“He wouldn’t let it lie” – Con Keating returns to discount rates!

Earlier in the week, the Pension Play Pen had its monthly lunch, missing was Con Keating who was busy explaining to his daughters why they had nothing to fear from the supposed deficit in the USS pension scheme. Con’s voice could and should have been heard at the meal but we all were shy of articulating his position.

Now that Con has got his mojo/internet connection back – he is setting the record right. This is what we would have got – two barrels – full blast – plus your questions answered if you are Steve Tiley, Bob Compton , Mike Ohtsuka. The publication of this blog is retrospective and remote chairmanship, probably the easiest kind!

Discount rates – Yet again!

This is yet another attempt to explain a fundamental misconception with respect to the discount rate applied in DB pension scheme valuation. It will do so in the most basic terms.

Suppose we have a counterparty who wishes to have a particular sum of money at some date in the future. They ask us what they will need to pay us now to deliver the desired sum at that date. We decide that we are prepared to offer this sum and set about pricing it. The pricing process chosen is discounting that future sum at a rate of interest. This is a non-linear process – the rate is constant but the amounts by which the price varies over time are strictly increasing with the passage of time. There is an inverse relation between price and discount rate.

The rate we choose is arbitrary in the sense that it might be anything. There is no natural rate that might or should apply. Many of the advocates of financial economics have been known to claim that the government rate should be used, with their argument being that this is a “risk-free rate”. This is a misreading of the tenets of financial economics. The government rate may be default risk free but it varies over time and this is risky. More importantly the rate at which we discount a sum does not transfer any of the properties of the instrument that generated this rate. Using a default-free discount rate does not make our contract default-risk free. It merely increases the price of the contract.

The rate we choose to use would normally reflect the desired profit that we wish (or need) to earn over the life of the contract. This usually involves an assessment on our part of the assets we may acquire with the price paid to us and the returns we may expect those assets to deliver over the life of the contract.

Once we have agreed a price and this consideration has been paid, this rate does not change; it cannot. If rates go up or down, the parties do not subsequently alter the consideration paid. This is and should be exactly the same with DB pensions contracts, though the benefits ultimately payable are defined parametrically rather than as a cash amount.

There are exceptions to this invariance. Obviously, things may change by the mutual agreement of the two counterparties – the contract may be explicitly varied in this way. A contract may also be varied due to the insolvency of the obligor, a subject to which we will return later in this note.

The rate at which the contract was struck determines uniquely the intermediate progression of the amount of the liability. At inception it is the sum received and at maturity the agreed sum payable. In days past, this intermediate valuation progression was known as the pace of funding.

This is a valuation from and for the perspective of the writer of the contract, the sponsor of a DB plan. For the beneficiary, the value depends upon their circumstances; a question of utility. However, this is the amount of their claim in insolvency proceedings or other liquidation. It is equitable to all other parties involved.

If we write multiple contracts, the situation does not change fundamentally – though collectively we need to consider the weighted average rate. This is the rate that we advocate should be used in DB pension scheme valuation.

It is clear that as market conditions change so will the rate that we want to apply for new business, but the overall book of business in force is determined by the terms on which those liabilities were acquired, or equivalently DB pensions awarded.

The problems with the current methods enshrined in legislation are that they treat all business past and present as being new business. These are counterfactuals to the real position. They tell us no more than what the contracts or schemes might be worth if these rates applied.

Reverting to the simple contract, suppose that there are reasons to have qualms about the financial standing of the obligor, then in this circumstance the beneficiary may seek to mitigate the possibility of their failure by taking security. The amount of security which may reasonably be taken is the value of the contribution at the time of inception, and that plus an accrual at the rate implicit for the time elapsed. Securing the obligation to an amount higher than this is pointless as in liquidation any amount over this accrued claim will be repayable to the administrator or liquidator of the obligor.

It is also worth noting that insolvency courts operate on the basis of the balance of probabilities, which means that claims which are uncertain are valued at the probabilistic expectation (also known as the best estimate).

Suppose that the simple contract was recognised by the issuance of a bond. Once the contract is agreed, it does not change. If rates fall, the beneficiary is not asked to pay more and the obligor cannot pay less. The bond may change hands in a market at a premium or at a discount but the obligor is not a party to those transactions.

Of course, there are exceptions to this, notably with derivatives. Margin calls may arise as a reflection of the market value of a contract, or more precisely as a result and mitigant of the credit risk arising from changing market conditions. These instruments may be long in duration, but they are fundamentally short-term in nature. The contract is not finalised by the initial payment, but varies with market conditions. This is simply anathema in the context of DB pensions.

This contractual accrual rate is clearly the rate which pension liabilities should be valued. It has not been used to date. It is possible to reconstruct it from historical records: I have done so for three schemes and failed for six others due to the paucity of scheme records. Notwithstanding that, there are approximations which may easily be utilised and which will converge to this rate with the passage of time.

Responses to comments on earlier blog –

Regrettably I could not make it along to the playpen lunch, due to prior commitments. I am informed that there was an emergent consensus that the discount rate applied to DB pension liabilities should vary with the quality of the sponsor covenant. From the accounts I have received, it is not clear whether this was meant to be a higher or lower rate. In financial markets, it is customary and correct to demand a higher yield to compensate for the possibility of default. That would imply lower values for the pension liabilities. It would also not reflect the fact that the higher yield of markets is a diversifiable risk and priced on that basis. If we are to apply a lower discount rate, then for consistency we should also accept lower benefits terms – perverse at best – a subsidy to the employer.

In private correspondence, a number of other points were raised. The method was described as retrospective – it is in fact time-continuous which means that it applies both retrospectively and prospectively. This is the rate underwritten by the sponsor. Neither of the current methods possesses this continuity property; no exogenously chosen discount rate could (unless it coincides with this rate, a vanishingly small likelihood.) Funding based upon such exogenous and discontinuous measures will add redundant costs to the cost of production of the promise.

A further suggestion: the amount payable to some third party for transfer – the problem here is that this is a new business valuation, not valuation of the business in force. Of course, such transfer is precisely the concern of the PPF and a source of conflict for the Regulator. For an ongoing scheme, it is no more than an interesting counterfactual. Transfer of liabilities is an interesting case in that unlike the asset, the bond earlier, all three parties are involved. Clearly the sponsor and the entity assuming the liabilities, but also the beneficiaries as they have an interest in the acceptability of the standing of the assuming body.

Another suggestion: “an estimate of the level of assets needed now based on the returns expected in future allowing for the assets to be held in future. Ignoring “prudence” this is scheme funding”. The problem here is that this scheme investment return rate is not the rate underwritten by the sponsor. Indeed, this rate might be applicable to a standalone scheme, but shortfalls would need either capitalisation or recourse to the sponsor. We have discussed it at length in earlier publications – it could and usually would raise the sponsor cost above that underwritten with no ultimate benefit to scheme members. There are also issues arising with residual orphan assets.

Moving to the comments posted on this blogsite

Mike Ohtsuka

The problem with the idea that scheme actuaries might set the expected rate of return on assets at the level of the contracted rate that I advocate is that for many schemes this rate is rather high – in one scheme that I advise there are still liabilities outstanding where the rate of return exceeded 10% and the weighted average is close to 7%. It is difficult to justify using a 7% expected rate of return in current conditions. It would return an accurate valuation of the liabilities but fail to capture the ongoing reliance on the sponsor. I do not believe that in general it is a good idea to distort or misrepresent anything.

Stephen Tiley

If only I could think about the unknown unknowns … though I wonder, might I not need a closed mind to do that. Incidentally, I first came across that complete aphorism as a needlework Arabic wall-hanging in the office of a Crown Agent in Bandar Abbas in 1973. What you say about the changing nature and cost of DB promises is undoubtedly true. We cover many of these issues in our response to the DWP DB Green Paper and one new one, which is almost as large as all of the old put together, transfer values. This is an article which will appear as a Playpen blog after it has appeared in my column in Portfolio Institutional.

Bob Compton

My blog on USS and the earlier scribbles here, above and earlier, should answer the points raised.

About henry tapper

Founder of the Pension PlayPen, Director of First Actuarial, partner of Stella, father of Olly . I am the Pension Plowman
This entry was posted in pensions and tagged , , , , . Bookmark the permalink.

One Response to “He wouldn’t let it lie” – Con Keating returns to discount rates!

  1. Thanks Con for your reply to my comment. I see now that I didn’t state my proposal clearly. So I reproduce my comment below with the addition of “*THE VALUE OF THE ASSETS IS EQUAL TO THE DISCOUNTED VALUE OF THE LIABILITIES WHEN*” to clarify. Would you still reject what I say, even so-clarified?

    Reproduction of my earlier comment with added clarification:

    Here’s how a scheme actuary could implement Con’s approach to the valuation in a manner that satisfies current pensions regulations:

    If the scheme is 100% funded on a best estimate of the return on the assets in the scheme (where this best estimate is the weighted average of the best estimates of the internal rates of return on the various asset classes), this best estimate of the return on the assets will be equal to the rate that Con identifies. Since that latter rate is the rate of return on their contributions that members have been promised, let’s call it the ‘promised rate’.

    As Con mentions, “The likelihood of this [promised] rate being achieved is a measure of the extent to which the scheme may have to rely upon the sponsor.”

    If a scheme is 100% funded on the basis of a best estimate of the return on the assets, there is a 50% chance that this promised rate will be achieved.

    The regulations require that the discount rate “used in setting technical provisions must be chosen prudently”. We might interpret prudence as a higher than 50% likelihood that the promised rate will be achieved. How much higher than 50% this must be will depend on the strength of the covenant. Suppose that the required likelihood is 67%. Then the scheme will be fully funded on a technical provisions basis if *THE VALUE OF THE ASSETS IS EQUAL TO THE DISCOUNTED VALUE OF THE LIABILITIES WHEN* a discount rate is set as a rate of return which it is judged that the assets in the scheme have a 67% chance of achieving.

    Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s