This blog looks at how index-linked gilts came to be and asks what they are doing in LDI portfolios – its authors are Jon Spain and Con Keating
In 1980, the Wilson Report was published, recommending that the UK government should issue index-linked gilts for pension funds. Announced in the Budget on 10 March 1981, the first issue was for £1 billion at par on 27 March 1981. Initially restricted to approved private sector pension schemes, ownership was relaxed a year later. Within a year, 4 index-linked stocks had been issued, three (including the very first) bearing a coupon of 2.0 % with one bearing a coupon of 2.5 %.
Since then, using the over 15-year index data available, real yields (relative to RPI) on long index-linked gilts have fluctuated enormously, reaching as high as 4.5% (Dec 1991) and as low as -2.6% (Nov 2021). Why have they been so strongly recommended by actuaries and TPR?
Initially touted as ideal for pension schemes with future cashflows tied to RPI movements, little thought seems to have been paid to an ideal structure for the intended buyers. For one thing, few private sector pension rights are even fully linked to inflation, with different limits applied to different pension tranches, suggesting that ILGs provide inflation protection which may be unneeded. Also, for substantial periods, the protection offered by linkers was below inflation.
Until 2030, the payments receivable on ILGs will be indexed in line with RPI, with CPIH to be used after that (see ukrpi.com). The ILGs have an inflexible structure. If trustees are looking for inflation-proofed income in order to support pensions in payment, then they must buy an extremely large capital sum. Alternatively, if they really want inflation-proofed capital, for reasons we don’t understand, then they must also buy driblets of income.
Maybe ILG investment returns have been so good as to offset these disadvantages?
Well, that is not really the case. Over periods of 15 years, the average annual return was 8.18% for long index-linked gilts against 7.91% on long conventional gilts, a slightly better performance by 0.27%. During 2022, both types of long gilts returned very high negative returns, being 46.92% (index-linked) and 40.05% (conventionals). So, as well as being inflexible and not aiming at producing the investment cashflows likely to be useful to private sector pension scheme trustees and sponsors, this asset class’s performance has not been persuasively attractive.
One aspect of the attraction of ILGs has been that TPR has very strongly encouraged actuaries and trustees to use index-linked gilt inputs for actuarial valuation assumptions, and this may well have fuelled the demand rather than any higher returns fitting the trustees’ actual cashflow and funding requirements.
As indicated above, the first four index-gilts were issued with coupons of around 2% (one a little higher). Since then, looking at the index-linked gilts in issue as at the end of 2022, they have been issued with coupons ranging between 0.125% (virtually zero) and around 4%. This emphasis on coupons is important in several respects, one of which is their role in determining the relative modified durations of these securities, namely their relative volatilities. For example, at the time of writing, the modified duration of the 1.125% conventional due 22/10/73 was 29.7 years, while the 0.125% index linked due 22/3/73 had a modified duration of 48.2 years. The ILG volatility in this case is expected to be 1.6 times that of the conventional.
The modified duration of a typical scheme was 18.5 years. The passage of one year, with rates unchanged, will leave the modified duration of this scheme as 18.2 years while the ILG will have a modified duration of 47.2 years. These differential rates of change carry important consequences if trying also to hedge discount rate sensitivities, and it is evident from the results now being reported by schemes, that many were, in effect, materially over-hedged.
It should be noted that no one seems to have a clear picture of where the ILGs are actually held now. In June 2016, Schroders estimated that UK private sector defined benefit schemes already owned an estimated 80% of the long-dated index-linked gilt market, assessing their potential demand at almost five times the size of the market. At year-end 2022, ILGs had a total market value of £561 billion, 29% of gilts outstanding. Direct holdings by defined benefit pension schemes at September 2022, as reported by the ONS, were £264 billion, 43% of the total outstanding. Of course, DB pension schemes had much further exposure to ILGs through pooled funds, which could plausibly have doubled the sector’s exposure to ILGs. Notwithstanding this uncertainty, it clear that ownership of ILGs is highly concentrated in DB pension funds.
Concentrated ownership of a security results in extreme price behaviour in terms of both highs and lows; an exaggerated form of the risk on / risk off behaviour. This was particularly evident in the year ended October 2022, when the ultra-long linkers fell by more than 85% in price, over 10% more than conventionals. This behaviour may be illustrated by considering the rolling 12-month correlations of returns, as shown in the chart below:
The conclusion to be drawn from this is that, when correlations are high (close to 1.00), ILGs may from time to time be a good surrogate for conventional gilts, but there are also substantial periods of time when they are not.
This may also be seen in a regression of these conventional and linker indices. While much of the regression coefficient can be explained by the differing relative durations, not all can be. The low R squared, 62%, indicates that ILGs are a very imperfect hedge of conventionals – which poses some acute challenges for LDI strategies using ILGs.
When LDI struck in September 2022, gilts came under pressure from the margin calls of leveraged positions. The Bank of England intervention was initially limited to purchases of conventional gilts, which is consistent with their actions in quantitative easing which was limited to conventional securities. The decision to expand the facility to include ILGs can only be interpreted as being driven by a desire to preserve a viable ILG market and that channel of government financing. There really are no substantial arguments for the systemic importance of ILGs.
Given all of the above, we are left with a question: Why have actuaries and TPR so strongly favoured ILGs for so long above all other assets?
