Index linked gilts as ‘power assets’
Index linked gilts (ILGs) are different from conventional gilts or any other fixed income bond because the British Government promises to uplift both the principal amount (the ‘par value’) and the interest amount (the ‘coupon’) by inflation (using the RPI). The only other inflation-protected securities with a government guarantee to date have been 3 and 5-year certificates issued by National Savings & Investments (NS&I). ILG maturities are spread out as far as 2055. This government-backed inflation-proofing makes them unique building blocks in a private investor’s portfolio of financial assets, as they perfectly match, or ‘hedge’, a liability or cash need in the future that the investor thinks of as being defined in terms of purchasing power, whatever happens to inflation. Most individual financial objectives should be defined this way, in ‘real terms’, to avoid all the different forms of ‘money illusion’ that badly distort personal financial decision making.
This is rarely the way ILGs are actually used in private client portfolios, even when managed by highly-qualified professionals. This is a testimony to the ubiquitous hold of money illusion but also to the intellectual arrogance of most professional money managers in the face of investment uncertainty. Arrogance is a form of behaviour conducive to higher fee generation when clients believe market risks can be successfully exploited for gain by skilled agents. But humility is more likely to be conducive to higher return generation when market risk is not easily exploitable or costs of trying are high. Both money illusion and skill illusion explain why so little use is made of ILGs as unique ‘power assets’, as distinct from just another asset class to throw into the diversification pot.
Risk free or just another bet?
This debate, which rarely reaches the wide audience it deserves, burst into the open in the columns of the FT this month. The American finance professor Jeremy Siegel is well known as the author of ‘Stocks for the Long Run’, a book (now in its 4th edition) which has helped establish in both professionals’ and the public’s mind the essential characteristics of different asset classes that make them useful for investment objectives.
Professor Siegel wrote a guest column for the FT titled‘Inflation-linked bonds face a headwind of many risks’ in which he nodded towards the special benefit of ILGs, particularly in pension plans, but also listed reasons why they were over-priced (both in the UK and their equivalents in the USA) with real yields of barely 1% (‘levels that would have been unimaginable just a few years ago’). He concluded by warning of large losses on the horizon for holders of such bonds as economic activity and inflation both picked up. Equities, he felt, offered ‘much better long term protection against inflation than today’s low-yielding inflation-linked bonds’.
Outrage from the ‘power asset’ camp! It provoked a letter signed by another American finance professor, Zvi Bodie, three frequently-published consulting actuaries familiar to UK pension funds and ex-corporate finance head turned pension consultant, John Ralfe. They argued: ‘The conservative pension saver, especially those with little or no other capital, should avoid the worst outcomes in retirement by holding inflation-protected bonds, even if it means giving up the possibility of the best outcomes by holding equities.’
They also argued that the risk of equities is typically misrepresented by the investment industry, which (either out of mischief or ignorance) understates the risks and overstates the likely payoffs when investing in equities.
Much as I welcome the argument, as it helps to banish popular illusions about risk and the management of risk, I have to say there are fallacies on both sides of this debate, as well represented in these two FT pieces. The faults lie partly in the assumptions that lie behind the maths and partly in a disconnect with personal risk preferences in the real world. In this Insight we will try to find the firm ground between these two camps and show how it can be used as a foundation for individual portfolio choices.
I am not particularly concerned with a less meaningful aspect of the debate which is that there will always be arbitrageurs in the market and investors with a particular opinion about future inflation who will trade between ILGs and conventional gilts as if both were risky assets. On a short term view, when the position taken is not matched to or defined by the maturity of the bond or the duration of the investor’s objective, they clearly both are bets, not hedges. If this is largely what Professor Siegel was referring to, he is right.
Assuming a non-random equity ‘system’
The sort of historical evidence in support of equity investing made famous by Professor Siegel is broadly supportive of the view that equity returns are a fairly intuitive output of adaptive capitalist systems. They do broadly what you would expect as an economist, if you only knew that stock markets recorded the performance of publicly-traded companies in some form of ‘market economy’. However, it is important to distinguish between nominal returns, which are distorted by each country’s and each period’s inflation experience, and real returns, deflated by a national retail price index. Inflation is one of the externalities we expect market systems to adapt to, more or less well.
Interpreting behaviour from real returns, without separating between income and capital, it makes surprisingly little difference if the market economy is Anglo-Saxon, Swedish or Japanese. There is in fact (yet) no theory of why millions of individual decisions by real investors based on idiosyncratic views of personal welfare leads to either similar trends in equity ‘total’ returns (both income and capital) or why the very large deviations from trends that emerge in all the data histories are both bounded and also similar between countries, regions, different stages of economic development and different periods of history. There are many sources of differences, including how they behave together, at the same time, but these may contain less useful information (or none at all) compared with the similarities.
The absence of a theoretical explanation of sustainable trends and of reversion to the trend is a concern for many and a particular concern for academics. But even if there was a theoretical explanation, it would rely on the same data as the only practical basis of experiment. If you have little choice but to participate in the system, you might as well leave proof to others and work out for yourself what view of the world and the system you are willing to hang your hat on. I decided a long time ago that, with respect to real total returns from equities over long periods, it was not a random world.
If the system were random
This preamble is important because many of the critics of the representation of equity risk believe it is in fact a random world. These letter writers, for instance, argue that ‘the proper measure of equity risk is the cost of buying insurance against underperformance versus the risk-free return – a “put option” on a stock market index. If risk reduces over time the cost of equity put options against any shortfall should reduce. But the cost increases the longer the option period, reflecting increasing not decreasing risk. The theoretical price, based on a standard option pricing model and actual prices charged by banks, is about 25 per cent for 10 years and 30 per cent for 20 years.’
The mathematical truth in this statement, that risk expands, not shrinks, with time conceals an assumption that it expands at the rate of a random time series (the square root of time) rather than at the slower rate of a mean-reverting series. The ‘theoretical price’ of the risk derived from option markets is not a proff of anything because it extrapolates from short-period nominal return behaviour which is (probably) random. There is no market in 25 year equity options and, if there were, arbitrage from the physical equity market would cause it to reflect the same assumptions of mean reversion widely made by participants in that market.
ILGs and mean-reverting equity returns
Hanging your hat on a sustainable but uncertain trend in long-term real returns also means that you will assume that the mean expected equity return will rise faster than any risk free rate, such as the 1% real yield of ILGs. The insight that risk rises with time means the band of probable outcomes around the trend increases slightly for every year the investment is expected to be held. (This will still translate into a declining standard deviation of short-period returns the longer the expected holding period, which was the source of the widespread error that risk falls with time, as it does not compound each single-period standard deviation over the relevant length of time.)
The concept of a rising and expanding range of possible real wealth outcomes around a known risk free rate with a lower slope is expressed in the diagram below. The principle to which the letter refers, of both lifting the floor or worst-case outcome but at the cost of a lower slope for wealth generation, is illustrated here by substituting half the equity portfolio by a risk free ILG. Depending on the values assumed, there is a point at which the distribution of outcomes lies entirely above the risk free rate. At any point, you can adjust the level of substitution of equity risk by ILGs to achieve a tolerable worst-case outcome. You would find the same general principles explained in Professor Bodie’s widely-used text book ‘Investments’.
Choosing between risk free and equity bets
The FT letter writers express the opinion that ‘conservative’ investors, worried about the consequences for living standards of a bad outcome, should not make an equity bet at all but accept the certain but very low real return and outcome of ILGs. This is a massive over-simplification. If it has a paternalistic purpose, it is thoroughly misguided.
The question investors at any level of wealth, whatever their instincts about risk, need to ask is ‘how much risk should I take, given the consequences of bad outcomes and the floor below which these become personally intolerable, and what will that cost me in resources?’
If, like other reluctant participants in the capitalist system, I want a certain standard of living but do not want to bear much risk to obtain it, I think I would like to know whether I could in fact afford to avoid risk. If the implications of building more certainty into my life was a mediocre living standard, either by making spending sacrifices whist still earning or in retirement (or even both), I might accept a bit of a capitalist gamble to lift most of my probable outcomes above the mediocre.
If, on the other hand, I were extremely wealthy and had no children, and was not hard-wired to keep creating wealth (as many enthusiastic participants in the capitalist system are), I can imagine I might decide to stop taking risk, or would only take enough to secure improvements in my spending power that I was likely to value. I have no need to gamble and no satisfaction from gambling. I can leave the table and live on gilt coupons (index linked, of course).
These practical examples of risk taking preferences, in a real-world context with particular personal consequences, illustrate an important role of ILGs in both financial planning and investment management as the market benchmark of the cost of avoiding risk, in both resources and outcomes.
Choices at retirement
The benchmark for avoiding all risk at retirement is the income provided by an index linked annuity, as this deals with longevity risk as well as capital market and inflation risks. Using a retirement plan as an example:
On the basis of current real yields, a capital sum of £1,000,000 will produce an annuity income before tax (for a 60 year old male with two-thirds spouse benefit) of £25,250pa. This illustration ignores the practicality that, if it were in a pension plan, you would apply the tax free cash to a purchased life annuity which unfortunately does not come with full indexation. Being an annuity the real payments will be level each year, even though you might prefer higher spending targets early in retirement and less later, but with equivalent real outcome certainty.
With controlled risk taking, in current equity market conditions, the same sum might generate a profile of preferred spending starting with a gross equivalent draw of of £40,000 pa at the start of retirement. Depending on the payoffs from risk taking and how quickly or slowly they emerge, the starting rate might increase to a mean expected rate of £90,000 pa in real terms, sustainable to age 95. On worst-case real investment returns, the rate of draw might taper, in line with the time preferences, from age 75 to about £26,000, so never less than the level real annuity.
At retirement, it would be illogical to hold ILGs and draw down from them instead of buying an inflation-indexed annuity. More of the resource would need to be assigned to funding the extra years of life beyond the actuarial expectancy of the pool of lives in the annuity fund – the illustrations above assume 35 years instead of about 27 years. To match the same worst-case gross draw as above holding only maturity-matched ILGs would require £400,000 more capital and to match the median draw would require a further £600,000.
No free lunch
An important insight of the FT letter writers is that equity investing is typically presented as a free lunch. They say: ‘The higher expected return of equities over inflation-protected bonds is simply a reward for the risk of holding equities; it is not a “free lunch” or a “loyalty bonus” for long-term investors.’
Provided the trade off between resources, certainty of outcome and risk taking is properly explained and quantified, there is no illusion involved in making a case for equity investing. On the contrary, it avoids the illusion that avoiding risk is cost free and affordable. In the example above, the actual rate of sustainable draw is consistent with 99% confidence of not breaching the floor, there is no spending of payoffs from risk taking before they have been earned and the cost of avoiding different forms of risk is explicit.
ILGs perform an invaluable and indispensible role both in communicating the principles and quantifying the trade offs. They should feature in every financial plan and every portfolio review. But they should also be the main method of risk control in the portfolio itself.