“What’s the purpose of a pension scheme?” – Con Keating

collective scheme



Con Keating argues in this piece that pension schemes aspire to “self-sufficiency”, giving them the status of  insurance companies. His thinking centres on what claims we have on schemes if the covenant to fund these schemes is lost. His conclusions are that there is no certainty of outcome either in DB or DC, he is calling for a Royal Commission to clarify the situation.


In our response to the DWP’s DB Green Paper, we raised the question of the purpose of a scheme and fund and suggested that this was a topic worthy of extensive discussion. We called for a Royal Commission to investigate this issue and its treatment in UK legislation and practice. This note will address some, but by no means all, of the issues.

In simplified fashion, this is a debate as to whether we should consider the scheme and its fund as a stand-alone or independent entity with the duty to provide the pensions promised, or whether the obligation resided with the sponsor employer, with the scheme as little more than administrative convenience which holds security for the promise made and generates income to offset of defease the employer’s underwritten cost.

A related issue concerns the certainty and security of a DB pension promise. Should the promise be immutable, that is to say, require performance after sponsor insolvency or should the promise merely be secured. In this latter case, the member would receive the accrued value of the promise at the date of insolvency. This may or may not be sufficient to buy similar benefits elsewhere, and indeed would usually not be. Indeed, in order for members to receive the accrued value the scheme would need to be fully funded (at best estimate).

This really is a fundamental issue on which the legislation not clear; there is support for both views in that. It has arisen again in the context of recent articles on scheme valuation. Accounting treats the scheme liabilities as an obligation of the sponsor firm. Here, we will draw out some of the consequences of the two approaches.

If we consider the scheme as an independent entity, which has assumed the pension liabilities, its life is not conditioned on the survival of its sponsor; if the scheme is effectively an agency of the sponsor, then its life is conditioned on the status of the employer sponsor. In both cases, the scheme has recourse to the sponsor should it find its fund is insufficient. In other words, it holds an option to call upon the sponsor, recurrently if necessary, for further funding.

In this ongoing viable sponsor situation, the scheme could never be in true deficit since the value of this option would always offset that circumstance, unless, of course, the sponsor is insolvent and unable to pay. But as long as there is an ongoing viable employer, the scheme has no need for capital buffers or indeed “prudence” in its estimates. Good faith would require it to operate on the basis of a balance of probabilities and expectations (best estimates). “Prudent” funding levels, if applied, would simply reduce the value of the scheme’s call option on the employer sponsor.

The sponsor would report the value of the option written to the scheme – this is not a trivial calculation. Of course, the option contract crystallises with sponsor insolvency and its value at that time is the value of the scheme’s claim in insolvency proceedings.

The difference between the two views of the purpose of the scheme and its fund are of little consequence when the sponsor is viable. In our opinion, there is a gross over-emphasis on sponsor viability – sponsor insolvency is actually a rare event. However, problems do arise with sponsor insolvency. This has given rise to arguments that discount rates and much more should be conditioned on the sponsor “covenant”. We would note, here, that the concept of the sponsor covenant does not appear anywhere in the rafts of DB pension legislation.

This covenant conditioning idea is fundamentally misconceived. Indeed, it is not clear whether this was meant to be a higher or lower discount rate. In financial markets, it is customary and correct to demand a higher yield to compensate for the possibility of default. However, that would imply lower values for the pension liabilities, and lower security for members. This would make the obligations less likely to be replaceable elsewhere on sponsor insolvency. It would also not reflect the fact that the higher yield of markets reflects a diversifiable risk and is priced on that basis. If we are to apply a lower discount rate, then for consistency we should also accept lower benefits terms – perverse at best – potentially a subsidy to the employer. The pricing problem is a classic trilemma – the contribution, the implicit investment return and the final pension benefit are integrally linked and of course feed into the overall labour costs.

One variant to this idea is that the scheme should provide for the insolvency likelihood. We will return to this idea later, when considering the scheme as an insurance company, but content ourselves here with observing that this would be provision against a specific undiversified risk, and would for most schemes and sponsors be extremely small. If the likelihood of insolvency is around 1% pa, which may be thought of as being typical of a BBB credit, then the provision should be only 1% of the loss given default, and that is likely to prove grossly insufficient.

If the pension obligation were no different from other secured commercial debt, the amount of the claim in those insolvency proceedings would be based upon the accrued value to the date of insolvency. No account is taken of potential future developments or indeed of future events which will go unfulfilled. If the security is insufficient to satisfy this accrued value, then the deficit or shortfall is the amount of the claim in the proceedings. This claim may or may not be paid in full – usually not. If the security held exceeds the value of the accrued claim, the excess must be returned to the administrator or liquidator.

This would be the treatment of securities held by DC pensioners. It is equitable among the stakeholders of a company. It is also possible for these stakeholders to allow the alteration of their claims in order to rescue or restructure a company.

There have long been issues with another aspect of DB pensions, the treatment of different classes of members. If we look back at the historic record, distributions on sponsor insolvency were originally set by scheme rules, and only later by pension regulation. Preference under this was given to pensioners, with the consequence that some long-serving active members close to retirement could find themselves left with little or nothing – a situation which proved a public relations nightmare for government. The “cock-up” rather than conspiracy view of history would attribute many of today’s issues to a confusion between the treatment of a scheme among other creditors with the treatment of differing classes of scheme members among themselves.

In retrospect, this preference can be seen as the first step towards the rather strange idea that DB pensions should be inviolable, in the sense of being paid in full no matter what the condition of the sponsor. It also shifts the role of the pension scheme unambiguously to being that of stand-alone provider. A company’s promises die with it.

It is interesting that this view was first discussed by actuaries in 1986/1987 when it was simply asserted as the role of the scheme. No supporting evidence or argument was provided, and indeed the idea that this was the proper purpose of a scheme was contested within the actuarial profession. Here the scheme is offering inviolable immutable benefits; a situation which brings with it the problem of how it should approach this ideal.

The argument for such benefits seems to revolve around the magnitude of the DB pension in an individual’s wealth at retirement. For the long-serving member, the DB pension is usually the largest part of that wealth. However, the changing employment patterns, which many advocates of DC use as support for the argument that DB pensions are now inappropriate, actually mean that the employee would be rather well diversified. If employee tenure really is just 4 years, then the employee would be very well diversified. Indeed, rather better diversified than many employees with DC, given the findings of various studies of employee allocations to securities issued by their employer in these arrangements.

A related argument is that sponsor insolvency represents a compound risk: loss of employment and loss of some of their pension wealth. This is true only for active employee members of the scheme. Retired members have no employment and deferred members are usually employed elsewhere. In other words, this argument would have validity for only a minority of scheme members. We would note also that many rescue, restructuring and recovery plans are precisely concerned with maintaining employment.

The open question is what should the treatment of scheme members be in insolvency, and more broadly. This is genuinely a potential source for moral hazard, unlike many of the other situations described as moral hazard by the regulator and others. A scheme member may wish to have both a low initial contribution and a high and generous set of benefits. The low initial contribution opens the possibility of higher current wages – also an attractive proposition for members. The high set of benefits then implies a high rate of accrual or effective promised investment return. Using the accrual rate that we have proposed elsewhere for valuation would throw light on this situation. Using an exogenous discount rate, such as bond yields or the expected return on assets obscures this and may reward scheme members in an unwarranted manner. This is actually the position with cash equivalent transfer values (see Portfolio Institutional: “Transfer Values, Discount Rates and Reality”)

If all members are assured their full benefits, no matter what, they and their trustee representatives have an incentive to accept overly aggressive and generous terms. Note that this does not apply to benefits already awarded – those terms have already been fixed.

If we regard the scheme as a stand-alone entity, then it is effectively an insurance company – a DB pension is a deferred life annuity, a form of insurance. As such it needs to be capitalised to reflect this situation. As long as the sponsor is viable, this may take the form of the option to call upon the sponsor as needed. But post sponsor insolvency, there is no option to be called upon. The obvious solution, with funded schemes, is for the stand-alone scheme to be capitalised prior to sponsor insolvency.

This is a single risk insurance company, a rather odd and usually risky proposition. Multi-employer schemes have their roots in an attempt to diversify and mitigate these risks. Asset pooling may be expected to deliver economies of both scale and scope, but it is the combination of risk-pooling and risk-sharing which delivers the greater value. It is unfortunate that many of the larger, stronger participants in many of these schemes now view the risk sharing as purely negative, and fail to recognise the benefits which have accrued and should recur in the future.

As an aside, it is notable that there are two models for multi-employer schemes – some offer similar benefits to all members and have a common contribution rate, while some admit disparate terms among participating employers. The creation of multi-employer schemes was very much an early attempt to address the issue of sponsor insolvency intended to enhance the security of the pension promises made to members.

Capitalising a scheme as if it were an insurance company is an expensive business. By virtue of its single risk specificity, it needs to be better capitalised (proportionately) than an independent insurance company writing many such risks. In other words, it will be more expensive for a scheme to be self-sufficient at a given level of risk than for a specialty insurer to provide these benefits at that level of risk.

With employer returns on capital employed averaging a little over 12%, this is very expensive capital. It really is remarkable that schemes are calling for ever more in contributions which they then invest at rates as low as one or two percent when the employer is consistently earning returns on capital in excess of ten percent.

“Prudence” and the desire for “self-sufficiency” are simply forms of excess capitalisation of the promise actually made. Indeed, the Achilles heel of all funded standalone solutions is that they will generate orphan assets. The best estimate or expectation is the neutral level of funding – in fifty percent of outcomes it will provide more than sufficient and in the other fifty percent less than sufficient. If we provide capital to ensure that schemes may continue and discharge all pension liabilities as they fall due over time at say the 99.5% confidence level, then after their discharge all of these schemes will be in surplus – and the average amount will be very close to the value of the best estimate.

It is very interesting to note that schemes do not, in practice, continue to operate beyond sponsor insolvency; they discharge liabilities by transfer and wind up. Suggestions recently made that schemes might continue and operate as stand-alone entities, with no sponsor to support them, were not well received by the authorities. The legislation really does not envisage this – on sponsor insolvency the scheme either buys coverage from an insurance company or if it cannot buy more than Pension Protection Fund benefits, enters the PPF.

The claim of the scheme is based upon the shortfall of assets to the cost of buying out liabilities with an insurer (s75 PA 1995). This inflates the claim relative to other creditors, and leads to the pre-emptive reorganisation of their status and much “pre-pack” activity. It is clearly inequitable to other stakeholders, and might well not survive legal challenge. This s75 valuation is a new business valuation.

The problems of sponsor insolvency are not well-addressed by funding. As this is a problem of contingencies, it is well suited for an insurance solution; low probability high consequence events are routinely insured against by both individuals and companies. In the case, the specific risk, sponsor insolvency, is well-understood; it is the heart of corporate bond markets and bank lending. This is well understood in some overseas markets, where insurance even extends to coverage of unfunded book-reserve schemes. It is also far cheaper to apply and monitor than funding. It is the risk underwritten by the PPF – though in that case inadequately.

There is much confusion and ambiguity in the legislation and trustee practice. Legislation and practice have added greatly to the costs of provision of DB pensions; many of these increases have not benefitted scheme members. Some have advantaged members at the unwarranted expense of the sponsor employer. Against this background, it is hardly surprising that so many schemes have been closed by their sponsor employers.

If we are ever to see a revival of DB provision, and there are good reasons why we should want that, then these costs need to be kept below the level of the tax concessions that accrue to the sponsor. If they aren’t, and it appears that the costs of regulation are currently several multiples of the value of the concession, cash wages or DC contributions are and will remain the more attractive option for employers. Most aspects of the increasing costs due to regulation have attracted comment, but the lowering of the value of the tax concession as corporate and personal tax rates have fallen has passed largely without comment in the world of pensions. The two are a very toxic mix.

About henry tapper

Founder of the Pension PlayPen,, partner of Stella, father of Olly . I am the Pension Plowman
This entry was posted in actuaries, pensions and tagged , , , , , . Bookmark the permalink.

Leave a Reply