Tools of the risk managment trade
Never before have the investment portfolios controlled by global asset managers received the attention they are getting at the dawn of the twenty-first century. Controls and restrictions are being swept aside as the investment community attempts to defuse the real millennium time-bomb: meeting the retirement aspirations of the world’s rapidly ageing population.
This ever-increasing search for excess return is compelling practitioners to look ever wider for assets that meet their exacting investment criteria. With this impetus behind us, it is scarcely unexpected to find our portfolios increasingly invested in widely dispersed geographical regions, broader and broader ranges of credit quality and stocks for which traditional valuation measures are meaningless.
Clearly we need to examine carefully what is being done to control and manage the risks created by these increasingly abstract investment strategies. Whether one is seeking to determine the optimal level of currency risk in a portfolio, the appropriate level of investment in emerging markets or the appropriateness of an alternative asset class in your fund, risk management tools are now an essential part of the investment manager’s armoury.
Fortunately, superior investment technology has developed rapidly in parallel with this industry trend of increased demand for complex, global investment strategies. Inverse floaters, knock-out options and total return swaps are all examples of very recent developments that have made the tail-end of the century such an interesting time to be in the investment management industry. Much of the commercialisation of the academic thought that underpins such ‘derivative’ products has only been possible since the advent of widely available analytical resources, in the shape of the humble personal computer. It is ironic that these words were composed somewhere over the mid-Atlantic on a laptop computer with more processing power than Boeing had to design the 747 it’s riding in!
This article explores some of more important risks facing the global investment community and the work that we have undertaken to assist in managing and controlling them.
Given the myriad of risks that a global, multi-asset class portfolio exhibits, perhaps it would be instructive to consider the cycle that each investment passes through on its way to becoming part of such a portfolio. We can consider this investment cycle as having three, distinct phases:
q Pre-trade Decisions that must be made before the transaction to buy (or sell) the security occurs;
q Trade Decisions made in the course of executing the transaction;
q Post-trade Decisions that occur as a result of the trade having occurred.
Note that this is indeed a cycle. It is to be expected that the pre-trade phase is significantly informed by the results of the post-trade phase. Hence many of the techniques overlap.
In the pre-trade phase, the broad investment objectives must be considered. Here we often find that risk management is considered not just from a defensive point of view, but also as a tool to understand and perhaps even increase a portfolio’s risks concomitant with the associated increased returns. Furthermore there is an increasing appreciation in our industry of the benefits such a pro-active risk management strategy can have in marketing a fund to prospective customers. Not only does this demonstrate the prudence of the investment firm, but may also provide useful and meaningful information to the ultimate beneficiary of the fund.
To support this “communication” aspect of risk management – not just doing the right thing, but being seen to be doing it – the information produced by such a risk management system must be not just correct but also meaningful. We have invested significant resources in making such communication possible through a risk management toolkit that we also use internally.
From the perspective of the investment management firm also there are serious risks to be considered at this stage. It has been said by some with far greater experience than I, that the worst position for a portfolio manager to be in is “wrong … and alone”! Doubtless we have all railed at one time or another against peer group universes, league tables and the inexorable rise of indexation – and particularly their inappropriate application to a particular investment scenario. Nevertheless, business risk is something with which we must all deal to survive and thrive in the new millennium. In addressing our clients’ concerns in this area, we exploit a unique resource – a continuous database of security transactions recorded since 1994 and analysed to provide insights into global market holdings and flows. This research enables us to provide investment managers with a picture of how their market in aggregate perceives current aspects of financial risks, such as the degree of contagion present in Asian markets or the degree of liquidity in the European bond market. Investment managers are then able to compare their perception of such risks with those of the their peers and competitors to establish their relative risk tolerances.
At the individual portfolio level, it is perhaps only too obvious that risks can come in many dimensions, many of which may be in conflict. For example, a pension fund will be compared to similar pension funds in its country – or across Europe – but also must seek to achieve a particular level of return to meet its investment objectives. A currency manager might be obliged to beat both his explicit benchmark, such as 50% hedged, but also the “do nothing”, unhedged alternative. Such risks can change significantly during different market regimes. For example, our analysis suggests that asset class risk (will equities outperform bonds?) is much more important than currency risk (will the yen outperform the euro?) during turbulent market conditions, whereas the absolute converse is true when markets are quiescent. To put all these concerns together into a tool that can help the portfolio manager make the necessary rational trade-offs between such conflicts and resolve them in a transparent and objective manner, we recently launched an optimiser, coupled to a vast on-line historical database of market variables, updated daily.
It is most important with such a model of the investment process that the phases are joined seamlessly. History is littered with examples of investment strategies that looked good on paper but fell down upon execution. Perhaps the most memorable is still the failure of ‘portfolio insurance’ to provide any such thing during the market crash of October 1987. In this case, the finest brains in finance were of no use in the face of a market without buyers for the securities the models were trying to sell.
For this reason, we have combined the pre-trade research and optimisation tools with our trade tools in a single electronic platform. The objective is to unify the information flows throughout the investment cycle to minimise the potential for human error and reduce to time from the creation of an investment idea to the confirmation that its implementation has been completed. Time is critical to ensure that any errors or misunderstandings that may nevertheless still occur are identified and corrected as quickly as possible to minimise any adverse price movements or market risks.
It is in the post-trade arena – often known as the middle office – that risk management tools are most frequently considered. Clearly, however, there is a danger of “locking the door after the horse has bolted” if this is the only arena in which risk is considered. For this reason it is, I believe, incumbent upon organisations such as ours to provide solutions throughout the investment cycle and to encourage their use within the global investment management community.
But the big advantage that the post-trade phase has over the earlier two is that at this point everything that can be known is known! The results that can be derived are therefore somewhat more tangible and can be used to cross-reference the necessary assumptions of the earlier phases. As an example, we have launched the industry’s first risk-adjusted performance measurement product. This is a technology-based decision support application and is based on well-established statistical principles including value-at-risk (VAR) analysis. This allows investors to gain a much greater insight into the success of their own securities lending programme than the more usual, quarterly revenue stream permits.
To help plan sponsors, as well as their individual managers, understand their risk profile, we have developed a unique range of post-trade risk management products. These allow us to inform, on an aggregate basis, our clients’ investment strategy committees – so that it can form a precise view of the “big picture”. Conscious that the users of risk measurement products may be aiming to satisfy different objectives, and may subsequently have different definitions of risk, the approach is tailored to two main areas.
First, adapting an approach originally developed by JP Morgan for use in banks and trading operations, the VAR calculator allows plan sponsors to translate ‘risk’ as a generic term into a money value. This value indicates the maximum potential loss over a given period with a given level of confidence. This allows the risk characteristics of globally invested holdings to be compared, as the calculator uses a consistent approach to capture the exposure to, and volatility inherent within, a range of asset classes or security groupings.
The VAR calculator provides a top-down view of risk. It allows the user to understand not only the impact of the current asset allocation decision, but also the greatest contributors to total risk. In addition, through the use of marginal and relative VAR, the impact of changes in the allocation decision and the nature of the relationship between groupings can be identified. The VAR concept cannot capture ‘event’ risk, or the potential loss that could be suffered in the event of a market crash. However the use of an appropriate confidence interval in the calculation allows the user to expect that the actual loss will not exceed the VAR in the specified time horizon, 95% of the time, for example.
The second area within which we have targeted our risk product range relies more on the traditional definitions of market risk. An analytics reporting package attempts to address the risk monitoring requirements that arise within an individual portfolio or mandate. As the traditional measures of risk tend to be asset-specific, risk reporting is tailored to cater for both equity and fixed income portfolios. The fund beta is used as the fundamental risk measure for all equity portfolios, whereas duration becomes the risk proxy in the analysis of fixed income portfolios.
In the analysis of equity portfolios, the use of risk statistics may form part of the control function – for example, monitoring the adherence to mandate in a passive portfolio, through the use of the tracking error characteristic. The use of risk-adjusted measures such as the Sharpe ratio can also be instrumental in comparing individual managers with similar mandates.
As we look into the next century perhaps one forecast that is not unduly controversial is that investment strategies are not going to get any simpler. Risk management is still very much in its infancy and much has yet to be done to identify and build an industry-wide consensus around the appropriate measures to use for investment-related risks in the portfolio context. Clearly the biggest risk is always the one yet to be identified – and so a consideration of the phases you use in your investment cycle may assist in identifying possible weaknesses and the possible risk management solutions available.
Jeremy Armitage is vice president at State Street Associates in London