Funding for What? (1972-1997)

Actuarial thought on pension fund valuation prior to the 1970s was almost exclusively focused on valuation with the objective of determining the longterm contribution rate, expressed as a constant percentage of salary roll, that would be required to fund liability cashflows as they ultimately fell due. Similarly, thinking on investment strategy was centred on the objective of generating long-term asset income that would be used together with these contributions to fund those liability cashflows. Pension funds were managed to provide a stable funding plan for the pension liabilities on a going concern basis. The pension fund’s function as a source of security for employees’ accrued benefits had, of course, been recognised by actuaries since the nineteenth century. But it did not figure prominently in the first 100 years of actuarial thought on how to manage defined benefit pension funds.

Funding with the objective of providing security for current accrued benefits started to command greater prominence in British actuarial thought in the 1970s. This was partly driven by an increased incidence of corporate mergers and takeovers, and the tendency these had to lead to wind-ups of some of the involved companies’ final salary pension funds (in which circumstance, the sponsor was not obliged to make any further contributions to the fund). In occasional high-profile cases the pension fund wind-up exposed the inadequacy of the asset fund to secure the pension benefits that members expected.

D.F. Gilley, whose work we already encountered above in his 1958 paper with Funnell, published an important paper on this subject in 1972.[1] He discussed what occurred on the wind-up or dissolution of a defined benefit pension fund. There were several aspects of the dissolution process that could place considerable responsibility with the scheme actuary. The trust deed of the pension scheme would, in theory, set out what the members’ benefits would be in dissolution. But these rules would generally be quite vague. Members might only be entitled to the benefits they would have obtained if they had withdrawn from the fund on the date of the dissolution of the fund. If any surplus or, more pertinently, deficit existed relative to the cost of securing those benefits from an insurance company with the pension fund’s assets, the actuary’s advice on how it should be equitably distributed amongst the members would usually be decisive.

The cost of securing the accrued dissolution benefits could diverge significantly from the long-term actuarial valuations undertaken to set contribution rates. The cost of the dissolution benefits was driven by the then-prevailing market gilt yield (which determined life companies’ pricing terms for the deferred and immediate annuities that would need to be bought to secure the benefits). As we have seen above, this market-based rate could vary substantially from the valuation interest rate used in the pension valuation (which was based on the actuary’s estimate of the expected long-term return on the pension fund assets and hence could include the actuary’s estimates of the equity risk premium and the long-term expected interest rate). Although market gilt yields would typically be lower than a pension fund’s valuation interest rate, the cost of the dissolution benefits would not necessarily be greater than the long-term ongoing valuation—the withdrawal benefit that would typically be viewed as the minimum accrued benefit would break the link with real salary growth, thereby significantly reducing active members’ liability values (the degree to which this impacted on the valuation would therefore be a function of the maturity of the pension fund). Overall, contribution rates and asset strategies were not set to target dissolution benefit funding, and the funding position on this basis could therefore vary markedly across pension funds.

Gilley called for the dissolution benefit funding level to be calculated as part of the regular actuarial management of defined benefit pension funds. Moreover, he proposed this funding result should be published and communicated to members. This was viewed as a fairly radical notion at the time. Indeed, the entire idea of taking into account the dissolution benefit funding level when setting contribution rates was provocative to traditional actuarial thinking. However, the economic turmoil of mid-1970s Britain prompted actuarial pensions thinking to consider some even more radical questions: should private sector defined benefit pension liabilities be funded at all?

This far-reaching question was raised in no less august and authoritative a setting than a Faculty of Actuaries presidential address—that of R.E. Macdonald in 1977.[2] The economic and political environment for defined benefit pension funds was extremely challenging in the early to mid-1970s. Annual price inflation reached a peak in excess of 24 % in 1974/1975. This high-inflation environment naturally fed through to salaries and drove pension liability values higher. Real investment returns over the first half of the 1970s were negative. There was political pressure to inflation-protect pensions in-payment, and to maintain (in real terms) the accrued benefits of withdrawing members. Taken together, and with the possibility that the high-inflation/negative real return environment could be a long-term feature of the economy, it was perhaps unsurprising that radical solutions to defined benefit pension scheme funding were being put on the actuarial table.

In this context, Macdonald noted the outspoken presence of ‘those who regard funding [of defined benefit pension liabilities] as an outdated shibboleth of the actuarial profession’.[3] Macdonald expressed some sympathy with this position and, in particular, questioned the reasonableness of advance funding pension liabilities in a high-inflation, negative-real-interest rate economic environment. His point was two-fold: in such economic conditions, the funding of inflation-linked liabilities would be very demanding for sponsors, especially if pensioner inflation increases were included, and this would likely be a time when employers could least afford to contribute the additional cash; and secondly, he found it unconscionable to invest in assets that offered a negative real interest rate. In his own words, ‘the acquisition of new assets during such periods [of negative real interest rates] is really impossible to justify’.[4] Macdonald suggested that, in such circumstances, pension contributions should be limited to those required to fund immediate benefit outgo, and no contributions for the funding of the accrual of longer-term liabilities should be sought.

Macdonald’s ideas were quite controversial, especially coming from a leader of the profession. After all, he was proposing that contribution rates ought to be reduced at the very time when an economic view of the pension fund would suggest contributions were required more than ever—both because the funding level would be at its weakest and the credit risk of the sponsor would be at its greatest. An economist working from the premise that the primary rationale for advance funding was to ensure a high level of pension fund member security might be forgiven for thinking the actuarial dark arts were becoming too clever by half.

But President Macdonald was not the only actuary expressing reservations about advance funding of pension liabilities during this period. The same year saw the publication of a paper by J.R. Trowbridge in the Journal of the Institute of Actuaries which covered similar ground.[5] Trowbridge looked overseas for examples of how unfunded private sector defined benefit pension provision could function. He found that the French had been running unfunded occupational pension schemes with some success for several decades. Under this system, which was translated as assessmentism, the employee security concern of a non-funded approach was mitigated by having industry-level pension schemes, potentially with hundreds of different employers as sponsors collectively contributing (together with members) whatever contribution amount was required to meet each year’s pension outgo as it arose. Thus the singlename credit exposure that was usually borne by British pension schemes was diversified (though an industry-level credit exposure would remain). These ideas naturally met with some suspicion amongst the UK actuarial profession. There was a concern that much of the French system was based on a voluntary cross-generational transfer that relied on notions of socialist solidarity that did not naturally fit with the British culture of self-sufficiency. Above all, there was a hope and faith that the economic norm of positive real returns in excess of salary growth would return to Britain’s pension funds, bringing with it a return of financial viability for traditional British pension schemes.

The actuarial debate of this fundamental aspect of defined benefit pension provision—to fund or not to fund—continued in the following years. Boden and Kingston followed their President’s lead and further discussed the relative merits of advance funding and pay-as-you-go financing of final salary schemes in a paper published in the Transactions of the Faculty of Actuaries in 1979.[6] They agreed with Macdonald’s central point that in a negative real interest rate environment the case for pay-as-you-go was greatly strengthened. However, Boden and Kingston went on to argue that the 1974-1975 inflation experience and the poor real equity return performance of 1968-1978 were aberrations that did not alter their forward-looking expectation of long-term positive real returns on pension fund assets. They also reviewed the French assessmentism approach discussed by Trowbridge and concluded: ‘Once we are reasonably happy that it is possible to accumulate interest on capital at least as fast as inflation, there seems no compelling argument for going through the trauma of adopting the French system.’[7]

Boden and Kingston were even sufficiently sanguine about future real interest rates to argue that explicit allowance for post-retirement inflation increases could and should be made in pension fund valuations and contributions. This position was supported by influential actuaries such as A.D. Wilkie, who made the same argument in another Staple Inn discussion a couple of years later:

I see no reason to assume a long-term real rate of return at present [1982] market levels of less than 4% in excess of prices. At that rate of interest, how many pension funds are in surplus, and how many cannot afford index-linked pensions?[8]

The degree of uncertainty and variation in long-term economic expectations amongst senior actuaries over this period of 1977-1982 is striking: are long-term expectations for real asset returns negative or 4 %? There was little doubt that, whether using advance funding or pay-as-you-go, the economic cost of defined benefit pension liabilities would be unsustainably high in a permanent negative real interest rate environment. A 4 % real interest rate, on the other hand, would make generous inflation-linked increases beyond the level of guaranteed benefits quite affordable. This highlighted the fundamental challenge of setting a funding objective based on subjective expectations for the behaviour of economic variables over exceptionally long time horizons.

But the optimists prevailed. British actuarial orthodoxy on the advance funding of final salary pension funds survived the economic turmoil of 1970s stagflation and its consequent financial market volatility. Moving from the late 1970s to the 1980s, actuarial debate shifted from whether or not to advance fund pension liabilities and on to funding methods and what the objectives of funding should be—back to the debate that Gilley started in 1972 which had lain largely dormant for a decade. Rather like the contemporaneous experiences of their life office brethren, influences and demands from outside the profession were the main catalysts for pension actuaries to revise and evolve their traditional approaches. In particular, and again with parallels with the life office experience, the accounting profession and government regulators had a telling influence (the Accounting Standards Committee was working on an exposure draft on treatment of pensions in company accounts; the Occupational Pensions Board was considering solvency disclosure and inflation protection for early leavers and pensions in payment).

The accounting profession was increasingly active in setting new standards for the reporting of the pension expenses in financial statements. Historically, the pension expense may have been accounted for simply as the amount the sponsor paid in contributions over the accounting period. As contribution payments became more volatile, accountants became more focused on a measure of the cost of pension benefit accrued over the accounting period. In the UK, actuaries’ preference was generally to keep funding and accounting expense methods as similar as possible. There was therefore a general preference to base this measure of cost on the stable long-term percentage of salary roll that the actuary assessed would be required to meet the current members’ future benefits net of the existing asset fund (the ‘aggregate method’). However, accountants were increasingly aware that this contribution rate did not necessarily equate closely with the cost of the liabilities that accrued over an annual reporting period. In particular, the cost accrued might be expected to increase later in the life of a pension fund as its membership aged. So methods such as the projected unit credit method, which focused on the contribution specifically required to meet the increase in the pension liability that resulted from a year’s further accrual, were preferred by accountants.

These unit credit methods were in increasingly common actuarial use in the 1970s but were initially somewhat unpopular with the British actuarial profession as it challenged the traditional view that the assessed contribution rate should be a stable long-term percentage of salary roll rather than a figure that would be expected to increase as the scheme matured. Colbran’s

1982 paper reviewed this situation and called for more professional guidance to be provided by the Institute of Actuaries on pension funding methods.[9] He suggested a list of acceptable methods should be created with appropriate terminology and definitions. This suggestion was acted upon and a ‘Report on Terminology of Pension Funding Methods’ was jointly published by the Institute and Faculty of Actuaries in 1984.[10]

Colbran also argued in favour of a market value approach to asset valuation rather than the discounted income approaches that dominated actuarial pensions practice at this time. The Staple Inn discussion of the paper reopened the pension asset valuation debate that had lain largely dormant since Day and McKelvey’s 1963 paper. In the discussion, the established actuarial preference for discounted income values over market values was clear. But this view was not unanimous, and there was a wide recognition that the first priority of the valuation method should be to ensure consistency between asset and liability values. This reopened the door to the idea of market valuation of assets and market-consistent liability valuation, as espoused as long ago as 1921 by S.G. Warner. In opening the discussion of the Colbran paper, Paul Thornton (who would go on to become Institute President in 1998) stated:

It is possible to arrive at a set of assumptions which, when used to discount the future interest and capital proceeds from the investments, would reproduce market values. There is no reason why a valuation should not be made incorporating assets at market value if the liabilities are valued on consistent assumptions derived in this way.[11]

However, some senior pension actuaries in the discussion did indeed find reasons why this approach to consistent asset-liability valuation should not be adopted. D.E. Fellows noted:

To rely wholly or largely on market values could lead to much volatility in the rates of interest used from valuation to valuation.[12]

And C.W.F. Low similarly stated:

This [market-consistent liability valuation] would mean constantly changing valuation bases to reflect variations in the initial rate of interest obtainable ...

this approach is dangerous, as it would concentrate employers and trustees’ minds on short-term investment performance which is not appropriate to the real investment needs of pension funds.[13]

1970s stagflation faded into distant memory as final salary pension funds returned to financial health in some style in the first half of the 1980s. UK price inflation appeared to have been tamed: in the mid-1980s, the UK annual inflation rate moved within the range of 3-5 %. Equities performed exceptionally strongly—the FTSE 100 increased by more than 125 % between the start of 1984 and the October 1987 crash. Real dividend growth was around 5 % per annum. Meanwhile, large-scale industrial redundancies further contributed to pension fund surpluses (redundancies shifted active members’ benefits into deferred benefits that were no longer linked to real salary growth and hence reduced their value).

This turn of events may have afforded actuaries some latitude to think more broadly about funding objectives and targets. In any event, the fundamental objective of funding became a hot topic again. D.J.D. McLeish and C.M. Stewart were particularly instrumental in reopening the arguments presented by Gilley back in 1972 for focusing funding on ensuring that the cost of discontinuance benefits could be adequately covered by the pension fund at the time of the valuation. McLeish restarted this debate with a Faculty paper in 1983 which emphasised member security as the paramount rationale for advance funding of pension liabilities.[14] He expressly rejected the Faculty President’s suggestion that funding should be deliberately suspended in times of negative real interest rates:

In private occupational schemes I believe that funding is necessary to provide security. The higher the yield on assets the better but there is no particular significance in the point at which the difference [between investment yield and salary inflation] changes from positive to negative.[15]

McLeish’s funding-for-security arguments were further marshalled in a paper[16] co-written with C.M. Stewart in 1987 where they argued that:

The prime purpose of funding an occupational pension scheme must be to secure the accrued benefits, whatever they might be, in the event of the employer being unable or unwilling to continue to pay at some time in the future. To that end, the contributions would have to be sufficient both to pay the benefits as they fell due for as long as the scheme continued, and also to establish and maintain a fund which would be sufficient to secure the accrued benefits in the event of contributions ceasing and the scheme being discontinued, whenever that might occur.[17]

A practical complexity that arose in consideration of this type of funding target was the ambiguity that might exist in the pension scheme’s trust deed and rules as to what members were entitled to on wind-up of the scheme. However, by 1987, statutory minimum benefits for early leavers existed that linked accrued benefits to the lower of future national average earnings inflation and 5 %. This provided an unambiguous minimum benchmark for wind-up benefits.

The approach of funding for discontinuance benefits was still generally unpopular with British actuaries at this time. Colbran’s 1982 paper that discussed the need for greater professional guidance on funding methods had even called for this particular funding method to be outlawed by the profession on the grounds that it could produce a substantially lower contribution rate than other methods. But McLeish and Stewart argued that this would only be the case for immature schemes. They rejected the relevance of the level contribution rate that would ultimately fund all liability cashflows and instead focused on ensuring that accumulated assets were always sufficient to meet the accrued liabilities. They further argued that the level contribution rate approach would tend to result in the accumulation of unnecessary surpluses relative to the accrued liabilities.

The responses to the paper in its Staple Inn discussion were generally negative. Traditional actuaries were concerned that the specific targeting of the security of the accrued benefits would actually result in less security by slowing the pace of funding. But some recognition of McLeish and Stewart’s arguments was won. The influential Paul Thornton agreed that the funding of accrued benefits should be part of the funding objective:

I too believe that it is necessary for the funding policy to be such that the assets are normally sufficient to cover in full the winding-up priority levels. From the point of view of security for the members’ benefits, this defines a minimum funding level.[18]

Like other developments in actuarial thought that emerged in a climate of traditionalist scepticism, once the genie was out of the bottle, it would never go back in. The next notable contribution to British actuarial thinking on the funding of final salary pension schemes arrived in a paper, ‘A Realistic Approach to Pension Funding’, by Paul Thornton and A.F. Wilson in 1992.[19]

This paper was ostensibly traditionalist in outlook, continuing to advocate a discounted income approach to asset valuation and funding methods that were focused on producing long-term contribution rates rather than funding current discontinuance benefits. But, rather like McLeish and Stewart, they argued that typical funding calculations could lead to over-funding. They noted the ‘embarrassingly large surpluses’ that prevailed in many pension funds at the time.[20] Their central argument was that these surpluses did not arise because of any inherent limitation in the traditional pension funding methods, but because the parameterisation of these methods generally included numerous prudent margins that were cumulatively unnecessary and sometimes accidental. They argued that the typical valuation assumption of a real return of 4-4.5 % was lower than the expected real return on equities that most pension fund assets were invested in. They believed a 5-6 % expected real return on equities was reasonable based on historical analysis and forward-looking economic arguments. They suggested that salary inflation tended to be lower than the national average for the age group of over-40s, and it was this group that was most important to the valuation of active members’ benefits. They estimated that this over-stated salary inflation by around 0.5 %. According to Thornton and Wilson, these and other margins were generating unintended surpluses.

In a similar vein to Thornton’s comments in the McLeish and Stewart discussion, they regarded the discontinuance position as an important reference point, though not necessarily the central driver of the contribution rate. They even went a step further, arguing that a margin of 20 % over the discontinuance cost should be maintained in light of ‘the speed with which surpluses and deficits can build up’.[21] To paraphrase, they were arguing that if the purpose of the assets was to meet a market value of liabilities in the near future, it was not sufficient to have assets that met that value of liabilities today—an additional buffer would be required to the extent that the asset and liability values could diverge between now and the possible dissolution event. This logic is essentially that of the value-at-risk concept that emerged from the banking sector and, in the twenty-first century, would become the basis for insurance solvency regulation.

A number of speakers in the Staple Inn discussion of the paper noted that the cost of transferring discontinuance benefits to insurance companies had increased considerably in recent years (partly due to improvements in early leaver rights and partly due to lower interest rates). For the first time, the cost of the non-profit immediate and deferred annuities that would meet the discontinuance benefits might typically exceed the liability values assessed by the traditional ongoing funding assessments, and indeed the market value of assets in the pension fund. Between the publication of the paper and its discussion at Staple Inn, the Robert Maxwell affair occurred, placing discontinuance security more squarely in the public eye than ever before. Thus, actuaries’ ‘embarrassing surpluses’ had to be simultaneously reconciled with increasingly transparent discontinuance shortfalls. This was a significant challenge to the actuarial profession’s credibility and communication skills. It was a challenge they struggled to meet. Ultimately, it was taken partly out of the profession’s hands when the government legislated the introduction of the Minimum Funding Requirement (MFR) for final salary pension schemes in 1995.

The MFR was a controversial and contentious requirement from its inception. It was a result of a committee, the Goode Committee, which was established by the government in the wake of the Robert Maxwell scandal. However, the recession of the early 1990s was another driving factor for the interest in legislation to protect pension fund members’ accrued benefits. A number of pension scheme wind-ups occurred during this recession where members did not receive their expected benefits, and this experience further contributed to the sense that the security of accrued benefits was not being adequately protected by the prevalent actuarial funding methods. So the MFR was first intended as a robust measure of the adequacy of the value of assets in the pension fund to meet the market cost of accrued benefits. However, once the financial implications of using the insurance annuity basis in the pension valuation were realised, a political compromise was made and the MFR valuation basis was fudged: the expected return on equities would be used as the discount rate for accrued benefits (except pensions in payment), even though this would undoubtedly understate the cost of securing the accrued benefits in the event of dissolution of the scheme.[22] Consequently, being 100 % funded on the MFR basis did not mean the pension fund necessarily had close to sufficient assets to meet the cost of the accrued benefits in the event of dissolution of the pension fund. The result was unsatisfactory for all concerned: actuaries lost more of their professional independence and ability to self-regulate; government had created a costly regulatory framework that did not meet its intended purpose; members may have been led to believe their benefits were secure when in fact they were not. As with their life office brethren, the 1990s was proving to be a tough decade for British pension actuaries.

  • [1] Gilley (1972).
  • [2] Macdonald (1977).
  • [3] Macdonald (1977), p. 13.
  • [4] Macdonald (1977), p. 14.
  • [5] Trowbridge (1977).
  • [6] Boden and Kingston (1979).
  • [7] Boden and Kingston (1979), p. 412.
  • [8] Wilkie, in Discussion, Colbran (1982)), p. 410.
  • [9] Colbran (1982)).
  • [10] Turner et al. (1984).
  • [11] Thornton, in Discussion, Colbran (1982), p. 387.
  • [12] Fellows, in Discussion, Colbran (1982), p. 390.
  • [13] Low, in Discussion, Colbran (1982), p. 406.
  • [14] McLeish (1983).
  • [15] McLeish (1983), p. 275.
  • [16] McLeish and Stewart (1987).
  • [17] McLeish and Stewart (1987), p. 155.
  • [18] McLeish and Stewart (1987), p. 221.
  • [19] Thornton and Wilson (1992).
  • [20] Thornton and Wilson (1992), p. 263.
  • [21] Thornton and Wilson (1992), p. 264.
  • [22] See Greenwood and Keogh (1997) for a detailed description and discussion of the actuarial bases of theMFR.
 
Source
< Prev   CONTENTS   Source   Next >