Not the author
The question of the security of DB scheme members’ pensions is among the most vexatious subjects in the realm of pensions. Though volumes have been written, with many contributions clearly heart-felt, belief in some erroneous and fallacious analyses is widespread. This article is an attempt to clarify these issues. It will use the liability valuation techniques described in earlier articles.
The investment accrual rate is a property of the pension contract, defined by the contributions made and the benefits projected. It is objective and unambiguous. Both ‘true and fair’, and ‘going-concern’ standards are satisfied. The accrual rate is both the sponsor employer’s gross cost of production of the pension and the implicit investment return on contributions to the scheme member. A liability valuation derived, using the accrual rate defines the amount of a creditor’s claim in liquidation, no matter whether solvent or insolvent. This is the amount which is equitable among all stakeholders of the company. These are valuations of the contracted obligation for the employer and the pension expected by members; it is an implicit term accepted by the members.
This is the contracted value underwritten by the sponsor. The DB pension contract design is a masterpiece of risk-sharing and risk-pooling; a hybrid insurance and investment contract. It is a form deferred annuity, collectively constructed and delivered. These aspects, taken together, serve to lower the cost of provision of the pension to the benefit of both parties. However, as with all contracts, it is incomplete, and it is these gaps which have proved contentious.
If we interpret the motivations of the market-consistency congregation charitably, theirs is an effort to bridge some contractual incompleteness. However, the pension contract is dyadic; a committed, ongoing relationship between two parties. It is non-negotiable, not traded. In personal bankruptcy, it is privileged, inalienable property.
There is a refusal here to recognise the limitations of markets. Indeed, the very existence of a company arises because markets cannot offer certain contracts cost-effectively or in some cases, at all. The circularity and inadvisability of then trying to force the elements of a DB pension contract into a financial market setting should be obvious to all but the wilfully blind.
The market in deferred annuities, such as it is, differs greatly from the conceptual ‘free market’ of economic theory. Indeed, it only exists because a specialist institution, an insurance company has been created in order to offer the contract and profit from doing so. Circularity is again evident. Add to this the differing motivations for provision between an employer and an insurance company, together with its high regulatory cost, and the idea that this is ‘fair value’ is nonsensical. Peeling the pension element away, carving it out from the labour contract changes the nature not just of the residual labour contract but also the pension.
Some numerical illustrations may help. The contracted liability valuation of the illustrative scheme (I) described in earlier articles is £153.37 million (6.07%). For comparison, two hypothetical schemes are considered: one as in a previous article (A) where the contributions were just 50% of those made in the illustrative scheme, and another (C), where the contributions made were 50% higher. The accrual rates for these schemes are respectively 8.23% (A) and 4.79% (C), and contracted liabilities are £123.46 million and £177.59 million. The projected pension payments are the same in all three cases.
While, in general, scheme members should be concerned first by the adequacy or generosity of the terminal pensions promised, preferring more to less, they should also have a secondary, contingent concern with these values. This interest arises with both transfers among schemes and on liquidation. In the case of transfers, it might be argued that this is the value to which the member is entitled, and that if scheme assets are inadequate to deliver this, the sponsor underwriting might be called upon to make good that deficit. By contrast, in the case of insolvent liquidation, with an inadequate fund this defines the amount of claim that is equitable to other creditors. This valuation defines the contractual level for ‘full’ funding. If any other valuation standard is applied, it will distort and misrepresent the contract underwritten by the sponsor employer.
The most important thing to grasp here is that this liability valuation is the gross cost of production of the pension to the employer; it is the (gross) cost of production of the promised contracted pension by the employer. It is not the cost of replacement of the pension by some other means; it is not a cost of procurement in some imagined quasi-market.
This understanding raises a very important issue, that of prudence in the assumed values of factors entering the estimation of the ultimate pension payments. It would be sound risk management practice for these parameters to be conservative when the contract being valued is for supply of the pension by the corporate sponsor, but it is simply foolhardy when the contract is for the procurement by some other means. Sound business practice in the procurement of pension supply would, as with any other commercial service supply, require the minimisation of the cost of these services, not exaggeration of their estimated cost. This is also the motivation underlying arguments which find expression as ‘going concern’ valuation.
The arguments about the sustainability of the sponsor employer arrangement, the longevity of the institution are an expression of this issue. The cost of procurement by some other means is relevant only to the extent that the employer is unlikely to survive to pay the pensions contracted. While most see this as the likelihood of insolvency, it should be recognised that companies are twice as likely to merge, be acquired or cease trading when solvent as they are to enter bankruptcy or re-arrange their affairs in acute distress. The alternate procurement valuation is a conditional valuation; this should mean that if allowance for this possibility is desired, it will be a minor adjustment.
A numerical example is appropriate. The alternate valuation used here will be the buy-out cost of the illustrative scheme (I) which was quoted at gilts less twenty-five basis points (1.40%); this has a cash value of £286.28 million. Let us assume that the likelihood of failure of the sponsor of this scheme is 1% in the current accounting period. Then the valuation of this scheme, including provision for such an event, would be:
£286.28 *0.01 + £153.37 *0.99 = £154.70 million.
The DB pension is a very significant asset for most scheme members; its security is a valid concern from the standpoint of their social welfare. The primary concern revolves around sponsor insolvency. With this in mind, we might ordinarily expect regulation of the quality of companies permitted to offer this form of pension or regulation of the terms of pensions awarded.
However, there would be some real issues of feasibility involved with this form of regulation. Suppose regulation limited authorisation for the provision of these pensions to high quality companies; there would be arguments over the assessment of quality and the arrangement would also be exposed to the migration of the sponsor, over time, from high to low quality. With technological disruption now a very much in vogue topic, it should be obvious that some of these migrations or transitions might be very rapid indeed.
Regulation of the terms of award might be considered as helping to resolve the problem, evident earlier, of lack of comparability of scheme terms and transfer values. In the earlier example, a member of the high accrual rate scheme (A) would be unable to purchase, with their proceeds, comparable benefits in the illustrative scheme (I) or scheme (C). It should be noted also that the accrual rates quoted in this and previous articles have in the main been average accrual rates, when the accrual rate that is relevant would be the rate for new awards or the acceptance of incoming liabilities. Incidentally, the accrual rate for new awards for the illustrative scheme is currently 5.50%; it also accepts transfers in at this rate. The problem with such a form of regulation lies with the fact that accrual rates do vary over time. One particularly thorny issue would be that revisions to the accrual rate may arise from both experience and changes to assumptions and parameters underlying the liability projections. In an earlier article, the average accrual rate over the period since 1952 was shown. The highest average rate reported was 13.24% yet there were only three years in which new awards were made at implicit rates higher than this; most of the increase in this average was driven by experience and the upward revision of assumptions as they applied to the stock of prior awards. Regulation of these forms would require perfect foresight to be effective. Clearly, regulation of the terms of award is neither desirable nor feasible in this light.
An earlier article discussed the purpose of a pension fund, and this concluded that the primary purpose of a fund was to offset or defease the gross cost of DB pension provision to the employer. While it might be argued that the status quo arises from the tax concessions afforded to schemes and particularly the period of high historic rates of corporate taxation, which is then combined with inertia. However, the past few decades have seen radical changes to the ways in which scheme are regulated and managed, which make the assertion of inertia suspect. As achieved investment returns have fully offset or exceeded the cost of production of DB pensions over most of post-war history, this offset or defease explanation seems more plausible.
For the purposes of this article, which is concerned with member security, this motivation is immaterial. As was noted in that earlier article, funding a scheme is highly inefficient, very costly relative to the use of pension indemnity assurance and incomplete. Assurance can ensure total security of the member benefits, under an arrangement where the assurer steps in on sponsor insolvency, and then pays the full pensions as and when due. The premiums charged for such cover reflect the accrual rate cost of the scheme; they are based upon the assurer’s cost of production of the covered pensions in the event of sponsor failure.
The interposition of a pensions trust is a device to enhance member security; it ensures that scheme assets are bankruptcy-remote in the event of sponsor insolvency. With that said, assets held in excess of the contracted liability valuation are contestable by other stakeholders.
If we consider the purpose of the fund to be ensuring payment of the pensions post insolvency, a stand-alone arrangement where the fund runs off the liabilities of the scheme, then there is an issue of the level of funding necessary. Take the illustrative scheme (I), and suppose that this is funded as the level of contracted liabilities (£153.37 million), and further that it has market-based investment opportunities expected to return its accrual rate (6.07% p.a.) available to it. It faces uncertainty about the variation of this rate from time to time. The consequence of this uncertainty is that the scheme has a fifty percent chance, due to adverse temporary market developments, of failing prior to the complete discharge of all pension payments. In order to provide for this uncertainty, it needs additional funding over and above the expected liability valuation. The amount of this excess funding is determined by the variability of the investment portfolio expected to deliver the required 6.07%. If this portfolio has a volatility of 20%, the required level of funding would be £191.40 million, 128.4% of contracted liabilities. The risky fund (A) would also face the same funding requirement if it could also achieve the same uncertain 6.07% investment return. The difference then, of £67.94 million, would reflect the increased dependence of the scheme on the sponsor underwriting, the increased riskiness of that scheme’s contractual arrangements. Under these conditions, it requires funding at 155.0% of contractual liability value to assure stand-alone delivery. The most secure scheme (C) would require only £13.81 million of additional funding, just 7.8% of the contracted valuation.
As reported in an earlier article, the illustrative scheme has assets of £207.44 million, which implies under these assumptions that the scheme is more than adequately funded to stand alone. With the further assumption that the assets of the hypothetical comparator schemes would hold assets in similar proportion to their relative contribution records, then scheme (A) would have assets of only £103.72 million and a shortfall of £87.66 million. By contrast, scheme (C) would have assets of £311.16 million, massive overfunding (162.6%) of the stand-alone funding needed and indeed even sufficient to achieve full buy-out on the terms recently quoted (1.40%, £286.28 million).
The difference between the contracted valuation and stand-alone requirement is the value of the sponsor underwriting of the illustrative scheme liabilities, £38.03 million. Here, there is a possible argument for over-funding: one possible defence to the potential claims of other creditors for the return of funding in excess of the contracted liability. This stand-alone value of the fund can also be used to inform and evaluate decisions about scheme closure and wind-up through buy-out or otherwise
One thing is clear among all of the strands of this analysis: the worth of a DB pension scheme is and should be higher to its members than its cost of production to the sponsor employer. It may be that a funded DB system is less efficient than provision by some other methods, such as indemnity assurance, but it is also clear that the sponsor-underwritten, funded arrangement is far more efficient than any quasi-market ‘solution’. By providing these, the employer company is fulfilling one of its prime economic functions: adding value for the benefit of its stakeholders, in this case its employees. It is important to remember that this should be done in a manner which is equitable to those other stakeholders. The accounting and regulation should reflect this.
In recent times, it has become fashionable to discuss the possibility of altering the terms of benefits awarded. This arises from the belief that DB pensions are unaffordably expensive to provide, but there is no analysis beyond the current market-consistent method to support this belief. If we are to consider modifying the terms of DB pensions contracts, it is necessary first to value them in terms of their contracted values. This caution applies with equal weight when considering demands for additional ‘deficit repair contributions’. It seems likely that a pleasant surprise awaits many of the trustees and sponsors of many schemes.
