|
The opdu Report - Issue 20, June 2006 Nick Eyre, Group Secretary of the Co-operative Group, shares as a case study the Group’s move on ‘A’ Day from three final salary schemes to one career average scheme. The following are extracts from a presentation given by Nick Eyre at the Pensions Age Conference earlier this year. To give you some perspective of who the Co-operative Group is, we were founded as CWS in 1863 and in the 1920s were among the very first employers to offer, to all staff, a final salary pension scheme. We are perhaps therefore coming to the issue of pension provision from a particular, if you like, paternalistic perspective. We may, I think, say we have some pedigree in this regard. With approximately 70,000 employees across our retailing and financial services businesses (of whom over 30% are part time employees), and an average salary across the 3 principal employers in the Group of under the national average salary of £23,000. We do offer all our staff access to one of the Group’s pension schemes though historically we have markedly different take up rates across the three schemes. |
 |
|
|||||||||
|
The Group has assets of over £17 billion outside the long term business assets of £23 billion in the policy owners’ long term business or life fund. We are the UK’s largest convenience store operator, largest funeral business and have significant businesses in Travel, Pharmacy and Farming as well as owning The Co-operative Bank and CIS. Within the Group, there were, until 6th April, three principal pension schemes. The Co-operative Group scheme has 60,665 members made up of 11,764 active, 22,780 deferred, 26,121 in payment assets of £2.4 billion and liabilities on an FRS17 basis of £2.1 billion this equates to a funding level of 118% on an FRS17 basis. On a solvency (or wind up basis), the scheme was approximately 83% funded. While a little less well funded and a little smaller, the CIS fund was also relatively well funded at 113% on an FRS17 basis or 78% funded on a wind up basis and equally mature in membership profile. The Bank scheme which had been “hived down” out of the Group scheme in the eighties had a differing membership profile, was much smaller and although perhaps relative to much of the UK private sector, not significantly so, relative to the remainder of the Group inadequately funded. These funding levels as of the beginning of April and the demise of the three schemes had varied from the position at the outset of our review, though the broad balance remained truly reflective of relative positions and therein, of course, lies the biggest issue for an employer who wishes to offer quality pension provision through a defined benefit pension scheme. Equity markets and Bond yields, on which under current accounting standards liabilities are measured, are very volatile and this impacts an employer’s balance sheet and profit and loss account. As a mutual, we are not subject to the immediate vagaries of the stock market so it is perhaps less a stark issue for the Co-operative Group than listed entities like British Airways, BAA and others who have had highly publicised difficulties but an issue it is. The Conundrum How can you address a desire to provide a defined benefit scheme but not be over exposed to the risks a defined benefit scheme apparently brings with it. Among my roles as Secretary of the Group, I chaired the Group’s Risk Committee which, reporting into the Group Audit & Risk Committee, had back in 2003 identified pension risk (broken down into its two most significant component parts of mortality and investment (but in practice equity) risk as the greatest risk to the long term survival of the Group. To give you some quantification of this risk, using stochastic modelling techniques we had developed in addressing our exposures in our Life Fund to meet our BASLE regulatory obligations, we estimated the downside risk (at the 95th percentile) over 10 years, when interest rate and inflation risk were added, as £1.4 billion. Therefore there was a 1/20 chance that over the next 10 years we would lose £1.4 billion with current bond equity splits, if we remained “naked” on mortality risk. This was significantly greater than competitor risk to the Co-op Group (and we face along with the rest of the sector significant competitor risk in retailing) credit risk (major bad debts in the Bank) or equity risk through interest in CIS our composite insurer’s General & Life businesses (where steps were already being taken to address these issues). In fact, we identified the pension risk as four times greater than any other significant risk to the Group as a whole.
Lightening mortality rates means state as well as private sector provision has to move on. The slide below shows the changing mortality rates for 65 year olds over the last century.
We demonstrated multiple funding scenarios using differing return seeking assets/bond splits with the range of outturns. While inevitably the upside is significantly greater with higher levels of equity investments the damage sustainable at the 95th percentile is frankly unacceptable to an organisation that is planning to be around for another 150 years. The risk is asymmetrical - the employer carrying the downside risk as effective guarantor but not the upside which would remain in the pension scheme. In the Co-operative Group and in any defined benefit scheme, the two major risks inherent in the pension schemes are plainly mortality and investment risk but in our case they are compounded by our business. As mortality rates lighten not only do our pension liabilities expand but our exposures (within the Life fund) lengthen and if you think about it, the turnover (if that is the appropriate term) and therefore the profitability of our funeral business declines. While volatility in equity markets impacts not just our £5 billion of pension fund assets but £2.5 billion of shareholder assets in our general insurance business and £23 billion of assets in our life business in so far as they are invested in equities. As a result of the unsustainable risk profile, we set up a working party across the Group to analyse just how we should address pension provision across the Co-operative Group in the 21st century, balancing the need to reduce the unacceptable cost and risk to the Group, its members (our owners) and its 70,000 employees that are created by the three final salary schemes, against the other side of the conundrum which was created by our strong desire as a, hopefully, progressive and socially responsible employer to offer quality pension provision in which we did not abdicate mortality and investment risk to our workforce who are in reality not well positioned to address it. This at a time when it is increasingly obvious that the state will not be able to make effective pension provision for an ageing workforce. How have we sought to resolve these difficulties as a responsible employer who takes pension provision seriously? This internal Pensions Strategy Review Group made up of senior figures from each of our businesses as well as Pensions, HR and Finance professionals rapidly concluded, not surprisingly, that the status quo was simply not a viable option. It was simply financially not sustainable. The Bank scheme, as had been normal in 1980s financial services, was a non-contributory scheme for employees, while the others had differing contribution rates (averaging at 6%) for employees. The Bank had a retirement age of 60. The Group and CIS 65 (but a practice of allowing retirement at 60 without actuarial reductions). There was a significant discrepancy in funding levels: on FRS17 basis (plus 6%) in the Group to a 29% deficit in the much smaller Bank scheme, in part of course because of the more ‘generous’ scheme design, but also because of a practice of weaker contributions from the Bank over a number of years had created this divergence. While in absolute terms these shortfalls were minor, £264m in total on an FRS17 basis, but more worrying was the future (true) economic costs which we calculated, on an 80% gilts, 20% equity basis (the lowest risk basis our asset/liability matching study had shown us). This showed a true economic cost of £125 million per annum employer cost across the three schemes, if continuing the then present arrangements, assuming then current membership rates. This in a business which in 2003 had retained profits earned for the year and transferred to reserves (after payments of £49m to the pension schemes) of only £309 million. In the Bank, the underfunding position was immediately addressed as an interim step by a special contribution in 2005 of £14.5m. Conclusions of the Initial Review The first conclusion (the conclusion I would respectfully suggest that all employers, who take their responsibility to their workforce (and their pensioners) seriously would reach) was that any plan design should be affordable on a “best liability matching” basis and not dependent on the (uncertain) continued out-performance of equities. Employer contributions and asset allocation are separate decisions to what is and is not affordable. In fact, it is astounding how few UK defined benefit schemes appear to calculate pension costs on anything other than the ongoing valuation basis, assuming (I’d suggest perhaps recklessly) significant future equity returns. Our conclusion after lengthy discussion with the Director of Finance, who like me has to grapple with the classic conflict of being both a Trustee and on the management committee (or Board in plc speak) of the Group, was that a scheme costed on a minimum risk basis at 16% of gross pay (if contracted out) this equated to 13.4% if contracted in due to increased employer NI contributions was affordable on a sustainable basis (ie adopting prudent assumptions concerning increases scheme membership). The increased costs and financial, legal and governance risks of running three separate schemes and the inappropriate use of scarce resource, Group capital, which the differing funding levels of the three schemes represented all drove us to the conclusion that a merger of the three schemes was highly desirable. While the reduced costs of internal administration, lower legal, audit and actuarial investment manager costs and the simpler communication message which one scheme offered. As we formulated these recommendations the Pension Protection Fund details started to emerge and we assumed (correctly) lower PPF levy would be payable under one scheme. However, we were driven by the more subjective derisking issues. To our mind, the investment, contribution, tax and governance risks in particular were all significantly multiplied by having three independent trustee bodies responsible for significant liabilities (in excess of £4 billion) underwritten by the Group. Scheme Design Having identified an affordable cost and the need to move to one scheme, we turned our attention to scheme design setting this against the Group’s business background, our desire to understand better and control risk, competitor practice, inevitable difficulties in implementation and our Co-operative ethics. Our risk averse approach pushed us to a defined contribution approach in which we would be committing a finite cost. This would give certainty to the employer in costs, and transfer almost all risk to the employee who with uncertain benefits would bear notably the mortality and investment risk. With mortality rates lightening and stock markets increasingly volatile and both bond and equity returns perhaps entering periods of diminished returns on both an absolute basis and relative to wage inflation, if we take a longer term perspective, this at one level offered us an attractive option; an option much of the private sector has been seduced by (in what has become a paradigm shift), especially for new employees. Many have, of course, also by sleight of hand chosen to reduce significantly their levels of contribution as an employer. However, we were not minded to offer differing pensions provision for ‘seasoned’ and ‘new’ employees or indeed for senior Executives and other staff. This appeared to us wrong headed on three bases. Firstly, it would be divisive in the workforce having one person working with another doing the same job with one paid, in overall terms, because of differential pensions significantly more than the other. Such differentiation would also, I believe, have been a potential area for future equal pay claims. Secondly, with turnover of staff in the pension scheme at under 10%, the impact on the true cost of funding of changes for new starters only would be slow to take effect on our pension funding and our balance sheet (we estimated at least 15 years to have a full impact). Thirdly, it hardly offered the leadership necessary to drive the necessary pension change if we had one scheme for the top tier of management and another for the mass of our employees. We should offer to all staff as a Co-operative employer, good pension provision in which we, as the party most able to do so carried the significant investment and mortality risks appeared to us philosophically right. We were uncomfortable offloading these risks to our employee stakeholders who are not equipped to adequately address them. We concluded that we should, if at all feasible, seek to offer some form of defined benefit scheme. Assuming we could reduce the downside risk to acceptable levels, though given our cost parameters this would need to be at a reduced accrual rate.
In our composite insurance business, CIS, we had in its life fund to some considerable degree addressed the same risk issues and were successfully derisking our business in terms of mortality and investment exposure. Within CIS, significant elements of the mortality risk with some liability for policies in payment had already been sold into the market, while work was underway on reinsuring even risk of those policies not yet in payment. With the assistance of stochastic modelling the CIS life fund had, as we had already in the largest of the three pension schemes sought to understand and manage our equity risk and more closely match our bond profile of future pension payments by increasingly moving to better asset/liability matching with a more bond weighted asset base. In 2004, our proposal was that by the end of 2006 we move to a more risk averse liability matched asset profile in a merged scheme. Our current allocation of assets left so great an exposure to unmatched assets with an allocation of 67% return seeking assets, the downside risk was too great and we needed to move into this box. The downside risk would be further reduced with improved alpha ie upside leverage of downside risk with a more diversified series of return seeking assets including for example currency management and commodities as an asset class. This diversification further reducing the downside risk of investment underperformance to the pension scheme. So, pulling all this reasoning together, our conclusions were that we should move to a career average scheme rather than stay with a final salary scheme. This appeared to us to have a number of advantages. While not in itself addressing either the mortality risk or the investment risk, it would have a significant impact on the salary inflation risk. It was more equitable in that it recognised performance (and therefore salary) over a career with the Group rather than the last year or best of the last three years, important we felt in a business with a high proportion of part time workers and a likelihood of more flexible life/work patterns. It removed the inequality of the cross subsidy of those who are lower paid (and statistically likely to receive lower salary increases) to those who are higher paid (still in the future statistically more likely to be higher paid) which is inherent in any final salary scheme and it allowed us to retain 60ths accruals (with some judicious variation to non-core benefits and small changes to definitions of pensionable pay) which was important in getting staff and union buy in. Our 1/60 career average would have broadly equated to 1/80th final salary under our old Group schemes. It is also, for a mildly social reforming organisation, attractive as broadly redistributive from the more affluent management population to the less affluent shop floor. It was, in our view, however, that such a bold an approach could only be sustained if significant steps were taken to derisk pension provision. Limiting Pension Provision Risk Firstly, while the treatment across the three schemes had been inconsistent, broadly in each scheme, retirement at 60 was possible at no extra cost to the employee. We recommended that this move to 65, a move which, of course, shared the lightening mortality risk with our scheme members, though this and other changes we proposed should only affect future accruals. Secondly, to further derisk pension provision, we proposed that our pension provision contracted in to S2P, we estimated that this could reduce our pension risk exposure (if the government continued to offer this) for future service by up to as much as one third. Through the use of salary sacrifice the impact to pay packets could be ameliorated. As I will explain, this was perhaps the only significant part of our proposal we were unable to deliver. Thirdly, a closer matching of assets and liabilities moving over a period of time from equities to bonds better matching our liabilities and a further diversification of asset classes to include both currency and commodities, as an investment category. In addition, by utilising swaps and other derivatives in our non return seeking assets, we could better match our risk exposure. Fourthly, we should look to offload through reinsurance some of the very significant mortality risk we were carrying. These risk controls, in our view, had to be an essential part of this programme if we were to sustain a defined benefit scheme.
While our modelling suggests that our risk/return or alpha is maximised with a greater level of return seeking assets than the 20:80 equity/bond cost model we have moved and will move further over a period of time to a model in which only 50% of our assets will be return seeking the optimal model remains a matter of debate in our Trustee Investment Committee though as recent rises in the stock market and declining bond yields have shown timing of this switch is critical, though not as straightforward as it might at first appear. A steady rise of 20% over 12 months in the FTSE index has been almost matched by a very significant decline in bond yields driving up not only pension liabilities measured on IFRS and wind up bases but, of course, the cost of matching these liabilities. Implementation Our recommendation to buck the trend, which was rapidly becoming an unsightly clamour, away from defined benefit schemes for future pension accruals for both new and existing employees rapidly found favour across the respective boards of the Co-operative Group, CIS and the Bank when linked to a raised retirement age of 65, better asset liability matching and scheme merger. However, such an ambitious programme of change inevitably faced difficulties with Trustee bodies, staff and unions and potentially the new pensions regulator. The Boards were clear and determined however this had to be a package. Only if all of this risk limiting approach was achievable could the Group embrace the future risks inherent in a Defined Benefit Scheme. The differing stakeholders: 1. The Trustees The proposed merger of the Bank scheme with its £62m deficit on an ongoing basis and 61% funding on a solvency basis was never likely to be problematic as it merged into a scheme with a £359 million surplus on an ongoing basis and a solvency funding ratio of 81%. Even for the CIS Trustees, the second of the three schemes, overall the picture was positive. The scheme post merger would be better funded 81% merged (79% in the old scheme) on a solvency basis. The underlying employer covenant (at Group level) would be better and with the employer only prepared to offer linkage to future salary for existing active members’ historic accruals if a transfer was approved they concluded unanimously that the transfer of scheme assets and liabilities to a new “merged” scheme, should be approved. The issue facing the Trustees of the largest and best funded scheme was, at least on the surface more problematic. Each of the three sets of trustees had been provided with extensive independent legal and actuarial advice and while as an employer one shudders at watching expensive city lawyers and actuaries argue the points, this did allow the Trustees to approach the issues with a degree of self confidence. The effect in solvency dilution of merging the three schemes from the Group Trustee perspective (if we focus on the wind up basis and I think as a Trustee that is now the only important analysis in this situation), the solvency would be diluted from 88% to 81% a shortfall of £291m again on a wind basis would be transmogrified to a shortfall of £969m. As a matter of interest, on an FRS17 basis, a surplus of £172m would become a shortfall of £16m. Why should the Group Trustees agree to this? Well, firstly the employer wished them to do so and that is more powerful than perhaps at first cut you may think. Even in this relatively well funded scheme, the Trustees are reliant upon the employer covenant, in extremis, as there would be on a wind up basis a shortfall. There are clear and undeniable savings to the employer in these proposals. While the employer presented the case that by bringing about these changes that across a wider spectrum of issues, the covenant would be strengthened. The Group, the principal employer in the Group scheme, already essentially underwrote all three schemes. Secondly, the linkage of future salary increases to historic pension accrual (costed by the employer at £290m if all existing members continued to normal retirement) was a clear benefit to active members over whom the Trustees acted as fiduciaries. Thirdly, if this was approved, the employer committed for each of the next three years to increase cash payments to the scheme to match the economic costs (on the 80:20 basis) at 16% or £80m of pensionable salaries significantly increasing cash contribution from 10% or £50m. As a footnote to Trustee ‘sign up’ we sought simultaneously with the creation of the new CARE scheme to move, in the case of the Group scheme, in essence from a set of 1920s rules to a 21st century pension scheme. This involved the giving up of rights by pensioners to be balloted in certain circumstances. To concede this change on behalf of the pensioners, the Trustees demanded and received one-off compensation from the Group of £120 per pensioner at a total cost of £3.2m. In the end, we were able to secure unanimity, as I say, in the Group pension trustees also. 2. The Regulator The shock horror exposé that if the Regulator objected to the proposed merger the schemes would not merge, demonstrated little more than sensationalism in broadsheets. In reality, The Pensions Regulator expressed satisfaction with the proposed arrangements. 3. The Employee Stakeholder The last players in this case study, our employees, and the trades unions, were a further stakeholder who we needed to bring with us. The major trades unions had been notified and involved in the process for over 12 months ahead of our public announcement though as under our collective bargaining agreement, we have with them pensions are not included and they had not been involved in the initial pensions review body we respected their request not to call this process consultation. While certainly our larger unions USDAW, AMICUS and NACO, I believe, ultimately accepted the overwhelming case for change, acknowledged the desirability of the defined benefit/career average approach and accepted that in this changing pension environment we were actually the “good guys” and indeed we had shown some flexibility. They had negotiated some changes in the overall scheme design (including to our surprise and indeed our disappointment, the continuation of contracting out on the basis that for the lowest paid it did keep the total initial cost of pensions down). One has, of course, to acknowledge it is exceedingly difficult as a trade union to publicly come out in support of a diminution in overall remuneration packages to members, but their acceptance of the overall package was reflected in their low key press profile and lack of any industrial action. Ultimately, I would characterise the approach taken by the trade unions as responsible. As a quick tour of the horizon of this case study, I hope that this is of some interest and sets out the risk, cost and practical issues which the private sector faces in relation to pensions in the context of trust law, employee relations and employer responsibility. Nick Eyre |
|||||||||||
|
|||||||||||
