Periodic Systemic Risk & Investment Strategy

   Products & Services   About PIMCO   Press Center   Bond Resources   Career Information   Content Archive   PIMCO Foundation

Viewpoints

Introduction My aim in this note is to raise some investment strategy issues that are brought to the forefront by focusing on two things: the presence and dynamics of systemic risk, and the presence of illiquid assets in the portfolios of many large investors. Investors have been hit hard by the current crisis. Lessons are being learned and investment strategies revised. The center of this learning process has to be rethinking models of risk. The central lesson of the crisis seems to me to be that a component of risk is not static. It evolves in a way that is not yet completely understood. Static risk models are not so much wrong as incomplete. An important component of risk is not stationary. The possibility that a component of risk in the system as a whole can rise in a manner that is difficult to detect is important. When it happens it causes the "normal" correlations of returns among asset classes to shift rapidly upward even as asset values rise (usually to unsustainable levels) and then fall (sometimes very suddenly). That shift causes the diversification and hedging models and risk mitigation strategies to malfunction.A subset of large investors, including universities, foundations and pension funds, have income or payout requirements. The crisis highlights the question of how much volatility in income is tolerable or desirable. The rules governing payouts may need to be adjusted in light of the crisis experience, and the possibility that short and medium run returns are misleading signals about long run returns because of rising systemic risk.In this and other areas of decision-making in complex environments, choices are strongly influenced by implicit models that we carry around with us. A model in this context is just a set of assumptions that we use to translate choices in an attempt to predict outcomes. These models have complex origins: Typically they are learned initially from others and then adapted in response to experience. Debates and disagreements about strategy and policy choices often have their origins in differing assumptions about the underlying reality, while the model differences remain unstated. The efficiency of the debate and the quality of choices can be materially improved by taking time to be explicit about the underlying assumptions and models.1Non-Stationary Components of Overall RiskExperience and a growing body of evidence and research indicate that risk may have two distinct parts: a non-systemic and stationary component and a systemic and non-stationary one.2 On the systemic side, the intertemporal dimensions are important. Major systemic disruptions do not occur every year. Instead, instability builds up and then the system is shocked and resets, with the exact timing being thus far quite unpredictable.3 As a result, wrestling with systemic risk requires a reasonably long timeframe, longer than that associated with the stationary risks that are the focus of most of the attention.In Table 1 below, I take a ten-year period and assume that there are nine years of "normal" average returns at various rates followed by a "bad year" caused by the systemic risk component. On the left side are a range of average returns in the nine "normal" years. The top is the percentage decline in the year of the shock, running from 0% to 25%. The table shows the average returns for the ten year period with the shock factored in.

 Source: Author's CalulationsHypothetical Example for illustrative purposes only. The normal return range provided is intended for illustrative purposes and is not indicative of the past or future performance of any PIMCO product. Returns do not take into consideration fees or expenses.

Source: Author's CalulationsHypothetical Example for illustrative purposes only. The normal return range provided is intended for illustrative purposes and is not indicative of the past or future performance of any PIMCO product. Returns do not take into consideration fees or expenses.

Albeit stylized and hence only illustrative, if this is an aspect of the kind of environment we live in, then it has a number of implications. Long run investors should probably think of the returns as more like those depicted in Table 1 rather than the unadjusted "normal" returns. An endowment might consider basing its payout decisions more on long run shock-adjusted average returns (more like Table 1) than on a simple trailing weighted average of past returns or ending asset values, the normal practice now.4 Of course there are competitive and other consequences of adopting this more conservative approach and they will vary across institutions. The point here is not to suggest that there is one right answer, but rather that thinking about the longer time horizon dimensions of risk is a material consideration in investment strategy and in spending rates and decisions.This is not an argument against the high return strategies. Investors may very well take the position that the strategies that look to generate the higher normal and post-shock average returns are worth it. I will say more about that in a minute. But the high return choice should be accompanied by a thoughtful companion decision about the need for and pattern of payouts and the use of the returns. There may be investors (a minority, I suspect) who do not have payout or income requirements and can focus almost exclusively on the long run returns. For these investors, hedging options that are designed to smooth returns at some cost in terms of the level may be of less interest. Even then, the issues having to do with liquidity and flexibility (discussed below) are relevant in that they impact long run potential returns.One can think of all this as an inherent tendency toward mean reversion. There are long run returns associated with various investment strategies and capabilities. The short and medium term returns can deviate considerably and for extended periods from the expected long run returns. Put another way, when returns seem abnormally high for an extended period, they probably are and something is likely to bring the returns down even if we don't know in advance what it will be. This is particularly relevant to large investors who cannot totally segregate themselves from the macroeconomic fundamentals without abandoning a reasonable pattern of diversification. Traders can do better than that. The trading superstructure is partly market-making and -deepening and value creation. But a lot of it is a zero sum game in which some win and others lose. High returns in that part of the system do not therefore generalize or add up to the potential for high returns in the whole economic and financial system. Successful value investors may enjoy some degree of risk mitigation by virtue of an unwillingness to invest in what they consider to be overvalued assets (absolutely overvalued rather than relatively). They are in effect shifting the portfolio away from "over-valued" assets as they become more numerous in the run-up to a potential crisis. Nevertheless, they are not invulnerable to systemic risk. Fair and under valuations are not exempt from the downward pressures of a crisis or the resetting of asset values after a build-up of systemic risk.Responding to Periodic Systemic RiskIf one accepts that periodic bouts of systemic risk are likely to be a feature of our future as well as our past, what can be done to mitigate the impact of the shocks that occur? First, if one thinks that one can or may be able to detect a pattern of rising systemic risk, then the investor can take action in the form of portfolio adjustments and/or the addition of hedging instrument"“type holdings (tail risk hedging) that have the effect of helping mitigate the shock.5 Done successfully, that can have a significant effect on long run returns. Of course certain instruments on the hedging side are dependent on the counterparties not failing. One can bet on their stand-alone resilience or the likelihood of government support.There is a wide variety of views on this subject. Some say bubbles are undetectable ex ante. Others argue that all relevant information is priced in at any point of time. Still others disagree and make their reputations and returns in part on anticipating trouble in the form of instability or lack of sustainability. My view is that while the current theory of financial market dynamics and the evolution of its risk profile is far from complete, there are useful indices and there are investors and analysts who pay particular attention to systemic risk. In the present instance, the rising aggregate debt-to-GDP ratio, the rising ratio of real estate prices to rental rates, and the abnormally low risk spreads, though not definitive, might have nudged investors into a more cautious posture. It therefore seems to me that investment and portfolio strategy should include a careful monitoring and evaluation of these external inputs (think of them as assessments and warnings from those with a track record of focusing on the stability of the macro environment). Action can be taken when the weight of the arguments is persuasive if not definitive.Depending on the liquidity properties of a portfolio, defensive action in the form of portfolio adjustment may be more or less restricted. As to tail hedging options, the costs tend to rise as the evidence of imbalance rises. Early anticipation, to the extent that it is possible, could be quite valuable. Another possibly complementary approach is the purchase of a tail hedge as a routine matter and not conditional on the correct or incorrect anticipation of rising systemic risk. Table 2 would give some idea of the amount one should be prepared to pay in terms of forgone annualized potential returns. In fairness, the figures in Tables 1 and 2 depend on the time horizon between shocks. For example, if one assumes clear sailing for 19 out of 20 years, followed by a shock, then Table 1 shifts to the figures below in Table 3.

Source: Author's CalulationsHypothetical Example for illustrative purposes only. The normal return range provided is intended for illustrative purposes and is not indicative of the past or future performance of any PIMCO product. Returns do not take into consideration fees or expenses.

Clearly judgment is required. My personal preference runs in the direction of assuming that shocks arrive relatively more frequently but that they vary in size. So I tend to start with Table 1 and mentally assign probabilities or likelihoods to the shocks of various sizes. A sensible approach could be to hedge against intermediate-level shocks. That may also help mitigate the impact of larger shocks and, depending on the annualized cost, could help increase the net annualized return potential in the upper half of the shock range. Of course it may also lower the annual average return potential in the lower half of the shock range too. Finally, purchasing a hedge has the added feature of lowering the differential between annual and long run average return potential and takes some of the pressure off the spending rule side of the equation for those institutions for which it is relevant and at times internally politically challenging.6Regulatory ReformA major focus of post-crisis regulatory reform has to do with detecting and limiting systemic risk. Since we are in mid-process, it is difficult to say how it will come out. One could take the position that this will eventually solve the problem of periodic rising systemic risk, and that we will return to the more comfortable world of relatively stationary risk without the periodic imbalances. In my view this is not a good bet for a variety of reasons. The historical evidence suggests that the problem is persistent and resistant to previous attempts to deal with it. Financial innovation will proceed along with regulatory arbitrage. The dynamic sources of systemic risk probably lie in the evolving network structure of financial markets. These are subjects of intense interest and active research. But it is unlikely that dramatically enhanced understanding will occur overnight and then be reflected in effective regulatory systems. Finally, the international dimensions and sources of systemic risk are increasingly material, and we have only limited if any demonstrated capacity for dealing with them.In short, it is possible that non-stationary systemic risk will become a feature of the past. But it is not a good guess at this stage. Even if it happens, in order to be sure that it has, we will need an extensive period of stability in order for confidence to rise.Liquidity and RiskThe present crisis has caused distress across a wide range of institutions with respect to the cash flow aspects of liquidity management. For relatively illiquid portfolios, an unanticipated shift in the parameters of the cash flow models (reduced distributions, increased capital calls and collateral or margin calls on various instruments) created an extreme lack of liquidity. The effects were multiple. Distressed sales of assets were required, exacerbating the negative returns. Borrowing was sometimes possible and needed, but that changes the leverage and risk profile of the overall portfolio. Further, with illiquid assets, and large shifts in asset prices, portfolios can become unbalanced with respect to asset class targets and are not rebalanced easily in the short run.In these dimensions, the lessons of the crisis have not been missed. The focus, I think it is fair to say, has been appropriately on the cash flow challenges that come from a combination of fractionally large illiquid overall positions and large systemic shocks. There are two other aspects of liquidity management that deserve attention. One is that illiquid investments place limits on the investor's ability to respond to early warnings of a shift in systemic risk of the type discussed above. It does so by limiting the ability to adjust the portfolio. Presumably this lack of adjustability is what is priced into the differential returns relative to liquid investments with similar underlying properties. Whether this pricing adequately reflects the cost of inflexibility in the presence of systemic instability is an open question.Even with an illiquid portfolio, there is still the tail hedge option. Careful analysis of the relative costs of these two possible responses to rising instability (tail hedge versus asset allocation adjustment) should be an important input to the basic liquidity choice for the portfolio. It may vary across institutions by size and sophistication so that there is not necessarily one right answer. The second important strategic issue is that liquid portfolios create investment opportunities in times of widespread distress. These result from distressed prices or downward overshoots in asset values, combined with the capacity to invest while others cannot or will not. Liquidity therefore has potentially significant option value that rises with systemic problems. Conceptually this value needs to be added to the return that is normally attributed to various classes of liquid assets in "normal" times. That will affect the relative value attractiveness of liquid and illiquid assets and hence influence the target asset allocation choices made by various classes of investors.So you have three underappreciated (at least until recently) virtues of liquidity: avoidance of cash flow distress, flexibility in adjusting to overvaluations and signs of rising systemic risk, and option value in the aftermath of crises or a major resetting of values and the trajectory of the system.Incentives and Performance EvaluationPeriodic systemic risk, with considerable fat-tailed uncertainty about the break points, presents significant challenges for performance evaluation and compensation. The following seem to me important in meeting this challenge.

Concluding ThoughtsPeriodic increases in systemic risk are likely to be a continuing recurring feature of the world we live in. It should influence investment strategy in a number of ways.

1 For a discussion of making policy decisions in uncertain environments with incomplete models, the reader may want to look at Growth Strategies and Dynamics: Insights from Country Experiences, Mohamed A. El-Erian and Michael Spence.

2 See for example, Carmen M. Reinhart and Kenneth Rogoff, This Time is Different: Eight Centuries of Financial Folly

3 The unpredictability of the break point is one of the challenges facing contrarians, be they investors or analysts.

4 Approaching the payout this way is a way to help hedge against volatility in the income stream. It may or may not be the most efficient way to achieve this result. Part of the purpose of this note is to explore various alternatives.

5 The ability to do this is clearly affected by the liquidity of the portfolio. I will return to the subject of liquidity in a subsequent section.

6 The arguments are familiar. Why are we spending 4% of the endowment when the aggregate annual return including donations has been running in the 15-20% range for more than a decade?

Past performance is not a guarantee or a reliable indicator of future results.   Derivatives may involve certain costs and risks such as liquidity, interest rate, market, credit, management and the risk that a position could not be closed when most advantageous. Investing in derivatives could lose more than the amount invested. There is no guarantee that these investment strategies will work under all market conditions and each investor should evaluate their ability to invest for a long-term especially during periods of downturn in the market. No representation is being made that any account, product, or strategy will or is likely to achieve profits, losses, or results similar to those shown.  Hypothetical or simulated performance results have several inherent limitations.  Unlike an actual performance record, simulated results do not represent actual performance and are generally prepared with the benefit of hindsight.  There are frequently sharp differences between simulated performance results and the actual results subsequently achieved by any particular account, product, or strategy.  In addition, since trades have not actually been executed, simulated results cannot account for the impact of certain market risks such as lack of liquidity.  There are numerous other factors related to the markets in general or the implementation of any specific investment strategy, which cannot be fully accounted for in the preparation of simulated results and all of which can adversely affect actual results.This material contains the current opinion of the author but not necessarily those of PIMCO and such opinions are subject to change without notice. This material has been distributed for informational purposes only and should not be considered as investment advice or a recommendation of any particular security, strategy or investment product. This material is not indicative of the past or future performance of any PIMCO product. No part of this material may be reproduced in any form, or referred to in any other publication, without express written permission. Pacific Investment Management Company LLC, 840 Newport Center Drive, Newport Beach, CA 92660, 800-387-4626. ©2009, PIMCO.