Click here to download a pdf version of this paper.

Fans of absolute-return investing claim it represents ‘the future of asset management’, its attackers that it is just a fad. Most investment fads are intellectually lazy: they describe concepts but appeal to emotions. The most appealing of investment concepts are versions of the free lunch: the ‘something for nothing’ culture. For absolute-return products, the claimed USP is that you can generate much more upside return than you risk in absolute loss and that this is more efficient and rational than harvesting long-term risk premiums from volatile assets by accepting large interim losses. In this position paper, we put the claim to the No Monkey Business ‘no nonsense’ test.

Absolute-return investing is faddish. It is not the new industry paradigm. It is mostly not even absolute. But a hybrid form of absolute and relative investing is emerging as a lasting addition to the diversity of approaches, whose mix is ultimately decided by investors (as much because of their experience as theory).

Fund managers compete as if the two approaches are mutually exclusive – either philosophically or economically. They are not. Investors and their advisers do not need to look for inherent superiority. The differences need to be clear enough to choose between them or decide how to combine them. They can do this by relating them either to suitability for different tasks or to their own view of how the world probably works, since this ‘model risk’ is at the heart of the differences. To be this discriminating, investors have to get beyond the sales pitch.

No Monkey Business offers an alternative for each of the best of absolute-return and the best of conventional asset allocation. What makes each the best is a focus on risk management. We reject as suboptimal the vast majority of other industry solutions, both conventional and new.

Scope of the paper

The investment products that are typically marketed under the umbrella of absolute-return investing are

• capital-guaranteed structured products
• ‘target-return’ unit trusts and OEICs and enhanced cash funds
• hedge funds and funds of hedge funds

Whilst the new 130/30 funds use some of the same techniques as hedge funds, they are not intrinsically absolute funds as the net exposure to some benchmark asset is neutral relative to its long-only counterpart.

The first part of the paper, summarising our position, assumes some familiarity with the industry as well as with investment theory.

The second part starts by explaining the context, and the language, for the benefit of individual investors who need that for clarity about different approaches. The arguments are then developed for their benefit – but also for professional investors willing to be challenged.

Our position

Absolute-return investing is not what it appears to be, either in theory or in practice. It is no different from any investment approach that reduces downside risk to tolerable levels by traditional asset-allocation techniques – unless it can live up to its claim that it reduces upside volatility more than downside volatility.

Lower short-period downside risks, and the effects on compounded long-period wealth, are often advanced as reasons for giving up on harvesting conventional risk premiums with a buy and hold strategy (as in ‘relative-return’ investing). The latter describes fairly well the state of the art in the UK retail investment marketplace, where both the return estimates and the risk-management dimensions of asset allocation are poorly developed. This is not surprising since advisors and wealth managers have only recently adopted asset allocation as either a theoretical or practical basis for key decision making.

However, absolute-return investing also has to compete with modern approaches to liability-driven portfolios that are fairly conventional in terms of their asset building blocks but apply better-developed risk-management techniques.

For instance, if the asset-allocation approach is one based on an assumption of ‘mean reversion’ (which harvesting risk premiums logically implies), then a dynamic allocation approach that responds to those changing market conditions and changing expected returns, whilst always observing a defined risk tolerance, will generate much higher risk-adjusted, compounded, wealth outcomes. This is the basis of the No Monkey Business dynamic asset-allocation model.

Adding exposure to more risk sources, and changing the distribution (or weights) can certainly produce more efficient risky portfolios, whether we choose to call these exposures ‘conventional’ or ‘alternative beta’. But that can still be matched by a combination of a less-diversified risky portfolio and a risk free asset, particularly if the risk free asset guarantees a contribution to a targeted wealth outcome (in the way, for instance, that an index linked gilt guarantees a certain real outcome at maturity). We refer to this combination of risky and risk free as dilution, as a deliberate contrast with the reliance on diversification alone as a means of managing risk.

To be superior, the alternative of absolute-return investing has reliably to generate asymmetrical risks, as neither a passive nor a dynamic conventional asset allocation, constrained by the same risk tolerance, can beat that.

Downside risk can be limited in a purely passive way using options, as do capital-guaranteed structured products. Their payoffs are a function of interest rates and volatility, both of which are time-varying but ultimately systematically tied to market efficiency. That means that stringing together a series of short-period structured contracts cannot generate the same wealth outcomes for a given level of risk as unprotected, but diluted, market exposures. Structured products represent opportunities for active managers at times but not an efficient long-period wealth-building strategy – or even a long-period alternative to rolling over cash deposits. This inconvenient fact should become apparent when low interest rates combine with higher volatility. But the full impact, in opportunity cost, will be slow to emerge and too late to reverse.

The claim of absolute-return investing to asymmetrical returns must therefore rely on forms of active management. To be different from benchmark-constrained asset management, these must mostly relate to ‘market timing’. The ability to short securities or swap exposures (new forms of ‘relative bets’) is not enough to transform the shape of the payoffs.

We have about 15 years of hedge-fund data we can check to see whether active management has produced asymmetrical payoffs. Provided we limit ourselves to diversified groups of funds, the evidence is just about supportive – and that even after high (but partially contingent) fees. But it mainly depends on one period: the 2002-3 bear market.

Moreover, the results are not consistent with the avoidance of loss implied by the use of the term ‘absolute returns’. Asymmetrical payoffs to active bets have required much more, regular exposure to volatile assets than is consistent with that claim. Since 2003 (a bull run with relatively short-lived sell-offs), the payoffs look very similar to a diluted conventional risky portfolio with a relatively moderate risk tolerance (but certainly not highly risk averse).

This also suggests that in migrating from banks to hedge funds, proprietary traders formerly constrained by a bank’s absolute risk budget have gradually slipped into the relative-return world, albeit its region represented by lower risk tolerance.

Does this represent a sustainable basis for generating the payoff characteristics capable of supporting endless growth in hedge funds assets, let alone a new paradigm for active management? From an industry-wide perspective: almost certainly not.

So what do they represent? They are definitely credible alternatives to conventional asset allocation, at some equivalent risk tolerance, where implemented using active, but benchmark-constrained, managers. No Monkey Business will always prefer a long/short equity manager to a long-only active manager, within some limited active management budget for both risk and cost.

Provided an investor can estimate the likely future risk level of an absolute-return fund of funds (or portfolio of funds), it will be superior to its conventional dynamic asset allocation, if it achieves its hoped for alpha, as this is the source of its winning asymmetry.

If the investor has no access to a dynamic asset allocation matched consistently to a given risk tolerance and utility function (such as if relying on IFA-advised portfolios of packaged products), an absolute-return approach with the right risk level is probably better. However, the same lack of adviser skill that means you could do better also prevents the degree of risk in the alternative approach (though almost certainty greater than ‘absolute’) from being correctly identified.

We can conclude, therefore, that the successful adoption of this alternative approach requires agency changes:

• improvements in the identification of risks by agents, from pension consultants to IFAs
• better targeting of risk levels by managers themselves

There is always the risk that, though necessary for greater market penetration, these agency changes also interfere with the management process itself and, ironically, maim or kill the goose that lays the golden alpha egg. Such self-defeating dynamics may account for the durability of conventional approaches to harvesting risk premiums.

Finally, our own experience, offering two different approaches to risk management for the same risk tolerance, suggests that intuitive preferences will ensure both survive to compete with each other.

Our argument that conventional asset allocation can be dynamically matched to a quantified risk tolerance, using dilution of predictable risky assets by a risk free asset, implies a particular view of the world based on ‘mean reversion’ in real equity returns. Reliance on a multi-asset class approach implies a different and more anarchic view of the world, such that the more risks are added the greater the chance of controlling downside risk. Each involves a particular ‘model risk’: that the view of the world will prove wrong. We observe that new clients select between these model risks based either on objective suitability or on pre-existing preferences. We also observe that some prefer to split their hand and adopt both.

1. Language defines the debate

A feature of No Monkey Business thinking is that clarity usually starts with the terms: distinguishing between the language of the sale pitch and the language we can consistently rely on to break down investment products and styles into understandable components. This consistent language tends to be anchored on investment theory.

If you are confident about meaning, you can skip this section and go straight to the arguments about the ‘transformation’ of risk and return (page 8).

Conventional use of the term

Before ‘absolute return’ was hijacked by a new concept, within the last decade, it had a single meaning and usage. It was, quite simply, the return earned (or expected to be earned). As such, it hardly needed a term except to differentiate it from ‘relative return’. Relative return was the absolute return expressed as a ratio of a comparable return, usually that earned by a ‘benchmark’ that is representative of the pool of all investment opportunities from which the absolute return was derived, as a function of the actual investments picked from the pool. Whatever the benchmark, its return was also an absolute return.

An absolute return divided by another absolute return becomes a relative return. Relative returns in consecutive periods can be compounded to create a multi-period relative return, such as to show comparative performance over three months, three years or thirty years.

Absolute returns are by common usage, but not by definition, ‘nominal’: expressed in money terms. A nominal return deflated by inflation, which makes it a ‘real’ return, is technically a relative return. A measure of inflation becomes, in this case, the basis of comparison. This is not very constructive, however, when considering this investment fad. For this purpose it is better to assume that an absolute return for an investment or a benchmark can be either nominal or real. For example, a real return on an equity investment can be compared with a real return on a bond to capture the different effect on each of inflation and the difference in effect according to the period considered.

The common use of relative return that justifies a single meaning is some form of performance assessment, whether the basis of performance is competition between:

• asset types
• managers of the same asset type
• an active manager and a ‘passive’ investment in the benchmark

Conventional usage and investment theory

The conventional use of absolute and relative returns is integral to the dominant theory of capital market behaviour.

This makes a separation of its own between three components of return:

• return explained by exposure to a pool of opportunities (‘the market effect’)
• risk relative to the risk of the pool of opportunities (the two combined, the sensitivity to market movements, being expressed by the Greek term ‘beta’)
• a residual, or unexplained, return component

If the benchmark used to measure the market effect really is representative of the pool (often a big ‘if’), the unexplained return variance is a reasonable proxy for manager skill (or value either added or destroyed), with which its Greek name, alpha, has become synonymous.

The separation of alpha and beta uses the maths of ‘regressions’ and also introduces relative risk to the return comparison. But otherwise it serves the same general purpose as absolute and relative returns: to provide information that clarifies contributions to achieved returns.

Both recognise the important truth that most of the absolute return is earned simply by exposing capital to a particular opportunity set with particular risk and return characteristics. The theory only admits the chance that a portfolio can generate pure alpha if its beta exposures have been perfectly hedged.

In modern portfolio theory, exposing capital to the absolute risks typical of equities is a conscious policy choice an investor makes. A diversified equity portfolio could be predicated on a set of expected returns and variance of returns based simply on historical evidence. Or it could further rely on a theoretical explanation, such as that the expected returns are a reward for the uncertainty of outcome and volatility in the short-term path of equities. Whatever the true explanation, the ‘systematic’ return behaviour of a market cannot be completely diversified away: it comes with the exposure.

If, as a matter of high-level policy, I choose to expose my money to the systematic returns of UK equities, represented, say, by the FTSE All Share Index, my absolute return will be largely explained by that policy decision. Logically, therefore, I should attribute to myself, and my policy choice, the performance of the index. To whichever UK equity manager I choose to implement my exposure, I logically attribute only the relative return against the All Share Index.

In this case I am making no value judgement between the policy decision and the implementation decision, between myself and the manager, or between absolute and relative return. Like beta and alpha, each serves a different purpose.

How investor preferences changed

During the long bull market that culminated with the bursting of the technology bubble in 1999, investors were generally happy to bear equity risk and harvest a risk premium that was very generous relative to cash and quite competitive with bonds (which enjoyed a bull run of their own). In this period, absolute returns from a wide range of assets were highly satisfactory. The conventional usage of absolute and relative returns described above was not much questioned.

Hedge funds represented the vanguard of a challenge to conventional approaches but at this time only those hedge funds with equivalent risk to equities, and similar high achieved returns, were at all successful in gathering assets. In fact, even conventional managers who left the party too early were punished by having assets taken away from them. There is no evidence that investors were prepared to give up equity-type returns for the sake of avoiding downside risk.

All this changed in the bear market of 2000-2003. Unless investors were heavily exposed to technology and media stocks (which they could not all be), the fall they endured was not in fact exceptionally large by previous standards. But it is possible that a long bull market and reduced volatility (derived mostly from lower and more stable inflation) had lulled cautious investors into more risk-taking than they could really tolerate.

Amongst institutional funds in many countries another set of changes was occurring that genuinely altered investors’ rational risk preferences: regulatory reform that called for explicit assignment of capital to different levels of risk taking (‘risk budgeting’) and accounting reform that created significant adverse impacts on balance sheets and income statements if funds experienced ‘paper’ losses due to short-period volatility. The importance of this shift in institutional ‘utilities’ (which simply define how decisions about risk should most rationally be taken) cannot be over-emphasised.

In long-term institutions and the upper levels of private wealth, hedge funds were the main beneficiaries of flows from conventional to alternative assets as they moved their investment objective to reducing volatility relative to equities and to increasing upside volatility relative to downside volatility.

In the private-client marketplace, it was structured products, guaranteed to preserve the nominal value of capital but paying out a proportion of gains in an underlying benchmark over the life of the structure. Guaranteed equity bonds, with a life of typically five or six years, were the main beneficiaries of ‘mass-affluent’ capital but more esoteric benchmarks, such as commodities, joined the array of products offering to provide upside potential with no downside risk.

How conventional language was hijacked

The language of the new innovation was important to its successful marketing. A key part of its sales pitch was to trash relative-return investing, as if it was dumb of investors to encourage managers to be lazy about doing anything to control or change systematic risk exposures. Why follow the market down when you could sell the market? Or why hold the market when you could replace it with options, to protect some or much of the upside? Who cares if a manager outperforms the market if the absolute return is still negative?

From the perspective of correct usage of the terms, these were all criticisms not of relative returns but of the chosen risk level associated with the absolute returns. Differences in risk appetite affect the policy choices made at a high level and so belong in the category of appropriate beta and absolute return risks.

What has happened is that the term ‘absolute-return investing’ has come to mean investing with a particular risk preference that emphasises preservation of capital, as if ‘absolute’ described only positive returns.

The second attribute commonly associated with absolute-return investing is a wider range of strategies, and resulting sources of risk and return. This can be viewed as bringing structural advantages (better scope for diversification benefits) or potential performance advantages (more bets for a manager to make).

The third attribute is that the returns from absolute return investing are mainly derived from alpha.

In hijacking the language that has served us very well for over half a century, the new fad risks a great deal of confusion about where returns come from, how to make choices and what we should pay for and why.

2. Transformation

What we have learnt from the distinctions using language based on investment theory is that the key to the sales pitch is a claimed transformation of investment risks and returns. It is time to examine what this really means.

How are the risks and returns altered?

The high-level choice to accept equity risk in order to seek equity reward assumes a trade off between risk and return in order to maximise some utility, such as a level of future desired wealth which is consistent with both tolerable paper losses along the way and a tolerable shortfall from the desired outcome. The trade off is only made because the possible wealth outcomes without risk taking are not satisfactory, since we all start out wanting to avoid risk.

Whatever the level of risk that maximises our utility, we can obtain it by controlling exposures to risk – in other words, by the dose of risk. Usually we do this by holding more or less of risky assets relative to risk free assets.

For any alternative investment to be superior to a conventional asset allocation with the right risk level, it has to alter the risk and return trade off. The implicit or explicit claim made by so-called ‘absolute-return investing’ is that it does alter the trade off: by giving up more risk than expected return compared with its conventional equivalent.

This claim relies on two types of effect.

1. Diversification benefits are increased, by introducing more return sources that have low correlations with other portfolio components: they do not behave the same way at the same time
2. Specific strategies are pursued that put a floor under downside risk that allows for capital preservation while still offering probable gains.

Diversification effects
The first assumes that, for a given level of satisfactory return achievable from either type of portfolio, the risk of the more diversified one will be lower. The risk-adjusted return is therefore superior. This claim relies both on the return predictability relative to conventional strategies and the assumptions about correlations.

Diversification benefits can be increased by adding conventional long exposure to unconventional systematic risks and returns. Provided the new exposure genuinely has lower correlations with the existing components, it should be possible to arrange the mix of assets to produce the same return with lower risk or to take the same amount of risk for a slightly higher return.

What might these unconventional additions be? There are actually few assets that can be relied on for this structural diversification effect.

• Direct property holdings through some packaged investment (as distinct from property shares) already form part of many conventional portfolios
• Private equity exposure can be easily acquired through a few quoted securities but these are also part of the broad market
• Commodities, though not an asset class, generate a genuinely different set of risks and returns when combined in an index-fund structure (combining cash collateral and index futures to create an ungeared portfolio of commodity exposures)

However, by using derivatives in a portfolio, a manager can further increase the range of risk and return sources that are not highly correlated. Ironically, the additional sources arise largely from relative returns: betting on the difference between two assets by being long one and short another, such as the relative return of oils stocks against the market or Royal Dutch Shell against BP.

To the extent it is relative bets that mainly differentiate absolute return portfolios, it is reasonable to argue that the expected risk-adjusted returns will be better than can be achieved by mixing conventional assets. Many of these relative bets work mainly by rearranging risk exposures already present in a conventional long-only portfolio, such as movements in interest rates, in option volatility and the pricing of credit risk. Thus the scope to add new risk exposures is less than generally assumed but there are still benefits from changing the distribution of exposures (since diversification is a function of weights as well as numbers of different exposures).

It is debatable whether relative bets have the same ‘normal’ payoff distribution as conventional asset combinations, which is what the comparison between risk-adjusted returns assumes. It is likely that the main difference is occasional losses much greater than normal. This is true of many so-called low-risk, low-return asset strategies in finance that are as old as markets: bank lending, insurance underwriting and arbitrage. In the 15 years or so of true history of alternative investing, these ‘fat tail’ events have arisen within individual products but have not been in evidence in diversified exposure to alternative products.

If, consistent with the loss-averse utility, we assume diversified exposure is the most rational way to use alternatives, it is reasonable to assume that the underlying mix of risk factors in this richer mix produces normal and symmetrical returns. In that case, the payoffs can be matched by those of a mix of conventional risky assets and a risk free asset at some equivalent risk tolerance.

Something other than diversification is required to transform the distribution of payoffs.

Achieving asymmetrical payoffs
The desired portfolio effect, increasing upside relative to downside risk, requires one or both of two quite different activities on the part of the manager:

• active management that consistently pays off
• option strategies

Each has shaped the absolute-return industry to date in quite different ways and each poses a very different threat to the conventional asset-management industry. They may both produce asymmetrical payoffs but investors adopting either will experience quite different opportunity costs.

Active-manager skill

Asymmetry relies on managers being able to judge when to be ‘long’, ‘short’ or ‘neutral’ in terms of the conventional or alternative risk factors in their opportunity set. Research has shown that market timing in conventional investment management is unpredictable, in other words past wins do not predict future wins. So for skills to be a predictable means of reducing downside more than upside, they have either

• improved dramatically (perhaps because of being freed to operate without the constraints of a long-only, benchmarked portfolio), or
• migrated from activities outside mainstream portfolio management where they already existed (as in banks’ proprietary trading desks or commodity firms)

A less obvious implication of presuming skill is that the investors on the wrong side of trades with predictably skilled investors do not learn. This was possibly a factor in the early days of hedge funds but it has probably already been negated. Three examples can be cited of ‘sucker’ money wising up:

• central banks are more careful about advertising their exchange-rate policy
• equity investment funds have sharpened up their dealing capabilities
• index-tracking funds have tried to avoid predictable rebalancing patterns

Option strategies and capital guarantees

Options are the only systematic and predictable means of ensuring asymmetry, by providing upside potential coupled with limited downside risk. Whereas futures and swaps can be used to implement exposures (long or short, with more or less gearing), or to create relative exposures (or differences between two absolute returns), we have decided they probably still generate symmetrical payoffs, with equivalent upside and downside risk.

Whereas hedge funds can use both active management of their risk exposures and option strategies to try to generate asymmetrical payoffs, structured products that guarantee the capital subscribed are limited to option-based strategies.

The issuer of a structured product that guarantees to return your capital, on a worst case basis, at the end of, say, six years, is not offering you a free lunch. On the contrary, the structure uses assets priced in some of the most actively-traded, liquid and efficient markets in the world.

The first element is a discounted six-year bond, ie a bond that at the prevailing interest rate will repay both the capital and rolled-up interest at par in six years. If, as a function of interest rates at the time, this requires only 70% of the money you put up, 30% of it is available for two other elements:

1. the issuer’s and promoter’s cost and profit
2. a call option on whatever underlying asset or index the structure gives upside exposure to

The interest rate (on which the discount depends) and the implied asset volatility (on which the option cost depends) together dictate the possible upside participation and the probability of gain.

Since the discount is equivalent to the opportunity cost of a cash deposit, the value of the structure must be a function of volatility. If we assume that over any long period the market’s errors in estimating actual volatility will even out, there is no long-term value to be obtained from rolling over call options. The best forecast for a long-term compounded return from a series of rolled-over structures is the cash return less issuers’ and promoters’ margins.

What there may be are rational trading opportunities for investors backing a particular view on volatilities relative to the prevailing interest rates.

Structured products might initially have been popular because interest rates were very low and investors were desperate for yield improvement. It did not require much upside potential to look enticing. But they have exploded in popularity since interest rates have risen, as this has increased the participation rates. The only common element has been that volatility has remained low. The recent combination is perfect. It can only get worse, probably because credit-induced risks to economic activity allow interest rates to fall but only at the cost of higher volatility in financial assets.

Most buyers of structured products are probably not shown a probability distribution for the payoffs of each product at the point of sale. Such illustrations have become more common but are not a regulatory requirement. These would reveal the bunching of outcomes around the cash return and the very low probability assigned to eye-popping upside.

After the sale, profits on the contract open up downside risk that logically is inconsistent with the risk appetite at point of sale. For example, a successful Xinhua contract on Chinese stocks at one stage implied a possible loss (for supposedly loss-averse investors) of 36%. Even though some of the contracts are reasonably liquid, many buyers and their agents act as if unaware of this and so do not trade out of in-the-money contracts to new on-the-money options.

Even knowledgeable investors are unlikely to be aware of the compounded payoffs of a long-term strategy of rolling over products, as these require modelling of the two time-varying and partially dependent factors: interest rates and volatility. These cumulative results need to be compared both with cash and with some level of exposure to risky-asset returns, to demonstrate the potential opportunity cost over long periods.

Clearly, capital-guaranteed products will survive for those able to use them discriminately. But they cannot possibly replace either conventional asset allocation, for a given risk tolerance, or a dynamic alternative to absolute-return investing, such as hedge funds.

This reinforces the observation that absolute returns are a misnomer, since structured products have been the only saleable combination of absolute capital protection and upside potential. Hedge funds have required downside risk, to be saleable.

Target-return funds
There is a third category of absolute-return funds that has emerged in Europe, in both onshore and offshore jurisdictions for regulated funds. Marketed as actively-managed against a cash benchmark (typically either £, Euro or $ 3-mont LIBOR cash rates), these adopt absolute-return techniques but are constrained by the product rules of their jurisdiction. Until these constraints are removed, they are unlikely to challenge either their conventional equivalents or retail funds of hedge funds.

These funds are found within a number of categories according to the benchmark or the instruments mainly used:

• Money Market (funds normally also termed ‘enhanced cash’)
• Global Bonds
• Cautious Managed (and its life-company equivalent: Defensive Managed)

Their performance has been disappointing in their early stages but they set themselves a much harder target by including a return on cash instead of capital invested.

Valuation issues
One of the new risk factors that any diversified approach to alternative investing is likely to introduce is illiquidity. Portfolio theory requires this to be remunerated by a risk premium, so it is another example of an increment to the diversification of multiple risks.

It is something else as well. Holding less liquid or untraded positions, including derivative positions that call for some benchmark reference for ‘marking to market’, introduces issues about price discovery that have a bearing on the return path of the fund, and hence its observed volatility.

Confidence about pricing is also important for funds that remain open to new investors, to ensure equity between them.

The trend is for illiquidity to increase, for example as hedge funds match the strategies of private-equity partnerships. For absolute-return investing to challenge conventional asset management, these issues will need to be addressed much more convincingly than hitherto.

How we advise clients
We have adopted two alternatives that we believe cover all the investment needs we encounter when organising, co-ordinating or managing investments for individuals and their families.

1. Goal-based, model-driven portfolios dynamically matched to target outcomes and time horizons
2. Multi-asset class portfolios which are absolute/relative hybrids

1 Goal-based portfolios
Our model-driven portfolios use a lean mix of assets whose return behaviour is (unusually) evidenced by long data histories: the major equity markets. Risk tolerance for each client goal is determined in a collaborative, planning process in which some quantifiable terms of reference that unequivocally describe the goal are ‘solved for’. The process forces them to be internally consistent, excluding any possibility of a mistaken free lunch or unrealistic expectations:

• probable real outcomes (the targets) expressed in terms the client can relate to
• resources required or assigned
• the planning time horizons (as opposed to decision time horizons)

The risk tolerance results from the selection process, as the client directly exhibits how they value the different attributes of success and failure, such as how upside potential is traded off against a floor outcome.

Once discovered, that risk tolerance becomes the basis for us to manage the portfolio, as market conditions alter and time horizons grow shorter (unless set to roll forward).

Ensuring consistency with the risk preferences requires us to dilute the exposure to risky assets using a goal-specific risk free asset. We cannot rely on diversification because the risky assets are few and too highly correlated.

This is a much tougher competitor for absolute-return investing because it provides the formal, well-thought risk-management process that absolute-return managers rightly say is missing from conventional approaches to asset allocation. Absolute-return investors want to give greater value to loss aversion. We translate that into shortfall aversion: outcomes that will cause regret or hardship.

2 A multi-asset class portfolio
What we organise here is not an absolute-return portfolio but as we have seen is closer to the reality of much of the market, where downside risk corresponds to a risk-diluted portfolio with a relatively moderate risk tolerance for a long horizon.

Changing the risk tolerance further, up or down, requires either dilution (probably with cash) or gearing (which is usually part of a total balance sheet approach and therefore implemented by means of a mortgage secured on property, not the portfolio).

For its assumed risk tolerance we still want maximum diversification of risk exposures, hence the use of the term ‘multi-asset class portfolio’ rather than absolute-return portfolio or ‘alternatives portfolio’. We delegate both the risk-factor structure and the manager selection to a third party but if we cannot secure the full diversity we want we complement that core manager with some cost-effective exposure to the missing factors. At the moment, our solution requires some complementary exposures to property and commodities (which we select).

We hope for asymmetry due to active-manager skill, but we do not assume it. Our preferred core manager targets a mean annualised 3% of beta return and 2% of alpha return but we would be happy with 1% and overjoyed if it all arises when we most need it.

How clients select

They select by

• ‘Model risk’: which implied view of the way the world works they trust most
• Suitability for the goal in hand Cost (a model-driven portfolio relying on index tracking is much cheaper)
• How they value the ‘option’ of earning positive alpha

They also split their hand: diversifying between the two.

Model risk
Our goal-based model generates stochastic simulations of thousands of possible paths and outcomes. These reflect a view that equity real returns are volatile but not entirely random: they have a tendency to revert to a mean, or trend, real growth rate.

We observe that both the growth trends and the dispersion around the trends are remarkably similar for those markets with long histories of real total returns for a diversified index of securities. The paths are also quite similar most of the time but deviations can diverge at times, so the correlations between the markets vary – but not predictably. Projections requiring a high degree of confidence (as high as 99%) therefore need to allow for close convergence of correlations.

As noted, we control the range of outcomes by dilution: holding more or less of a risk free asset matched to the target outcome. If, for instance, the target is expressed in real terms (because its purchasing power must be maintained), the matching risk free asset is an index linked gilt.

This view of the world stakes a lot on the relative predictability of a few markets with long histories of their behaviour in real terms. It might be as few as four markets or regions: UK, Europe, USA and Japan. Together these represent 90% of the world’s opportunity set by value but much less by number of investable opportunities. It is a view that trades off predictability against diversification, where adding assets leads to diminishing certainty about the return, risk and correlation assumptions.

It is also a view that places more importance on outcomes than on paths, or interim volatility. This is self-evident in the use made of index linked gilts which, though they match the outcome with certainty, are nearly as volatile as equities if both long-dated and very low-yielding.

It is an approach more naturally suited to goals where predictability, horizon matching and quantified tests of goal progress are all highly valued.

The multi-asset class approach implies a radically different view of the world in which the future is much less predictable, even where evidence is most plentiful. The Darwinian process of change that is implicit in an index of equity securities over long periods (which though called ‘passive’ is actually constantly being refreshed) counts for less in this world view than a Darwinian process of manager adaptability. Managers compete, rather than markets competing.

It is a view that values being ‘in with a chance’ of outsmarting other investors, even other smart ones, even if there is relatively little total available alpha that all these investors are fighting over.

It embraces Grinold’s Law of Active Management that holds that the chance of earning positive alpha is a function of the number and frequency of independent bets made: bet little and often.

It is a view that suits an investor (or an investor’s goal) that requires low downside risk, not just because they want to avoid it (don’t we all) but because there are particular adverse consequences that weigh more heavily than lower cumulative outcomes. Examples are individuals (or bodies of trustees, including unknown future trustees) who fear that they will not be able to stay the course in the event of a very deep fall in portfolio values.

Splitting the hand

Not surprisingly, many clients choose to embrace both world views, as another contribution to their diversification. They may choose differently for the same type of task, but at the level of their entire balance sheet they benefit in similar ways from splitting their hand.

TAGS: asset allocation, Investment process, market timing, performance