I’ve blogged before about how putting employer covenant into establishing a funding target is as sensible a concept as zombie cats. Therefore, whilst the new code of practice published last week is markedly better than the first draft, it is disappointing from my perspective that such a flawed concept continues to have such high billing.
Probability and binary outcomes
To briefly summarise my thinking on this again we need to consider binary outcomes.
A binary outcome is one where there are only 2 possible results. A great, albeit morbid example, is that this year I will either be alive at the end of the year or dead.
Now I might want to think about the impact of those scenarios. If I am alive then (hopefully) things will be good and I’ll continue to earn money to provide for my family. However, if I’m not then my family could be in real difficulty. What should I do?
Well I could look to try and save up enough money to make sure my family is ok if I died (let’s say this is £500,000 for arguments sake). However, this is utterly impractical, it would take years and how would we eat in the meantime?!
The probability of me dying this year is quite low though (about 0.06% based on current tables) so really all I need is 0.06% of £500,000 which is £300. I can manage that. Now if I die my family will be protected.
Oh wait no they won’t. They’ll just have £300!!!
The probability of my death is irrelevant as it will either happen or it won’t. Saving the average amount needed doesn’t help in either scenario. If I survive I’ve put aside £300 I could have spent and if I die it is woefully inadequate to provide for my family.
Applying a probability for an individual event like this therefore doesn’t make any sense.
Of course the solution to my problem is simple. I should buy some life insurance.
By using insurance I can pay my £300 (plus an expense/profit allowance) and know my family is protected in the event of my death.
This works because the insurer is taking lots of £300 premiums. Let’s say they have 10,000 policies. It would now be reasonable to expect that 6 of those policies would involve a pay-out which should be covered by the £3m of premium income.
Insurance is an efficient way of covering such risks.
Pension scheme risks
Much is made in the new code of practice on funding of the significant risks involved in funding a pension scheme. From the member (and hence trustee) perspective though, there really is only 1 risk, the risk that the employer sponsor runs out of money. Provided the employer is around the benefits will get paid no matter what the current funding level.
This risk is again binary. Either the employer will survive and continue to make profits or it won’t. We might be able to make judgements on how likely it is but this is just a probability. The actual outcome is still binary.
So what is the solution? Well as is the case for managing my own mortality risk the solution is insurance. To fund schemes to the worst case scenario is (like me saving £500,000) completely impractical.
The government introduced such insurance in a compulsory format in the form of the PPF and PPF levies.
However, at the same time the Pensions Regulator introduced the concept of employer covenant strength.
Employer covenant strength is akin to the probability of insolvency.
An employer with a good covenant is like a person that eats well and goes to the gym a lot. An employer with a bad covenant is like a person that eats too much junk food and drinks too much. It changes the probability of outcome but not the range of outcome.
As I showed before, trying to use such a probability to manage a binary outcome does not make sense. Yet the code of practice says:
It (employer covenant) should help the trustees decide how much risk it may be appropriate to take (ie when they set their technical provision assumptions and investment strategy
With the intention that technical provisions should be higher for a weak covenant and lower for a strong one clear from some of the examples:
a low value for technical provisions based on a strong employer covenant assessment
But this is just like me putting £300 aside to cover the likelihood of my death. In fact worse, it’s like me doing it twice, once put aside and once as insurance (levy).
It won’t be enough if the employer fails and is too much if it doesn’t. It just doesn’t make sense.
The code actually acknowledges the impossibility of knowing whether an employer will be around in the long-term:
It is unlikely that trustees will be able (with any degree of certainty) to assess the employer covenant too far into the future
So surely in the majority of cases it is not relevant?? There is a bar to cross perhaps that some could fail, reasonable expectation to be around maybe, but after that, and for the majority of employers, a long-term view should be taken that doesn’t differ by employer.
Funding allowing for the PPF
Acknowledge the existence of the PPF and the single risk faced by members and trustees is largely managed.
There is perhaps some debate around the level coverage of benefits and I certainly wouldn’t be averse to increases in compensation but agree this and the possibility of employer insolvency can be ignored.
Schemes can then focus on running for the long-term on the assumption that employer support will always be there. Funding can be set at a level to broadly reflect the expected cost of providing benefits. Risk management becomes about managing risks for the employer rather than to protect members. Prudence used to allow an employer to reduce the volatility of its contributions rather than build a capital reserve.
In fact, as the reason for funding in the first place was to provide some level of security there is a reasonable argument to not require funding at all. (I don’t quite subscribe to this as I think there is merit in paying contributions at the time of accrual so cash cost and services gained are aligned somewhat and intergenerational issues mitigated.)
Can we allow for the PPF in decision making?
Acknowledging the existence of the PPF is an interesting concept. There is no law to state that the PPF should not be considered in scheme funding decisions but this is the common view propagated and believed by tPR due to its objectives. The reason for this is a court judgement Independent Trustee Services Limited v Hope and others in 2009. However, if you read the actual judgement the position is much more nuanced.
In particular, Mr Justice Henderson states that:
there is no single all-purpose answer to the question whether the PPF is a relevant consideration for trustees to take into account.
The judgement really states that the existence of the PPF cannot be used to justify something that would otherwise be “improper”. Additionally, a large part of the discussion of why this is the case relates to it being against public policy and not in the public interest.
In the context of scheme funding though, given the obvious efficiency and need to avoid double counting of capital and insurance, allowing for the PPF in the context of what it is there for would seem, to me, to be a very reasonable position to take and in line with the judgment made.
The current position is one that encourages excessive and often meaningless prudence which in turn often leads to excessively prudent investment strategies (if you’re funding to it then why take the risk – employer’s don’t generally think long-term enough to wait for money to come back).
There is much to like about the revised draft code of practice but the concept of employer covenant driving technical provisions doesn’t work. If the law allowed trustees to take account of the PPF it would be of great help. But even without this the funding code does not need to bring in employer covenant which is not mentioned in the law on funding.
Once insolvency risk is managed lower funding targets that are more reflective of actual expected costs can be used. The approach to funding can also be used to reduce the volatility of contributions (funding doesn’t impact on cost, just pace of funding don’t forget). In this sort of environment a highly valued benefit could be provided efficiently with members having reasonable levels of security in outcome. Isn’t this what we should be aiming for?