lord Young of Cookham

        My response was that we would not be allowed to under the terms of the amendment. We have flexibility, which the noble Lord would deny us. The amendment would mean that future cohorts of students and taxpayers would have to bear the risks of the scheme, because it would insulate current students from any change. Perhaps that is why the Labour Party did not legislate to prohibit changes to the terms and conditions of existing loans when they introduced the system of income-contingent loans in the late 1990s. As I said, his amendment would prevent the Government making any change to the loan agreement that would favour the borrower, were this ever to be necessary.

        It is also important that the Government should continue to be able to make necessary administrative amendments to the terms and conditions to ensure that the loans can continue to be collected efficiently. An example of this was the repayment regulations having to be amended in 2012 to accommodate HMRC moving to an electronic system to collect PAYE income tax through employers. Not being able to make this ​type of technical change to the regulations would eventually affect our ability to collect repayments through the tax system.

      • What I was also pointing out is that changes to the borrower’s detriment of sold loans are prohibited by the Sale of Student Loans Act 2008 but only if they wouldn’t have been made had the loans remained unsold. Once all pre-2012 loans are sold, a future government could for example, decide to increase the repayment rate on all income-contingent loans to 10% from 9% which would have to capture all those pre-2012 loans for it to be workable, or indeed reduce the repayment threshold which would have knock-on consequences for pre-2012 loan repayments. As the change is also applying to unsold loans, it can’t really be argued that the change has been made because the loans have been sold, so it doesn’t breach the Sale of Student Loans Act 2008 as although the borrower is in a worse position, it’s not because their loan has been sold.

        I don’t believe that’s how the Sale of Student Loans Act is to be interpreted. Can you show me the section which says prohibitions only apply ‘if they wouldn’t have been made had the loans remained unsold’? I think the test is only whether borrowers are worse off or not – regardless of what is happening to other unsold loans.

        You do realise that this blog has been warning people for 5 years or more that retrospective changes are legal? And that I warned Martin back in 2012 that he was underplaying the likelihood of a government exercising that right to the detriment of borrowers?

      • Dan permalink

        “Also see what happened with mortgage-style loans recently. The law does not specify a way of calculating the repayment threshold”

        Regulation 7 and Schedule 2 to the Education (Student Loans) Regulations 1998 (as amended by e.g. the Education (Student Loans) (Amendment) Regulations 1999) set out how the deferment level for mostgage-style student loans is to be calculated so it is set out in law:

        Terms of loans

        7.—(1) Every agreement for a loan made before or after these Regulations come into force shall include the terms set out in Part I of Schedule 2.

        (2) Every agreement for a loan made before these Regulations come into force shall in addition include the terms set out in Part II of Schedule 2.

        “deferment level” means 85% of the lender’s estimate of average monthly earnings of all full-time employees in Great Britain for the January when the level will apply based on figures published by the Office for National Statistics, or if that Office ceases to publish relevant figures, any other published figures;

      • You didn’t read my piece for moneysavingexpert. That’s exactly the point at issue – the sale contract defined the method to be used to calculate average monthly earnings.

      • Dan permalink

        “I don’t believe that’s how the Sale of Student Loans Act is to be interpreted. Can you show me the section which says prohibitions only apply ‘if they wouldn’t have been made had the loans remained unsold’? I think the test is only whether borrowers are worse off or not – regardless of what is happening to other unsold loans.”

        Should read: ‘if they WOULD have been made had the loans remained unsold.’

        Section 5 subsection (6) of the Sale of Student Loans Act 2008 (

        “But in amending loan regulations the Secretary of State shall aim to ensure that no borrower whose loan is transferred is in a worse position, as the result of the amendment, THAN WOULD HAVE BEEN THE CASE HAD THE LOAN NOT BEEN TRANSFERRED.” (“transferred” means sold to a loan purchaser).

        “You do realise that this blog has been warning people for 5 years or more that retrospective changes are legal? And that I warned Martin back in 2012 that he was underplaying the likelihood of a government exercising that right to the detriment of borrowers?”

        Absolutely – I’ve followed developments closely.

      • Dan permalink

        Copy of the terms that are on the loan request form is in this post. The rest of the terms are in the regulations.

      • And you think the test of that clause is what’s happening to other loans?

        I had only really thought about that clause in relation to things like RPI-indexing. IE, a sale requires everything to be nailed down as much as possible – indexing in perpetuity arguably benefits borrowers, therefore it’s an allowable amendment.

        I haven’t seen any purchase contracts – though I’ve been told about what’s in the mortgage-style loan sale contracts. I believe no one wants to have to argue these points out or have them tested in law and so any government changes attract penalties in favour of purchasers. Or even, treating such changes as breaking the contract.

      • Dan permalink

        The explanatory notes to the Sale of Student Loans Act explains:

        “36. This section also enables the Government to apply any future change in loan regulations to sold loans as well as those which are still owned by the Government. This includes loans which are taken out before such changes take effect, where the change would be retrospective. This is to ensure that all borrowers can be treated equally, for example if the repayment threshold (the level of annual income at which point borrowers are required to start making repayments) were to be altered by the Government.

        37. In making or amending such loan regulations, or regulations under section 186 of the Education Act 2002, section 5(6) requires the Secretary of State to give consideration to borrowers whose loans have been sold so as to avoid detriment to any such borrower resulting solely from the fact that the loan is sold. The Secretary of State must compare borrowers with their notional selves and aim to ensure that their position is not worsened as a result of any proposed amendment to regulations, as compared to someone with the same characteristics whose loan has not been sold.”

        Really, the Sale of Student Loans Act provides no protections for borrowers BEFORE their loan is sold, as the clauses about ensuring there is no detriment to the borrower solely because their loan has been sold only applies to loans already sold. There is nothing to stop the government worsening terms before they then sell them which surely should have been included in this Act.

        And after the loan is sold, it still allows for negative changes as long as those changes are equally applied to someone with the same characteristics whose loan has not been sold. I suppose you could argue that pre-2012 and post-2012 loans have different characteristics, so until all of the same type of loan has been sold, changes would need to apply equally (or not detrimental to the borrower with a sold loan) for everyone with that type of loan.

        The explanatory notes explicitly give the example of altering the repayment threshold for sold loans so that all borrowers are treated equally, whether their loan has been sold or not.

      • I’d be very interested to see this all tested in law, but I doubt purchasers or the government would. What is the status of legal explanatory notes? And is there a technical meaning to notional selves?

        My reading of those notes is that borrowers are compared to a hypothetical individual with the same characteristics but unsold loans – regardless of whether there are such individuals.

        RPI-indexing was explicitly justified in the regs with reference to a sale. That’s what makes it significant here. The government gave up some future flexibility – periodic administrative reviews – to facilitate a sale.

        I don’t think we should lose sigt of the fact that government will sign clauses which treat changes – detrimental or otherwise – as break clauses or penalty clauses. This won’t remove any rights future governments have in this regard but will dissuade anyone from doing so. If the sale is a success from the government’s point of view, then they have lost all incentive to tweak repayments. And vice versa, a bad sale would be one where they’d have to consider invoking these rights.

        The post-2012 loans were set up to be sold – that was the original intention of the 2011 White Paper – it’s a sign of failure that we’re going back to the Plan D of selling pre-2012 loans.

  17. Brian permalink

    I’d also point out there are all sorts of crossovers in terms between pre-2012 and post-2012 loans anyway for borrowers who have both. Under current ‘repayment slicing’ terms, there is the bizarre situation where making post-2012 loan terms more favourable by increasing the threshold leads to more repayments going to the sold pre-2012 loan. Conversely, lowering the post-2012 threshold would reduce the amount of repayments going to the pre-2012 loan.

    A borrower with a sold pre-2012 loan could take out a post-2012 loan for further study and immediately wipe out most of the repayments going to their pre-2012 loan for the loan purchaser, which suddenly all go to the government for the unsold loan!

    • I would expect current slicing arrangements to be reviewed as part of any sale – a purchaser would probably demand that all repayments went against pre-2012 loans until those balances were cleared.

      • Brian permalink

        If the Government retrospectively changed the slicing terms to apply all repayments first to pre-2012 loans to aid a loan sale, they’d be on very shaky legal ground indeed given that it would breach the Sale of Student Loans Act 2008 in leaving the borrower in a worse position as a result of having their loan sold than had it been sold, and the Government ruled out precisely that option when they considered the slicing terms last time round precisely because it’s both presentationally unfair and onerous for borrowers in not allowing them to repay the part of their borrowing that is accruing interest at the higher rate (and which, for post-2006 borrowers, would have the later write-off so it’s in the borrower’s interest to repay the post-2012 loan first):

        Click to access Annex%202%20Submission%20Dual%20loan%20repayments.pdf

      • The big problem the government has with the sale is that is has to get someone to buy the “junior” or equity tranch of the securitisation. That’s the tranche that gets paid last from the income stream or takes a hit if there is nothing left after the more senior tranches have been paid.

        If the government is left holding that risk, then the ONS may decide that a sale has not been achieved.

        There are a number of problems with a sale but why would I buy a junior share of pre-2012 loans if borrowers can take out post-2012 loans and thereby topslice repayments. I think you said there were estimated to be 200,000 already in this situation of holding both pre & post 2012 loans.

        I think it would be reviewed. You could be right that it wouldn’t be implemented – but Dan’s interpretation above would suggest ways round the relevant clauses of the 2008 Act.

Trackbacks & Pingbacks

  1. The cost of writing off tuition fee loans | Critical Education
  2. IFS on tuition fees | Critical Education
  3. Target impairments – a correction (of sorts) | Critical Education

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: