All entries for Saturday 07 April 2018
April 07, 2018
An article about the USS pensions dispute in the Times Higher Education by David Voas (Let's defend pensions not defined benefits) gives a somewhat inaccurate account of the issue. It also makes a comparison between Defined Benefit and Defined Contribution pensions that is very misleading.
At the centre of the issue between the union and the employers is what kind of scheme it is. The employers want to end the guaranteed pension that is based on years of service and earnings, and replace it with a defined contribution arrangement where instead of a pension on retirement one gets a pot of money that depends only on what has been paid in and investment returns.
The article argues compares a young person aged 30 near the start of their career with someone nearing retirement, and concludes that the youngster would do a lot better under Defined Contribution. This is very misleading. For most people (almost all in fact) they would get a lot better penson in retirement from a Defined Benefit scheme.
The article actually makes a number of statements that are wide of the mark. And it is not sufficiently analytical.
For example it says, "defined benefit schemes are defunct unless underwritten by the taxpayer." That is not really true. There are schemes in the private sector that are not defunct. And there is no reason why the USS should be defunct without taxpayer support. The problem is with the way the regulations are being applied and the weakness of the employers. While it is true that most of the DB pensions schemes are in the public sector, such as the Teachers Pension Scheme that is normal in the post-92 universities, what Voas is getting at, I think, is a repetition of City group think that applies simplistic generalities.
His description of the problem, though, goes from vagueness to missing the whole point. He writes,"The problem is twofold. Employers struggle to afford the cost of guaranteeing that the benefits promised in 50 years' time will be paid ... Employees are also hit, though, because everyone ends up paying above the odds." (emphasis added) What does this mean? It sounds very bad. It is not factually true. There is no conclusive evidence that employers struggle to pay the pensions in 50 years. All that is required is that the scheme stays open until then and invests in high return assets such as company shares (equities).
The next paragraph is an explaination but is in fact a gross oversimplification to the point of being misleading, revealing that Voas has not understood the arguments: "The basic equation is simple enough: contributions plus investment income equals benefits. Contributions - and even benefits - are fairly predictable, albeit affected by changes in life expectancy. The challenge is to estimate real returns on investment over several decades."
That is not where the problem lies. There would actually be no problem if it were as Voas says. The USS investment portfolio has a high return, an average of 12 percent per year over the past five years, for example, and is expected to continue to do well into the future, since it is skewed towards equities.
The problem is that the liabilities (present value of future benefits) are being valued counterfactually. That is, NOT using the real returns on investments as described. If things were as Voas says, there would not likely be a problem. The problem is that the scheme is being told to assume very poor returns - index linked gilts have a negative return. So the problem is a result of the excessive prudence being forced on the scheme by the regulations and the weakness of the support from the employers.
In order to compare DB with DC, he contrasts two academics with the same salary, a woman aged 30 with an older man nearing retirement. They both receive a defined benefit of 1/75 of salary for a year's work. He claims that for the older man it is a windfall: "he receives substantilly more than his current contribution would support. For the 30 year old it is very poor value. She could conservatively expect her contributions to grow by 2 per cent per annum, thereby doubling her money by age 65. This fund would be worth more than the defined benefits." This is very difficult to make sense of. Why compare these two individuals in this way, by just taking a snapshot of one year. What pension would the young woman get when she retires? How would that compare with the man? Surely what is needed is to compare the pensions obtainable from DB and DC over a whole career, not just one year taken at different stages in life. It is also very superficial. We are not told if the returns are after inflation. All defined benefits are uprated for inflation by the CPI.
And so it goes on. It reads as if the article were simply intended to make the case for DC over DB. But that goes against all the evidence from academic studies and practical knowledge from actuaries. It also goes against the studies that have been done for the UCU by its actuarial advisers, which show that for the great majority of members, it is very considerably more expensive to provide a pension of a given annual amount by DC. Estimates have put this at between 40 percent and 100 percent more.
This piece is extremely poor scholarship.
These comments on the USS proposals have been written by John Murray, who is neither an academic nor a member of USS, but has an expert understanding of the issues, having over fifty years of experience in the insurance industry.
A sustained large-scale academic assault on decades of mistaken focus on point-in-time market values and inappropriate discount rates may yield results that strike action cannot.’
(Extract from a letter published in the Financial Times on the 13th March 2018 on the subject of the USS dispute.)
His comments below relate to two recent documents provided by USS Managers, entitled ‘2017 Actuarial Valuation’ dated 1st September 2017 and ‘31 March 2017 Actuarial Valuation’ dated 8th December 2017
On the deficit
While the derivation of the discount rate applied is far from clear it does appear as though the apparent shortfall results from the managers’ deliberate intention to shift the fund’s investment portfolio to low-yielding government securities. Without this change there would be no deficit.
On ‘De-Risking’ and Collateral Damage
‘De-risking’ appears to consist of changing the investment profile as described above and then back-filling the resultant shortfall by use of the employers’ current deficit contributions of 2.1% of pensionable salaries. The consequent closure of what remains of the ongoing DB scheme seems to be regarded as collateral damage.
On Risk Generally
- Referring to long term investment in equities as ‘betting’ or ‘gambling’ is inflammatory, ill-advised and has no place in a discussion of this nature
- Investing in government securities is by no means risk free
- Nowhere is the risk inherent in the opportunity cost of missing out on potential equity returns addressed
- The risk that members may be deprived unnecessarily of a valuable benefit by reason of the adoption of a low-yield investment policy is never even mentioned
- The only risk that the managers seem to be interested in avoiding is the risk of a problem with the Regulator.
‘It would be difficult to convince the man on the Clapham Omnibus that one achieves self-sufficiency by reducing the income available to the fund’
On Achieving the Stated Objectives
On no common understanding of the terms will the managers’ proposed action achieve the stated objectives of de-risking and securing self-sufficiency.
On the Way Forward
- The direction being proposed by the managers needs to be abandoned
- USS is uniquely placed to seek a derogation from the much-criticised pension rules
- A joint case to make an exception should be put forward by UUK, USS and UCU
Observations on USS Proposals as set out in two documents entitled ‘2017 Actuarial Valuation’ dated 1st September 2017 and ‘31 March 2017 Actuarial Valuation’ dated 8th December 2017
In the notes that follow, the above documents are identified as 9/17 and 12/17 respectively. The latter document is to some extent an update of the former but is much less extensive.
1. Summary and Overview
1.1. Although neither of these papers is presented as a set of proposals for action, but rather as a review of certain suggested valuation assumptions, they very clearly contain implicit proposals that involve a radical restructuring of the USS (the Scheme)’s investment portfolio.
1.2. Boiled down to their essentials, the proposals are aimed (perhaps not directly but certainly effectively) at reducing the investment income that the Scheme will receive in the future, thereby creating a funding deficit. One of the consequences of this action will be the need to remove what remains of the final salary scheme going forward.
1.3. The reasons for this action are given as either ‘de-risking’ or creating ‘self sufficiency’. These notes will argue that the proposed action will achieve neither of these objectives while unnecessarily depriving members of what is left of a valuable benefit.
1.4. These notes also include suggested possible ways out of the current predicament.
2.1. The Scheme’s financial position is set out below:
USS Funding Position – Gap Analysis
Source: 31 March 2017 Actuarial Valuation dated 8 December 2017
Best Estimate Approved Basis Gap
Accrued Liability 54.8 67.5 12.7
Assets 60.0 60.0
Surplus/Deficit 5.2 -7.5 -12.7
Self-sufficiency 82.4 82.4 0
Mean discount rate
(above gilt yield) 2.31% 1.20% -1.11 (-48%)
2.2 The expression ‘Best Estimate’ in the first column should not be confused with the business term ‘Best Case’. In the Definitions at p 57 of 9/12 ‘Best Estimate’ is defined as ‘The trustees’ unbiased view of the future outcome for different variables without adjustment [or] with margins of any kind. It is consistent with the median (or 50thpercentile) outcome’. This might fairly be translated as business ‘Mid Case’. Moreover, the definition of ‘Discount Rate’ in the same glossary says ‘The discount rate for technical provisions is determined by the expected investment return less a margin for prudence’, so it may be expected that this mid case outcome does in fact contain some element of caution.
2.3 The term ‘Approved Basis’ might be better described as ‘Managers’ Selected Basis’. The deficit that appears in this column is the result of the lower discount rate employed. But this column should not be confused with ‘Worst Case’. It is not an attempt to see what will happen if the margin above the gilt yield happens to drop by 48%. What it represents is the expected deficit that will result from changing the approach in line with the comments made in the body of the documents. In other words the deficit does not come about as the result of some possible economic downturn but as the direct result of action that the trustees are, by implication, choosing to take. The actual discount rate is not made explicit but is expressed as a margin over the gilt yield. There are several sets of numbers given in the documents but again the mean rate is never made clear. The nearest clue is at p 10 (12/17) where, under the general heading of ‘Investment Return (discount rate)’ it says that ‘This approach therefore includes a provision for a gradual investment de-risking [i.e. moving to gilts or equivalent] to take place over years 1 to 20’. As far as it is possible to tell, this seems to explain the movement in the mean discount rate between the ‘Best Estimate’ and ‘Approved Basis’. The deficit would not, therefore, be expected to arise if things were to remain as they are.
2.4 The expected rates of return on different types of investment presently in use are set out below. The effect of the proposals, basically to shift from equities to gilts, can be clearly seen.
Projected Real Rates of Return (i.e. adjusted for inflation)
Source: 31 March 2017 Actuarial Valuation dated 8 December 2017
Asset Class 30yr Expected
Listed Credit 1.45%
Index Linked [Gilts] -0.76%
3. Why the Change?
3.1. The reasons given are described as ‘de-risking’ or seeking to achieve ‘self sufficiency’. The two concepts, as used in the reference documents, are closely interrelated. Self-sufficiency is described at p 58 (9/17) as ‘The value of assets that are required to meet the Scheme’s accrued defined benefit liabilities while adopting a low risk strategy. By a low risk investment strategy we mean one for which there is a low probability of ever requiring additional employer contributions to fund benefits accrued to date.’ Note that this does not include benefits that might be accrued in the future.
3.2. One might well ask why it is thought that self-sufficiency is best achieved by moving to a class of assets that has traditionally produced such a poor return. And also if it is correct to describe gilts as ‘low risk’. These matters are addressed at 4 and 6.
3.3. It is safe to say, however, that this notion (shifting to gilts or similar) runs throughout both documents as though it was a given. Sometimes it is denied, for example at p 43 (9/17) it says ‘The trustees apply a ‘discount rate’ to the liabilities which is based on assumed returns from current and planned future asset allocations. This does not adopt a formulaic approach to setting the discount rate linked to gilt yields’. But at pp 49 & 50 of the same document it says ‘For the ‘self-sufficiency’ and ‘economic’ bases the discount rate assumes a term structure derived from the yield of fixed interest gilts appropriate to the date of each future cash flow (extrapolated for cash flows beyond the longest available gilts) as advised by the Scheme Actuary. For the ‘self-sufficiency’ basis a margin of 0.75% is added’. So practice would seem to differ from theory.
3.4. The real reason for the proposed change is never articulated but it is not too difficult to divine. The objective of the Scheme managers is to switch the investments to what at p 28 (9/17) are referred to as ‘bond like’ investments aimed at producing the gilt yield plus 0.75%, and then to fund the resulting deficit by use of the (employers’) ‘current deficit contribution of 2.1% of pensionable salaries’. This deficit, it is estimated, will be eliminated in eight years (pp 24&25 9/17). After that they hope that the income from the ‘bond like’ investments will be sufficient to ensure that the employers will not need to shoulder any further increase in their contribution.
3.5. It is, of course, inevitable that defined benefits arrangements are discontinued at this stage otherwise, with such a poorly performing investment portfolio, the contributions required to keep these benefits going would be enormous.
3.6. There is a driver behind the driver and that is current pensions legislation which pushes schemes down the gilts route. The essentially flawed thinking behind this legislation is considered at 4 and 7. It is in this area that the real battle lies (see 7.2) and it would be valuable if the trustees could be brought on board. The managers’ intention is to reach some notional ‘safe haven’ whereby if things do go wrong they have assured the trustees that they will be free from any liability because they have followed the ‘safe’ route. This should secure a quiet life for all (except the Scheme members).
4. Investment, Gambling and Risk
4.1. Before moving on to consider a way out of this morass, it would be opportune to take some arguments off the table before they are employed as a counter. It is sometimes contended, for example, that investing in equities constitutes gambling and that using this form of investment is equivalent to ‘betting’ the scheme’s funds on the vagaries of the stock market. As we will see below, there is risk inherent in all types of investment but it is important to differentiate buying equities to hold for the long term and buying and selling in short order with a view to a quick profit: the former is investing and the latter may be considered a form of gambling (like the activities of a day trader). Moreover it is worth bearing in mind that equities, as a class of investment, have outperformed all other forms of investment, including UK house prices, over the long term (and what is a pension fund if not a long-term investment?) or at least in the period from 1900 to end 2017. See Global Investment Returns Yearbook: Credit Suisse (publ.). Talk of gambling or betting is inflammatory, ill-informed and inappropriate in this type of discussion.
4.2. Government bonds, on the other hand, are by no means risk-free investments as is sometimes claimed. Some of the risks involved are implicitly acknowledged in the reference documents but it would be as well to make them all explicit here.
4.2.1.Matching Term Risk
This is acknowledged by implication on p 10 (9/17) where it refers to ‘...cash flow (extrapolated beyond the longest available gilts)...’ Pension funds are very long–term undertakings whereas government bonds have end dates and it is often impossible to match the pension terms with government securities. This risk does not exist with equities.
4.2.2 Assumed Future Interest Rates Risk
There is an assumption by the managers that [bond like] rates will revert to ‘normal’ in ten year’s time (see the jump in the discount rate at year 11 in the table on p 51(9/17) and also p 5 (12/17). At p 8 of document 9/17 it is observed that ‘If interest rates do not in fact revert as forecast to the levels proposed (sic) by the trustees, then future contribution requirements could increase…...’ Well, yes and maybe a stronger word than ‘could’ would be appropriate. This risk is a heavy one and could be avoided. They got it wrong in 2014 when they thought that rates would have reverted by 2017. See p 9 (9/17) ‘Between valuations, long-dated index-linked gilt yields have fallen from already historically low levels by a further 1.5%, making them more expensive than in 2014. As a result the trustees could not de-risk the portfolio under the funding triggers agreed at the 2014 valuation’.
4.1.3 Reinvestment Risk
Because government securities are issued for a fixed period, they are automatically redeemed when the period ends. It is by no means certain that a similar rate of return will be available on issues then coming onto the market. See the remarks about extrapolation of cash flows under Matching Term Risk
4.1.4 Pro-Cyclicality/Availability Risk
If everyone is seeking to invest in bonds then this will drive the prices up and the returns down. This is implicitly acknowledged at p 9 (9/17) also see under ‘Assumed Future Interest Rate Risk’ (above). No mention is made of the fact that, if the government is looking to reduce the level of its borrowing, there is the risk that there may not always be sufficient bonds issued to meet demand. This is already a factor in Germany.
4.1.5 Sovereign Default/Concentration Risk
While the British Government has never so far defaulted on a bond, sovereign default is not unheard of and the future is difficult to predict. Going almost exclusively for government bonds introduces a risk concentration element which does not exist with a well diversified portfolio of equities
5 The Risk that Durst Not Speak its Name
There is a very significant risk that is not even alluded to in either of the reference documents, namely that the members may be deprived of what remains of a most valuable benefit for no good reason. Future history could well show that by maintaining a portfolio of mainly diversified equities sufficient income would have been produced to keep current benefits in place. This is the clear implication of the ‘Best Estimate’ valuation. Trustees owe a duty to every participant and not just the Regulator (as one could be forgiven for imagining reading the reference documents).
6 De- risking and Self-Sufficiency
6.1 So far as concerns the question of removing risk it is difficult to see how the proposals will achieve this, unless one regards equity investments as inherently so very risky that they must be avoided at all costs. But as has been demonstrated above, there are many risks associated with government bonds and it could be argued that the real risk that the planned action will run is the opportunity cost of not investing in equities and so missing out on the returns that have traditionally been available from this source. The only risk that the managers seem to be intent on avoiding is the risk of possible problems with the Regulator.
6.2 It would be difficult to convince the man on the Clapham omnibus that one achieves self-sufficiency by reducing the income available to the fund. It is only the very strange definition of self sufficiency contained in the document (first create a deficit and then seek to back-fill it) that will be secured by the proposed approach.
6.3 There is a powerful argument to say that what is being proposed will achieve neither the reduction of risk nor any degree of self-sufficiency.
7 A Way Forward
7.1 As has been demonstrated above, the proposed approach fails in its objectives and, by way of collateral damage, ends defined benefits going forward. It would be wise, therefore, to abandon ‘de-risking’ and revert to a more balanced portfolio of investments, including a substantial equity element. This will not be so easy to achieve, however. As explained at 3.4 to 3.6 (above) there are regulatory issues to consider. It is most probable that the managers will have intimidated the trustees and convinced them of the absolute necessity of reaching ‘safe haven’ if they are to be sure of avoiding any risk of personal liability. It would not be the first time that such a thing has occurred and the natural caution of the trustees is understandable in the circumstances. Further, the Regulator may be expected to support the managers’ position. The Regulator’s role is to enforce the rules, not to consider the best interests of the scheme members.
7.2 The easiest way to arrest the planned journey to Armageddon would be to seek a
derogation for the Scheme. There are grounds for this. At pp 35 et seq. (9/17) reference is made to the strength of the employers’ covenant. The heading at 1 on p 35 states that ‘The covenant is uniquely robust’ and at 2 that it is ‘........rated as ‘strong [the highest rating] by PwC’. For all the reasons explained in the document this looks to be a fair assessment. One could therefore argue that the Scheme, in its unique circumstances, should be allowed to derogate from the real or implied rules of pensions legislation and to continue with a (truly) balanced portfolio approach. It follows that the trustees should at the same time be granted immunity for any adverse outturn that might follow the granting of this derogation.
7.3 There is already support for this view. In a letter published in the Financial Times on 13thMarch of this year and referring directly to the USS, the joint signatories called for ‘A sustained large-scale academic assault on decades of regulators’ mistaken focus on point-in-time market values and inappropriate discount rates…’ And it should be emphasised that the requested derogation would cost the taxpayers nothing. To have the maximum chance of success the application should be made jointly by UUK, the Trustees and the UCU to the Pensions Minister.
7.4 Another, though less certain, approach would be to see if the trustees could avail themselves of what is known as ‘the business judgement rule’ which is available to members of company boards. This says effectively that directors cannot be held liable to their shareholders for an adverse business outcome if they made the decision in good faith. This can apply even to contrarian decisions such as assuming a rise in oil prices when, in fact, they fall. Broadly the claimant, to have a chance of success, needs to be able to show that the directors acted in their own interests and against the interests of the shareholders or favoured one group of shareholders over another. It is a powerful defence and it might be worth finding out if the trustees could take advantage of it or something similar. If the current advisers still insist on the proposed course of action then it should be possible to find some advisers who will take a different view. After all the shortcomings of the current pension rules and their consequences are well known.
Government regulations for defined benefit pension schemes, as they stand at present, under the Pensions Act 2004, are a major factor contributing to the pensions crisis that has seen the closure of scheme after scheme.
The problem stems from the rule requiring schemes to be judged on a simple comparison of market-based capitalized values of assets and liabilities every time there is a valuation, every three years. This rule is misguided for two reasons: because it is arbitrary in that it does not provide a meaningful comparison; and because it greatly magnifies risk, costing a lot of money to offset.
The rule is arbitrary because of the inconsistent way it requires assets and liabilities to be compared. The assets must be valued at market prices on the valuation date. But the value of liabilities, which is not priced in a market, must be estimated as some kind of present value of the pension promises. Since pensions consist of a stream of payments to retirees for many years into the future, there is no asset that can be traded in a market, and the regulations do not prescribe precisely how calculation should be done. The liabilities value is therefore essentially a notional sum.
The two sides of the balance sheet are calculated using very different methods. I will discuss the problems with the assets side later but first consider the liabilities figure. This is meant to be an amount of capital today that would, if invested, yield just enough to pay the pension benefits.
This thought experiment is a matter left to the trustees, with general guidance from the regulations and actuarial advice. They must decide whether – hypothetically – to invest in ‘safe’ bonds or ‘risky’ equities and other growth assets, or a mixture, subject only to the need for having regard to prudence. A sensible choice – you might think – would be to use the actual asset portfolio they have invested in. That would be consistent. However they often find themselves under great pressure from professional advice not to do that but to assume a ‘risk-free rate’ such as the return on long term gilts. Accountants would probably say it is mandated by FRS102. (By the way, even if they switched their portfolio into government bonds it would not be risk free.)
Thus, assets and liabilities are constructed using different – inconsistent – methodologies, making the difference between them – the ‘deficit’ or ‘surplus’ pretty arbitrary. A scheme that is open to accrual can be in ‘deficit’ even though it may have positive net cash flow and is not having to sell capital: in other words it is not in deficit in the way we usually understand the word. That a scheme is in ‘deficit’ is often – usually – reported as a bald fact without a proper health warning.
A notable example that is attracting a lot of ill-informed comment is the country’s largest funded scheme, the scheme for the older or pre-92 universities, the USS; the employees’ union, the UCU, is making my point but the USS executive show little sign that they are willing to listen. However the employers’ body, the UUK, have said they are willing to look at it in a joint experts’ panel with the union.
When a scheme is in notional deficit – as most currently are because near-zero gilt rates have blown up the liabilities – the regulator requires employers to make recovery payments. For an open scheme this often presents an existential threat if the employer cannot afford the payments.
Clearly, the valuation of liabilities is a serious problem which means that deficits or surpluses are arbitrary. Since they are the difference between two large and imprecise numbers they are also very volatile.
But things get much worse when we consider what the valuation of the assets means.
The regulations require assets to be valued at market prices. But, if the scheme is open, asset prices are the wrong indicator because pensions are paid out of earnings. In theory this should not matter because asset prices should exactly reflect the underlying earnings in dividends, rents, interest and so on. But markets are not perfect – in fact far from it – and this equivalence does not hold in practice.
This point is widely ignored by financial economists, even though it is well known to economists. Studies by the Yale economist Robert Shiller and others, have compared asset prices with their theoretical values based on discounted future earnings, and found them to be massively too volatile in practice. Equity prices, for example, are known to be many times more volatile than the economic fundamentals, that is, expected dividends, would suggest. The same is true of bonds and real estate.
It cannot be stressed enough that excess volatility is large. For example, in his book “Irrational Exuberance”, Shiller reports that “only 27 percent of annual return volatility of the US stock market might be justified in terms of genuine information about future dividends”. This would imply excess volatility of equity market prices by a factor of around four.
This does not matter much for a closed scheme where the assets must be enough to cover the liabilities whatever discount rate is used. The focus is on aligning finite assets and liabilities. Market valuations are inescapable.
But it matters a lot for an open scheme that holds its investments long term for income. Rather than at market prices, assets should be valued as the capitalized value of projected future expected earnings, using the same discount rate as for liabilities. This is much less volatile than asset prices.
But the regulations insist on using asset prices, with the result that volatility – which otherwise would be inessential to an open scheme – must be treated as a source of risk. The regulations therefore pose a major threat to an otherwise healthy and sustainable scheme because the notional ‘deficit’ will need to include an allowance to cover the risk due to excess volatility. Yet this market volatility is irrelevant to the financing of the scheme on an ongoing basis.
The regulations are intended to protect accrued pensions benefits against company sponsor failure. But there is two-way causation and schemes may close because the cost to the employer of plugging ‘deficits’ is too high. This has been a major factor in schemes closing. If deficits are over-stated because the rules treat market volatility as risk, then the rules themselves are unfit for purpose.
This therefore suggests two recommendations for reform. First, value an open scheme on an ongoing basis, rather than using market asset prices, which is an implicit assumption it is about to close. Rather than comparing market-priced assets against capitalized liabilities, trustees would get a better picture by comparing the profile of projected income and outgo. That would avoid both the arbitrary choice of a discount rate and the increased risk due to excess volatility of market prices. This reform would remove the reverse causality of an artificial deficit undermining the covenant. There still remains risk of the loss of employer support.
The second reform would deal with that risk. The Pension Protection Fund should protect open schemes as well as closed ones. A pension scheme does not have to close on employer bankruptcy if the work of the employee members continues with or without another employer sponsor. In extremis the PPF could act as sponsor. As a statutory body the PPF can afford to be a patient investor, investing for the long term to gain the equity premium. This means that the a scheme transferred to the PPF would have a lower deficit or surplus when valued as above.
The PPF will be able to support open schemes as well as closed ones. This would reduce the pressure on open schemes to close. The PPF should become a real safety net enabling companies to take a long view of their open schemes, thereby saving them having to make provision for short run market volatility as risk, reducing the cost of pensions. A consequence of this will be greater confidence in DB schemes many fewer scheme closures.