We have just updated www.ukrpi.com, with the statistics and various charts updated until the end of 2023. Just over a year ago, Henry Tapper kindly published our short piece (link below), in which we wondered why ILGs have been so strongly recommended by actuaries and TPR.
Apart from including a short piece on LDI, we have added a little more material about why index-linked gilts seem to have had no real economic purpose. We are not aware of any long-term institutional investor ever demanding index-linked capital. So we have done some simple sums, showing how inefficient ILGs have been at matching income.
Specifically, we have tried fairly pricing a 15 years index-linked (RPI) annuity-certain for years starting between end-1985 and end-2008 (there are 24 periods until end-2023). The price is based upon the initial ILG yield, with no expense or profits allowances made. Allowing for returns and indexed payments of £1,000 pa, if fairly priced, the fund at the end should be zero.
The returns are those on the long ILGs total returns index because the income generated is actually even less important than when they were first issued. Over the 24 periods, the mean experienced RPI increase was 3.10% pa, with standard deviation of 0.41% pa, a minimum of 2.71% pa and a maximum of 3.97% pa. The end-year ILG yields varied between 0.82% pa and 4.49% pa (see chart A below).
As an example, taking the 15 years from end-1985 until end-2000, the initial yield was 3.87% pa, giving an initial price of £11,217. After 15 years, the residual fund is £1,097 which is not what we wanted. To obtain the right outcome, the yield needs to be increased by 0.38% pa to 4.25% pa, giving an amended price of £10,923.
The price reduction of 2.61% for 1985-2000 is somewhat modest, averaging 13.72% over the 24 periods and being as (absolutely) high as 27.61% for 2006-2021. The figures for all periods are tabled (C) and the yields (A) and prices (B) are charted below. On average, over all 24 periods, in order to achieve fair pricing, the initial yield had to be increased by 2.14% from 2.77% to 4.91%, with the price being reduced by 13.72% from £12,202 to £10,467.
This hardly suggests that using ILG yields would have been efficient. Taking these results into account, we can discern even less evidence that ILGs should be taken as systemically important. Why were they so strongly recommended by actuaries and TPR?
Con Keating & Jon Spain 25 Jun 2024
Chart A (Long ILG Yields)
In the chart below, “YieldEst” refers to the original yield, “YieldAct” refers to the required yield and “YieldFix” shows the required adjustment. The table shows the mean, standard deviation, minimum and maximum over the 24 periods.
Chart B (Prices)
In the chart below, “PriceEst” refers to the original price, “PriceAct” refers to the required price and “PriceFix” shows the percentage reduction required, tied to the RHS scale. The table shows the mean, standard deviation, minimum and maximum over the 24 periods.
Table C (Full Results)
Henry, Con & Jon’s article demonstrates very clearly the folly of following the herd. ILG’s like any investment are subject to the law of supply and demand. When supply is limited in respect of the demand prices are high. A real return can only be achieved by selling at a higher price than that paid. The charts demonstrate by how much Trustees have been over paying, if ILG’s were held to redemption. Any Trustee that has bought ILG’s at a premium could be deemed to be breeching their fiduciary duty, as they are guaranteed a loss, unless they are able to find a buyer (mug) willing to pay more. South sea bubbles and parables come to mind, when dogma overide common sense.
Apologies for an aside to some excellent challenges but I’d like to ask the question – “What is the difference between index linked gilts and our other national debt for £1.5tn of unfunded public sector pension promises? The later are also index linked, but have longer duration, higher coupons and 5m+ votes attached. Servicing both crucially depends on GDP growth.
ILGs are particularly inappropriate for cash flow negative closed DB pension schemes due to the low coupon rate (0.625%) increasing the dependency on short-term capital asset sales proceeds. Since 2008 the price volatility of gilts has far exceeded the price volatility of a widely spread equity portfolio and has resulted in highly volatile funding reports with little relevance to actual scheme funding strength. That is of course unless the pension scheme (mistakenly) tried to match the funding assumptions.
The argument that matching funding assumptions mirrors buy-out costs falls down on two grounds:
It assumes the insurance company also matches the future liabilities using ILGs which does not appear to be the practice.
Secondly it is now recognised that buy-out is not the only end game available with run on and consolidators offering non eventual buy-out solutions now very much on the cards.
If pension schemes (and insurers) do decide to reduce their holdings in ILGs (and especially relatively to fixed interest gilts of similar duration) the valuation assumption of future inflation will rise. The fact that the derived inflation assumption has remained relatively constant (never outside a range of 3% – 4% RPI since 2006 for the 20 year duration) despite the widely differing inflation experiences suggests to me that the market makers (pension schemes) apply a pre-determined inflation discount to the price of the fixed interest equivalent without reference to actual inflation expectations in determining the price they will pay for ILGs. Will the inflation assumption sky-rocket if there is a substantial sell off of ILGs?
All this suggests that pension schemes have been particularly badly served by a regulatory regime that has failed to recognise that what may have been prudent in 2004 had become anything but prudent in subsequent market conditions. We need a fundamental reappraisal of the approach to measuring the financial strength of pension schemes.
I have previously suggested that managing against run on cash flow forecasts is the best solution and should be the subject of regulatory oversight, but we do need a fundamental review.
While it’s interesting to look at returns over a long period from 1985 to 2008, may I suggest that many trustees and their advisers, both then and now, tend to look at returns over shorter periods?
Comparing 10-year market returns to December 2008, for example, I see the following:
Over 5-year ILGs averaged 5.5% pa
UK equities 1.2% pa
Overseas equities 2.2% pa (with US equity averaging only 0.1% pa)
(best returning equity allocation was Emerging Markets averaging 11.2% pa)
UK property 7.9% pa
Cash 5.1% pa
Over 15-year conventional gilts 5.2% pa
(Unhedged Overseas bonds averaged 7.6% pa)
Over 15-year corporate bonds 4.3% pa
RPI 2.8% pa
Since I suspect many schemes would have lowered their allocation to equities and increased their allocation to bonds during this decade, I’m not persuaded that actuarial consultants favouring ILGs and other bonds look to have given poor advice, with hindsight knowledge at that time). Allocations to UK property and/or Emerging Market equity will have been very small.
In closing, I’m reminded that even George Ross Goobey (so-called father of the-cult-of-the-equity) “apostatised”, when Irredeemables yielded more than 15% in the mid-1970s. And when UK Index-Linked Gilts were first issued (i.e. available to pension schemes), in 1981, they were on a Real Yield of around 2.5% which rose fairly quickly to 3% and above, subsequently exceeding 4%; indeed, as late as 1996, I believe it was still possible to buy them on a Real Yield in excess of 3.75%.
I think we can agree that consultants could and should have changed their advice after 2008, but before then I’m less persuaded.