Con Keating and Iain Clacher
After critically reading and re-reading the consultation on the new DB Funding Code and its associated documents several times, we decided that we would not submit a response to it.
Our principal reason was that it read to us more like a marketing document than a true consultation.
But having engaged with numerous consultations to see no meaningful change in approach or willingness to steer a different course, regardless of the facts, we thought it would be better to write extensively and hopefully spark some meaningful debate in the run-up. We are always available if anyone including the Regulator wish to debate and discuss.
Regarding the consultation itself, it states that it has been prepared in compliance with the government’s consultation principles. It even lists nine key principles. By our reckoning, the consultation breaches more of these principles than it complies with.
The problem with marketing documents is that they do not invite challenge and criticism; in this case they strongly suggest that the Regulator is not interested in listening to or changing anything. Their reported reaction (one of being less than happy) to the Bowles Amendment supports that view. In other blogs we have criticised the use of misrepresentative language such as “objective”. The dominant question that a marketing document seeks to provoke is: ‘Where can I get it?’ It is the setting out of a stall.
One major issue with the current consultation is that there is no cost-benefit analysis in either the consultation or supporting documents, though we are assured, in the Executive Summary, that: “We will take account of…our assessment of impacts.” This naturally raises a question: if this has been done, why not share it with us? If it has not been done yet, on what basis is the support for so many of the assertions and assurances in the Consultation?
The Introduction of the Consultation implies an objective to: “…ensure DB schemes’ efficient run-off phase”, which naturally raised concerns for open schemes and led to the Bowles Amendment. Taken together, these points led us to investigate the cost, impact, and efficiency of the proposed Code.
We began by attempting to model the cost to the entire universe of pension schemes, which the Consultation tells us have current technical provision liabilities of £1.9 trillion [as at 31st March 2019 – para 98 of the Consultation Document]. Unfortunately, there is insufficient information in the Consultation or the public domain more generally to do this reliably. With many heroic assumptions and inferences, we simulated ranges of possible total cost outcomes.
The results ranged from £10 billion to £600 billion, with the 5% – 95% range being £80 billion – £200 billion and the most common outcomes lying in the range £100 – £120 billion. Perhaps the only significant result was that there was no instance of a net gain rather than cost. These results also take no account of the higher costs of administration, which we would estimate to be of the order of £500 million per year. We note that these simulations are far higher, perhaps orders of magnitude higher, than the risk exposures reported by the Pension Protection Fund. Given our low confidence in these results, we concluded that they could not form the basis of any further research or analysis and decided that instead we would investigate the costs and impact on a few small DB schemes[i]. We report here the results for one of these as it is extremely informative. We do not claim that this scheme is typical or representative.
We consider one small open scheme, which is large by comparison with its sponsor employer. It has technical provisions and assets of £25.8 million at December 31 2019, while the sponsor has equity of £8.3 million with annual sales of £11.25 million and pre-tax profits of £1.24 million. The sponsor employer is small but of prime credit quality.
Although the scheme is open, we treat it as if it were closed to new members and future accrual, and then consider the position 25 years from now as the long-term objective (LTO) horizon. The scheme is today in balance with an expected rate of return on assets of 6.1% pa; this is high by comparison with other reported discount rates, but it is lower than the scheme’s twenty-year historic returns performance of 7.6%. We report this historical investment performance figure for completeness but do not rely on it in any of the analysis which follows.
At this 25-year forward point in time, it should be funded to a self-sufficiency basis, which we take to be a discount rate of 1%. The amount outstanding at this future date is 30.4% of today’s total projected benefits but only 10.9% of today’s present value. At a 1% discount rate the scheme would be funded at 97% of the buy-out level and 91% of the ultimate projected liabilities would be 100% funded.
It is immediately apparent that the proposed Code is a far from comprehensive solution for any scheme, focusing on a small tail of member benefits.
At this time, the Macaulay Duration[ii] (at 1%) of the scheme would be 9.23 years and pensioners in payment would account for 63% of the projected liabilities. The scheme would be mature.
As an aside, we do not believe that duration is a suitable measure of scheme maturity, as its value is dependent upon the prevailing level of interest rates. The duration of a perpetual is the inverse of its yield, so a perpetual yielding 1% would have a duration of 100 years, while that same perpetual would have a duration of ten years at a 10% yield.
The present value of the residual liabilities at this future date using a 1% discount rate is £18.6 million. This is an increase of £5.7 million on the current projected funding required (a 45% increase). This has a present value at 6.1% of £1.3 million. This is the amount of special contribution required today to arrive at a 1% funding level, 25 years from now.
This is slightly more than one year’s pre-tax profits. It is 1.7 times current annual contributions. It would require shareholders to forego any dividend. This is well within the employer’s current liquidity of £1.1 million and existing unutilised overdraft facility of £2 million.
But the Code does not end here; it requires the scheme to have transitioned to a low risk, and low return portfolio (1%), in 25 years from now when this level of funding is required. This will further raise the cost to the employer. For modelling purposes, we choose a uniform rate of transition over the 25-year period. The portfolio’s expected return declines from 6.1% to 1% and estimated one-year volatility from 20% to 10%. We then make a very important choice: we assume that only the assets currently supporting these future claims (£ 2.8 million) are ’de-risked’. If we ’de-risk’ all, the cost rises dramatically.
The cost in this limited case rises to £11.9 million in future terms (2.1 times the earlier cliff-edge case). This would require a current special contribution of £4.97 million, four years’ pre-tax profits, and 44% of total annual revenues. In the opinion of management, with which we concur, this would simply not be supportable. It would preclude any new investment for at least three years, damaging future productivity and sales.
If we were to ’de-risk’ the entire portfolio, the position becomes even more dire. It would be a future shortfall of £15.7 million requiring a current special contribution of £6.6 million, 5.3 years of current pre-tax profits. This would be truly catastrophic; to quote one director: “If we liquidated the remains of the business, we might just be able to cover employees’ redundancy payments.”
These costs are material given the size of the scheme; respectively 19.2% and 25.5% of total scheme assets.
How much risk to the employer would be removed by ‘de-risking’?
We use as our metric the expected loss given loss. We assume one-year volatility of 20% for the existing portfolio and 10% for the ‘de-risked’ portfolio. In the current situation, the risk to the sponsor at the future date is £916,955, while for the ’de-risked’ case it is £1,068,535. In other words, the proposed new Funding Code would increase the cost to the employer of pension provision by a very substantial amount while also adding to the employer’s risk exposure. We wonder how this could be considered compatible with the Regulator’s duty to employers to minimise any adverse funding impact on the sustainable growth of an employer.
The employer and trustees are comfortable with a current risk exposure of £1.9 million, arising from the volatility of the asset portfolio; particularly so, given its odds ratio of 2.65 :1 (Expected gain given gain: Expected loss given loss). By contrast, they are not comfortable with £1.07 million exposure given its odds ratio of 1.35: 1. The employer has equity capital resources of £8.3million; its risk exposure is 4.3 times covered in the present case. By contrast, the risk exposure coverage in the two ‘de-risked’ portfolio circumstances are respectively 3.12 and 1.59 times. This is a deterioration of the sponsor covenant equivalent to that from AAA to B.
We find it incredible that the 30% tail of the distribution of projected benefits should be so risky that it requires an increase in funding of 20-25%.
The claims to efficiency of TPR’s approach are surprising and it is well-known that pure funding solutions are sub-optimal; as the age-old adage has it, prevention is better than cure.
With this in mind we next examine the cost of an insurance solution. This is a policy with deferred effect where, if the sponsor fails, the insurer steps in and pays the full benefits as originally promised. This is the cover which the PPF could and should have provided.
It has been suggested to us that it is not wise to provide the top cover via the PPF. However, we see the precise form of arrangement for the provision as a matter for later discussion. (See endnote[iii]) )
The cost today of this pension indemnity cover[iv] is £343,609.10. The policy also has a value as an asset of the scheme. The value today of the policy is £1,370,423.12. This policy asset will increase in value until the date at which cover commences and then decline rapidly as pensions are discharged. In addition, its value moves in a counter-cyclical manner; its value will increase when scheme assets fall in value. Perhaps the greatest attraction of such an approach is that it would allow the scheme to follow return-optimising investment strategies.
It is clear that TPR claims to ‘efficiency’ are, at the very least, open to debate.
If we return to our earlier attempts at macro-level estimations of the costs of the code. The most extreme outlier, £600 billion, would correspond to the apocalypse of all schemes failing entirely unfunded at the objective date. In fact, the low estimates of cost are as high or higher than the highest estimates of the risk posed by schemes. It appears that this low-dependency ‘solution’ costs five to six times as much as the likely losses arising from expected sponsor and scheme failure. The cost to the taxpayer in lost corporation tax receipts is also substantial (as tax relief is given to these ‘costs of the code’), and particularly meaningful in the current pandemic circumstances.
The Regulator has failed to make any case for a special regime for schemes in run-off, let alone open schemes.
We began this series of blogs and articles with a call for a bonfire of regulation; at the very least we should start with this proposed Funding Code.
[i] We examined four scheme in total, varying in size from the £25 million of the reported scheme to a maximum of £180 million of liabilities. The results for these other schemes were broadly similar to those reported. For two of these schemes we do not have access to sponsor financials.
[iii] A range of issues have been raised and some suggestions made. For example: All of the current members in the PPF will ask for an upgrade. Better politically to require a compulsory mutual insurance or support fund – perhaps with 10%-25% of the annual premium being born by scheme members and collected by reducing future revaluation/ pension increases- should be weighted by value of pension – and can put in a exemption for small pensions.
[iv] The cost of this cover may be calculated in a variety of ways from the complexities of forward start strips of credit default swap contracts to differences in single premium current start policies. The results are not highly sensitive to the method, though they might be relevant in an active traded securities market. The figures quoted in this blog are derived from the difference between two policies – one covering the entire period and one covering the entire scheme tenor.