Clacher and Keating
In this article Con Keating and Iain Clacher explain an alternative to the current way we require DB schemes to be funded. It challenges received thinking and offers a way forward to regulators struggling to find an acceptable funding methodology to both trustees and sponsors. A must read
Here is Derek Benstead’s comment as a summary
As discount rates have been progressively reduced over the past 20 years, the funding target set in actuarial valuations has been set progressively higher. Pension schemes don’t have a problem of persistent deficits. The improvement in funding achieved over the years has been hidden because the funding goal posts have been moved further away at each valuation.
The major advantage of the method advocated here is the fixing of the goal posts. The contribution to benefit accrual implies a discount rate which values the benefit awarded at the contribution paid. The premise is a simple one. These are the terms on which the benefit was awarded and contributed to, so these are the terms on which we should judge progress since then.
Had we used this method down the years, we would have a better understanding of how pension schemes have actually fared down the years since the original funding plans were made.
Contractual Accrual Rates – A Practical Illustration
In various documents we have described the contractual accrual rate (CAR) of a DB pension award as that rate of return which equates the contribution made with the projected benefits payable under that award. The contribution is the scheme asset and the projected benefits the liability. The contractual accrual rate of a scheme is the weighted average over time and members of these rates.
The CAR is both the correct rate at which pension awards should accrue or equivalently be discounted and is the rate of return on the scheme assets necessary to meet the liabilities on time and in full with a UK DB scheme, and this is guaranteed by the sponsor employer.
To estimate the CAR of a scheme using the historic records of contributions and awards would be a complex and tedious exercise, and for many schemes would not be feasible given the quality of these records. However, we may exploit the return on assets property to establish the current CAR of a scheme. If we take the current market value of assets and the associated projected benefits, we may establish the rate of return on those assets needed to discharge the liabilities; this is the CAR of the scheme, at the current time and going forward.
We shall take an illustrative open scheme as our pedagogic example. This has assets of £25,853,771, which we shall consider as our contribution proxy and projected liabilities totalling £ 67,181,556, which are distributed over the ensuing 70 years as illustrated in figure1. The CAR is 6.1%.
Although this scheme is open, we consider first the situation with no new awards in the first year. We show, first, the development of scheme liabilities at Table 1. There are no revisions to the projected benefits in this illustration.
Table 1
| Amount (£s) | Note | |
| Opening Liabilities | 25,853,771 | |
| Accrual | 1,577,080 | at 6.1% |
| Pensions Paid | – 1,537,896 | |
| Closing Liabilities | 25,892,955 |
The accrual is the increase in the present value of liabilities, at the CAR rate, due to the passage of time. Next, we consider the income and expense position and the evolution of assets as Table 2. We introduce the asset portfolio income (3.4%), from dividends and bond coupons received, as well as the mark to market gain in asset prices. We see that the scheme is cash flow negative, relying on the sale of assets to pay pensions. This would be the position if the scheme were closed.
Table 2
| Amount (£s) | Note | |
| Dividend Income | 879,028 | at 3.4% |
| Pensions Paid | – 1,537,896 | |
| Net Operating | – 658,868 | |
| Opening Assets | 25,853,771 | |
| Gain / Loss | 491,222 | 1.9% Mark To Market |
| Net Operating | – 658,868 | |
| Closing | 25,686,124 |
The solvency position is shown in Table 3. Unsurprisingly, there is a deficit as the asset performance (3.9%+1.9% = 5.3%) is less than the contractual accrual rate of 6.1%.
Table 3
| Amount (£s) | |
| Assets | 25,686,124 |
| Liabilities | 25,892,955 |
| Solvency | 99.20% |
The scheme was, in fact, open to new members and future accrual. The stand-alone characteristics of the new awards are shown in Table 4. The assumptions driving the projected values for benefits are the same as those used for the historic scheme.
Table 4
| Amount (£s) | |
| Contributions | 733,296 |
| CAR | 4.20% |
| Benefits Added | 2,706,706 |
The lower than historic CAR on the new contributions and liabilities added will reduce the scheme CAR marginally. Table 5 presents the asset position.
Table 5
| Amount (£s) | |
| Dividend Income | 879,028 |
| Pensions Paid | – 1,537,896 |
| Contributions | 733,296 |
| Net Operating | 74,428 |
| Opening Assets | 25,853,771 |
| Gain/Loss | 491,222 |
| Net Operating | 74,428 |
| Closing Assets | 26,419,420 |
Next, we consider the liabilities scheme as a whole, including the new awards in Table 6, and Table 7 shows the solvency position of the scheme.
Table 6
| Amount | Note | |
| Opening Liabilities | 25,853,771 | |
| Accrual | 1,577,080 | at 6.1% |
| Pensions Paid | – 1,537,896 | |
| New Liabilities | 733,296 | |
| Closing Liabilities | 26,626,250 | |
| New CAR | 6.03% |
Table 7
| Amount (£s) | |
| Assets | 26,419,420 |
| Liabilities | 26,626,250 |
| Solvency | 99.22% |
We now consider a further year in which contributions were made and new liabilities added. The contributions and liabilities added are shown in Table 8.
Table 8
| Amount | |
| Contributions | 782,182 |
| CAR | 4.19% |
| Benefits Added | 2,873,610 |
Comparison of these statistics with the earlier Table 4 shows a substantial increase in both liabilities and contributions, though the new awards CAR is almost unchanged. The cause of this was higher than expected salaries for new recruits and greater than predicted increases for existing actives. This led to a decision to revise the assumptions for the existing benefits to be consistent with those applying to the new awards. We shall return to this later, but first will address the income, expense and asset position, as Table 9.
Table 9
| Amount (£s) | Notes | |
| Dividend Income | 924,680 | at 3.5% |
| Pensions Paid | – 1,475,281 | |
| Contributions | 782,182 | |
| Net Operating | 231,581 | |
| Opening Assets | 26,419,420 | |
| Gain/Loss | 1,400,229 | 5.3% MTM |
| Net Operating | 231,581 | |
| Closing Assets | 28,051,230 |
The revaluation of projected benefits shows these to have a total extra cost of £ 861,694 and the CAR of the scheme rises to 6.13%. Table 10 shows the liability position of the scheme and the solvency position of the scheme is shown in Table 11.
Table 10
| Amount | Note | |
| Opening Liabilities | 26,587,066 | |
| Accrual | 1,629,787 | at 6.13% |
| Pensions Paid | – 1,475,281 | |
| New Liabilities | 782,182 | |
| Closing Liabilities | 27,523,754 | |
| New CAR | 6.08% |
Table 11
| Amount | |
| Assets | 28,051,230 |
| Liabilities | 27,523,754 |
| Solvency | 101.9% |
This illustration has shown how the contributions-based CAR may be proxied by the required return on assets and shown it in practice for a closed scheme, an open scheme, and an open scheme with revisions to the projected benefits.
These illustrations also show the low natural variability of the correct discount or accrual rate. The changes which do occur all arise from real world changes to the benefits offered by the scheme. They are not changes in liability present values arising from arbitrary changes in the discount rate.
As discount rates have been progressively reduced over the past 20 years, the funding target set in actuarial valuations has been set progressively higher. Pension schemes don’t have a problem of persistent deficits. The improvement in funding achieved over the years has been hidden because the funding goal posts have been moved further away at each valuation.
The major advantage of the method advocated here is the fixing of the goal posts. The contribution to benefit accrual implies a discount rate which values the benefit awarded at the contribution paid. The premise is a simple one. These are the terms on which the benefit was awarded and contributed to, so these are the terms on which we should judge progress since then.
Had we used this method down the years, we would have a better understanding of how pension schemes have actually fared down the years since the original funding plans were made.
Thanks for that very clear synopsis of the article which will appear as an introduction in future!
“In this article Con Keating and Iain Clacher explain an alternative to the current way we require DB schemes to be funded. It challenges received thinking and offers a way forward to regulators struggling to find an acceptable funding methodology to both trustees and sponsors. A must read.”
Does this only apply to open DB Schemes or does it also apply to DB Schemes that are closed to future accrual e.g The British Steel Pension Scheme (BSPS2)?
The only problem is that it does not work. Where do you get dividends and coupons at 3.5% per annum? 10 years ago probably, but not now. Dividends are cut left, right, and center.
This is a total return game, and you cannot use a return higher 3% – 3.5% per annum for the next 10 years. You may not even get that one.
We have this unfinished business with the USS, where members would like to increase the discount rate instead of pay higher contribution into the scheme. This reminds me of a video on youtube where the Illinois Governor and Chicago mayor did negotiate the discount rate for the Chicago Metropolitan workers. That scheme ended in disaster, tens of thousand of people lost their pensions.
Robert The first example with no new awards would be a closed scheme. So yes it can be applied to both
Thanks Con,
In that case, as ‘Table 1’ shows the ‘Closing Liabilities’ being higher than the ‘Opening Liabilities’, could DB schemes such as BSPS2 consider this example to keep the scheme running instead of going down the buy-out route?
Eugen.
I think you are missing the point. We are demonstrating the method not making any claims as to what assumptions should be made. The figures used inte illustration are those of a real scheme 5 or 6 years ago.
I will take the opportunity to let you know that there is no meaningful relation between gilt yields and equity returns. Not simultaneously, not one year forward, now ten years subsequent nor 30 years, nor any of the annual intervening terms. And if you really want to think about low gilt yields a linear regression intercept at zero of 8%.-this is for the ten year subsequent returns.
Con
Robert Yes of course BSPS2 could run of the liabilities, The problem with the new funding code is that it will require insanely high levels of coverage – self-sufficiency is just code for buy-out level funding.
Con,
I have faith in the Trustees of BSPS2 but I hope they read these blogs which concern DB Schemes?
There are many members (including myself) who would like the Scheme to continue if it was safe to do so?
Eugen
I will just add one further comment here, The 6.1%CAR is not an expected return on assets. It is not a choice or control variable, It is a fact. There is a CAR in here which is related to expected returns and that is the new award CAR and that is 4.2%
Con
Dear Con
The only thing that is important is that trustees use conservative assumptions to protect members’ benefits.
I do not like is what happens now with the USS, and I think the next one would be the Railways pension scheme, where Unions and some members started negotiating the discount rate. In the U.S. quite a few pension schemes failed because of that.
Contributions are the result of mathematical calculations based on life expectancy and expected investment return. They could be or not be affordable, but they are what they are, they are not subject to negotiation. Subject to negotiation could be a reduction of future accrual, change the retirement date to a later date, of if the law is changed And allows for it – transformation to a defined ambition scheme, giving up the defined element to some extent.
Some of the Universities and especially the Railways companies are weak sponsoring employers. This pandemic will expose further how weak these employers are. We need to be conscious of that.
Last thing, we cannot defeat mathematics and statistics with Lords’ amendments.