The transfer of DB liabilities from an employer sponsored scheme to a bulk annuity insurer involves no gain to scheme members – their pensions are unchanged before and after such a transfer. From the perspective of the member, it is a largely neutral transaction.

There seems to have been very little thorough actuarial analysis of these transactions, but the new actuarial standard, TAS 300 v2, looks set to change this. TAS 300 v2 states:

5 Bulk transfers

P5.1. Practitioners providing advice to a governing body or an employer which is considering a bulk transfer must consider the following:

1. credible alternatives to the potential transaction for the long-term provision of members’ benefits. Practitioners must consider whether the following alternatives are credible: a bulk transfer to a superfund or an insurer and retaining the liabilities within the existing pension scheme potentially with additional funding and/or security;

2. any material impact on the protection provided for members’ benefits in the event that the benefits are unable to be paid as intended;

3. any changes in the material risks to the benefits of the different classes of members; and

4. any changes to the governing body’s ability to make decisions which affect the level of members’ benefits.

The motivations for bulk annuity buy-outs have been a desire to reduce the exposure of the sponsor combined with some enhancement in the security of the scheme for members. We consider first the risk transfer aspect.

Risk transfer in buy-out

Let us consider the funding levels disclosed in TPR’s Tranche 17 (T17) analysis. We note that the cost of buy-out has increased somewhat since the period covered by the Tranche 17 returns. The pricing of pensioners in payment has risen from an average of around gilts +30 bp at that time to gilts +10 bp recently. For deferreds there has been a similar movement, from gilts -10 bp to gilts -30 bp. At these recent margins, the blended discount rate for most schemes would be lower than gilt yields.

TPR reports that T17 schemes were 97.2% funded on a technical provisions basis, and were 80.2% funded on a buy-out basis. The buy-out price is 121.2% of technical provisions. Technical provisions are prudent valuations, a 10% – 15% prudent margin above best estimate. This places best estimate funding at between 108% and 114.4%

Bulk annuity insurers are concerned with the best estimate price in their risk management and accounting. They are charging 121.2% of 108% – 114.4% which is between 130.9% and 138.6% of best estimate. The corporate sponsor is paying this price to eliminate risk, which is correctly measured relative to the best estimate (expected) value, so 30.89% to 38.6%.

The insurance regulatory standard insolvency likelihood in 0.5% in a year. The likelihood of the scheme losing the amount of the premium is comparable to an insurance insolvency likelihood. The likelihood of loss for the 10% and 15% prudence levels is shown for two Normal returns distributions N(5,15) and N(5,10).

The likelihoods are shown in the red cells. It is only in the low (10%) prudence and highly volatile (15%) asset portfolio case that the risk to a sponsor is greater when running on, in all other cases, transfer to an insurer is riskier.

Consultants have been advising schemes on the low-risk asset allocations that are preferred by insurers for in-specie transfer of assets in payment of buy-out policies. Given the conservative de-risked nature of these DB scheme funds, it is reasonable to say that there is no risk gain in the transfer.

The Opportunity Forgone

It is worth considering the opportunity forgone by buying out rather than running the scheme on.

Again, using the Tranche 17 DB funding analysis, we have the price of buy-out as 30.9% and 38.6% of the best estimate of liabilities transferred. Obviously, these amounts might alternately have been invested by the scheme or sponsor.

The historic volatility of returns of the ONS FSPS dataset is an annualised 12.5%. This falls to 8.5% if the two worst quarters, Q2 and Q3 2022, are excluded (Q2 -16%, Q3 -12%). The worst performance was the year Q2 2022 to Q2 2023, at -33.4%.

The average life of schemes entering buy-out is 15.8 years. The table below shows the total value of the buy-out premium invested over the remaining lives of these schemes, together with the amount arising from investment of the initial buy-out premium paid. It shows these for a range of normally distributed returns distributions, where the expected investment returns rise from 5% to 8%, and the volatilities rise from 7.5% to 15%. The figures shown are medians of a simulation.

The total value foregone by schemes entering buy-out is extremely high. For £100 million of best estimate liabilities, it varies from £64 million to £110.4 million. Put another way, buy-out raises the cost of discharge of £100 million of liabilities to between £164 million and £210.4 million depending upon the investment strategy which might be pursued.

With scheme liabilities of around £1.2 trillion, these sums are far larger than the total deficit repair contributions made by schemes over the preceding twenty years. They totalled some £300 billion while the current cost of bulk buy-out is between £360 billion and £460 billion, and the gains now forgone lie in the range £750 billion to £1.3 trillion. The cure is far more expensive than the deficit repair contribution ailment.

PPF coverage

It is true that some of the members of schemes falling into the PPF will receive lower benefits than schemes covered by the insurers’ Financial Services Compensation Scheme (FSCS). For T17 schemes the PPF s179 funding ratio was 112.7%, that is PPF payments were 86.2% of the full promised liabilities measured at technical provisions levels. Scheme members would have 13.8% of their promised benefits at risk, but given the level of surplus, many members could be bought out of the PPF on better terms.

The risk to scheme members is actually very small indeed and will decline as more scheme members transition to being pensioners in payment. The risk, as the product of the likelihood of scheme failure, and the amount, varies from the vanishingly small to just 0.1% of benefits promised.

We have also received the following cautionary note, a hypothesis on the sufficiency of the FSCS:

“With regard to the risk transfer annuity – yes the payment to the pensioner is protected under the FSCS.  However, if one of the three large participants in the risk transfer market was to fail with a funding deficit (e.g. a loss of funds to a funded sub-underwriter) it is doubtful the FSCS could secure sufficient funding by levy to allow the transfer of liabilities to other players.

In that situation, I guess the legal basis of a DB pension would be carefully examined.  As reinforced by the Virgin Media case – the DB pension promise is made by the employer at the time of employment/benefit accrual and the employer effectively sub-contracts the payment of that promise to the pension scheme with a guarantee that the employer will fund that scheme to the required level to pay the benefits as they fall due.

That guarantee of the benefits defined at the time of employment cannot subsequently be changed by the employer by Deed amendment or by the transfer of the payment responsibility to an insurer (certainly with a buy-in).  I do believe the Courts might regard this employer’s guarantee as still being in place when a buy-out has occurred.  In the situation of a failure beyond the capacity of the FSCS, rather than providing public funding a la Equitable Life, I do believe the Government of the day would look to the employers and through them to the PPF to provide the necessary support.”

One argument in favour of such a response would be:

“After all the employer has already paid substantial “insurance premiums” for Scheme failure to the PPF!”

It will be interesting to see if finance directors continue to regard their legacy DB schemes in the same negative light once scheme specific analysis of this type is available to them.