Marginal gains are key for efficient portfolio management, and Defined Returns can help
Small adjustments to asset allocation optimisation processes can yield important gains, improving the frontier of efficiency for investors with risk controls on their portfolio
Tom May, Chief Executive & Chief Investment Officer
Experienced portfolio constructors know that if there is one behavioural problem that can be the most deadly for fund managers it is a fixation with winning big. The belief that the goal is to find an idea or strategy which can blow away the competition invariably leads to hubris.
Indeed we measure this bias by examining the kurtosis of portfolios, in order to hunt for evidence that outperformance has been achieved by many small good decisions rather than one big bet that went right. It is the most illusive of behaviours amongst fund managers.
The same behavioural problem of course can affect portfolio constructors themselves leading them to hunt each year for the asset class which can outperform all others. For most professional investors, we protect ourselves against this behavioural bias with an asset allocation approach rooted in modern portfolio theory. We base our asset allocations on historically optimised mixes which deliver more consistent outcomes and ensure that short-term views do not sway us from following long-term patterns.
Yet such an approach, which admits the impossibility of calling the one thing that can work, must never prevent us from hunting for marginal gains. Small adjustments to asset allocation optimisation processes can yield important gains, improving the frontier of efficiency for investors with risk controls on their portfolio.
When Dave Brailsford took over the role of head coach of British cycling – with the mission of winning gold at the 2012 Olympics and on the grand tours – he realised quickly that one of the biggest problems he faced was that the team’s fixation with victory had become the thing stopping it succeeding. By focusing on the distant goal of victory it had stopped searching for the small marginal gains, which when compounded together, consistently lead to that victory. This involved details as small as instructing the cyclists how to properly wash their hands to marginally lower the risk of getting ill during the Tour de France.
Our core conviction at Atlantic House Investments is that sound portfolio construction is based not on predicting which asset will rise the most but in harnessing the power of diversification to its full potential, searching for and grabbing at any marginal advantage in that process that we can.
We have never argued that the derivative-based funds we manage will transform portfolio outcomes. We know these are ultimately determined by the path of shares and bonds.
Yet we do believe there is sound evidence that a judicious use of strategies such as defined return investments produces the sort of marginal gains which over the long-term can make all the difference.
A more efficient frontier
One key measure of the added efficiency produced by defined return investments comes from examining the impact that their use has on the efficient frontier for a range of risk-rated portfolios.
This optimisation process, the classic core of most asset allocation processes, requires investors to make judgements about the maximum allocation to make to any one asset class. We have examined the impact of defined return investments if we set the cap for an allocation at 0%, 5%, 10% and 20%.
We have taken the performance of the Atlantic House Defined Returns fund over the decade as a proxy for the performance of defined return investments in general. We should note that this means we are examining a very specific form of defined return investment. Namely instruments tied principally to the performance of the FTSE 100 with capital protection barriers set at around 35% and return on capital barriers of around 25%. The instruments are generally structured over six-year time frames with calls at each calendar year.
Here a mean-variance optimiser run over the ten years until the 30 November 2023 suggests that the Sharpe ratio of risk-rated portfolios have been progressively improved using defined returns.
The gain is most profoundly felt for medium-risk investors. This is in line with the evidence around risk parity investing which instructs us to allocate a larger portion of our ‘risk budget’ to medium-risk asset classes than has classically been the case for portfolio constructors.
Source: Bloomberg PORT Optimise 05-12-23
A mean variance optimiser mechanically borrows from a variety of asset classes to arrive at its Defined Returns allocation. However, for the medium-allocation approach where Defined Returns is capped at 10% it takes most significantly from high yield bonds where it reduces its allocation by 5% and from infrastructure where it reduces its allocation by 1.9%.
Optimised asset allocation compared
Defined Returns in a Mean Variance Model with a maximum 10% weighting for an investor targeting a gross before fees return of between 6-7%.
Source: Bloomberg PORT Optimise 05-12-23
Of course, all measures such as this have analysed the performance of medium-risk assets during a period where they have not failed. These asset classes such as high-yield bonds, infrastructure, or defined return investments may look like they carry less than equity-like risk for long periods of time. However, during significant market failures, they can behave at least as badly as equities.
At these moments of peak crisis correlations come together and the overall value of a mean variance approach is eroded for asset allocators – a lesson hard learned during the Global Financial Crisis.
Modern portfolio theory does not have a solution for protecting portfolios against the behaviour of medium-risk assets in left-tail events. Instead, investors must look to portfolio protection or the use of active uncorrelated strategies. At Atlantic House our Balanced Return fund, which allocates 60% of its portfolio to defined return investments, pairs them with left-tail crash protection strategies to mitigate for this issue.
We must be aware that defined return investments have this characteristic but also not see them as unique in this respect amongst other comparable asset classes. We believe left-tail protection has merit whether a portfolio constructor has allocated their risk budget to defined returns or conventional equities. Our Uncorrelated Strategies fund offers a potential solution to this problem.
The number of 'good' years
A second means of analysing the efficiency gain produced by a defined return investments comes not from looking at the overall impact over time but at the frequency with which a portfolio can deliver a ‘good’ outcome for investors.
The smoothing of returns so that positive results happen more frequently and less spectacularly has a number of important benefits for investors.
Firstly, it reduces the risk of behavioural mistakes such as over trading a portfolio or capitulation after a loss. Steady returns lead to more patient investors.
Secondly, it reduces the frequency with which an investor must withdraw money from the portfolio after a substantial loss over the course of a long period of decumulation and therefore reduces the impact of pound cost ravaging, improving the personal utility of a pot of money to that individual.
To understand the improved efficiency of ‘good’ returns that a defined return investment produces we have to agree on an alternative for comparison. Clearly a mean variance optimiser will objectively compare defined returns against every asset class in search for mathematical efficiency.
In this instance though we must instead choose the asset class we think defined returns are most ‘like’. A reasonable comparison would be UK large cap shares. The defined return investments in our portfolio are generally tied to the performance of the FTSE 100 and the correlation between Defined Returns and large-cap UK shares is 0.86 – the closest comparison to any of the asset classes used in the 15 asset classes within the mean variance optimiser.
When making a comparison to large-cap UK shares, we encounter some of the complexities of Defined Return investing. Over the past ten years, the maximum drawdown of Defined Returns is 36.7%, slightly greater than the 34.2% of the index.
This drawdown, occurring at the very start of the Covid-19 pandemic occurred because as the index approached the 35% loss barrier attached to the instruments in the Defined Returns fund, they fell more sharply than the index, having been more resilient in the early phase of drawdown.
In this instance the fund fell more sharply than the index. This is because the instruments do not carry the dividends attached to the stocks and therefore should perform slightly worse than a passive exposure to equities close to or below the barrier.
Time to recover following Covid-19 drawdown
Past performance does not predict future returns. Source: Atlantic House and FE fundinfo 31-12-19 - 31-12-20.
However, whilst this pain had to be taken in this instance another key characteristic of defined return investments was also exposed. Namely the pace at which a recovery from a drawdown can occur. The market had to only rally a little away from that 35% barrier in order for the defined return investments to re-price up significantly. This meant that whilst the drawdown in this instance was similar, and indeed slightly worse, than shares, the time to recovery from the drawdown was far less. The Defined Returns fund recovered its early Covid-19 loss within 197 days, slightly less than half the 405 days it took for the index.
Source: Bloomberg 05-12-23
Past performance does not predict future returns. Source: Atlantic House/Bloomberg - Atlantic House Defined Returns Fund B Acc shares daily rolling 5-year annualised performance frequency compared to Solactive UK Large Cap Index (Net Total Return) in GBP daily rolling 5-year annualised performance frequency. 04-11-13 - 30-11-23.
The Lehman's moment
Whilst all defined return investors should be conscious of the impact of the holding in extreme market events it is also right to note that in many bad, but not catastrophic, scenarios in recent memory the fund would have helped portfolio returns. This is not limited to the Covid-19 pandemic which is the most extreme scenario to occur in the 10 years over which we have managed the Atlantic House Defined Returns fund. Scenario analysis also sheds a light on how the fund would have behaved in other extreme events in history if it had held its place within an optimised portfolio.
For example, a balanced multi-asset portfolio constructed of passive instruments and sitting at point five on our efficient frontier would have drawn down 9.36% in the event of the Lehman Bank default in 2008 with no allocation to defined returns. This same portfolio optimised to contain a capped 5% weighting to Defined Returns would have fallen 8.69%, with a 10% cap 8.14% and with a 20% cap 6.72%.
Source: Bloomberg Scenario Analysis 05-12-23
The philosophy that underpins Atlantic House Investments is a conviction that careful portfolio construction making use of derivatives can produce marginal but real gains for multi-asset investors that enable them to build better, more resilient portfolios.
We hold to the view that there is no free lunch. Every asset class or investment we prefer has complex behaviours in different scenarios that cannot be accurately visualised on one graph or understood over one time period. However, we believe that provided we can clearly explain the behaviour of Defined Returns investments they can play a vital role in portfolio efficiency.