Alternative Investment Management Association Representing alternative asset managers globally
Ineichen Research & Management
The idea of diversification is very old. Supposedly, it’s the only free lunch. The idea has entered the English language as “don’t put all your eggs in one basket”. It has entered investment management orthodoxy via Harry Markowitz and modern portfolio theory (MPT). The idea of spreading risk by diversifying risk is much older than MPT though. The Oxford Dictionary traces the “eggs in the basket” idea to 1710, referencing an Italian source of proverbs from 1662. The idea of diversifying risk can be traced even further. The Talmud suggests: “Let every man divide his money into three parts, and invest a third in land, a third in business, and a third let him keep in reserve.”
This makes it somewhat difficult to date the origin of the idea of diversification by spreading the risks. Suffice to say, the idea of diversifying risk could be thousands of years old, potentially as old as civilisation itself; or if not, at least as old as men going about their business affairs. (The concept of diversification is of course much older. The biosphere for example has been diverse for billions of years. Diversity is life’s ultimate survival mechanism. Every now and then there is mass extinction on this planet with the more complex life forms being wiped out. The biosphere regenerates for surviving simple life to evolve into something more complex. And then there is mass extinction again. The parallel to the wealth of nations is clearly recognisable.) The funny thing about the Talmud quote above is that it suggests both a “margin of safety” as well as equal weighting. Equal weighting is not yet a trend in investment management but it has been discussed in some journals for a while now. The main argument for equal weighting is that we just know far too little about future returns, future volatilities, and future correlations for mean-variance optimisation to have any value. Furthermore, the assumptions behind MPT are onerous and silly; onerous because the input variables for many viable investments are not available; silly because most of the assumptions behind MPT have turned out to be false, misleading, or dangerous, or a combination thereof.
Potentially the idea of risk parity is a trend. Risk parity is a strategy where the allocations to various asset classes are not equal but the allocations are determined by the risk of the various asset classes. If risk is defined as volatility, an asset allocation of 50:50 between equities and bonds can result in a risk allocation of 90:10. This means moving from traditional asset allocation towards risk allocation (or risk parity), generally speaking, results in a smaller allocation to equities. (One idea associated with risk parity is to have each 25% of equity risk, interest rate risk, credit risk, and inflation risk, and then fill the first three buckets with hedge-fund-type risks, rather than plain long-only risks.)
Replacing a quantitative optimisation which is based on a silly measure for risk with another quantitative optimisation which is based on the same silly measure for risk is, well, silly. Given that the current regulatory trend is equity-unfriendly, and government bond-friendly, the advent of a new monster equity bull market — potentially — is exactly then, when the equity allocations have been reduced to multi-generational lows. MPT has its uses though; it’s a great marketing tool. MPT works very well for selling financial innovations, for example ETFs or baskets of subprime mortgages. MPT works very well for selling “alternatives” too. Long-only investments in the equity market were once an alternative investment; an alternative to bonds. Strong performance in equities post 1982 in combination with the diversification argument were the main reasons for institutional investors loading up on equities in the most amazing bull market of the 1980s and 1990s.
In the fantasy/model-world of MPT, combining equities and bonds makes perfect sense because the correlation coefficient is less than one. (Except when it really matters of course; in a major panic and sell-off, for example.) However, a historic perspective suggests that in real total return terms, long-only investments in equities and bonds move more or less in tandem. More precisely, in times of inflation equities and bonds decline in tandem while in times of disinflation equities and bonds rise in tandem. It is wiser; therefore, to operate in an “asymmetric returns” fashion. This means allocations should be a function of the opportunity set, rather than a combination of statistical variables entered into a faulty econometric supposed-optimiser.
Sharpe ratios are marketing gimmicks and volatility is — and this is putting it very nicely — an incomplete measure for risk. Risk is not perceived as volatility. Institutional investors do not perceive volatility as risk, private investors do not perceive volatility as risk; even relative return managers do not perceive volatility as risk. Losing large chunks of one’s capital, on the other hand, is more like it. Recent financial history has shown that at the end of the day it is losses that matter most. Risk, therefore, becomes the probability of what matters most, i.e., losses and/or non-survival. Another way of putting it is defining risk as “exposure to accidents”. The problem with accidents in finance is that they do not seem to be spread out evenly over time; quite to the contrary, they seem clustered. Take sovereign defaults for example: there is indeed such a thing as a domino-effect. Sovereign defaults do not occur one at a time evenly spread over time. They are contagious and occur in a bundled fashion. As risk managers wishing to survive this episode of finance, we need to become students of history, rather than disciples of MPT (Modern Portfolio Theory). (We need — unfortunately — also to become students of back-against-the-wall-welfare-state-politics too.)
One aspect of accidents is that they are said to be surprising by definition: if they were predictable, they wouldn’t occur. This logic might apply to slipping on a banana skin. However, this logic doesn’t necessarily apply to finance. The introduction of the euro for example is an accident that is unfolding as we speak. It just took a while until it became apparent to everyone; well, nearly everyone. An investor has the choice to participate in the accident or hedge or invest elsewhere. Japan is not yet an accident but is one in the making due to its current debt levels and unfavourable demographic trends.
It goes without saying that losses matter more to owners of capital than for the agents managing the capital. (The agents do care about their own capital though.) The idea of marrying the owner’s capital with that of the managing agent is still one of the most powerful attributes and differentiating factors of the hedge funds industry. Warren Buffett surely hit the proverbial nail on its head: “After all, who ever washes a rental car?”
The reason for mentioning accidents is our belief that risk management is a thought process rather than a quantitative exercise. Risk measurement, one could argue, is a quantitative exercise. If risk management is indeed a qualitative exercise where thoughtfulness matters, it is healthy to think about what could go wrong, even if that leads us away from MPT, VaR, and alphas and betas for a moment. (The Tunguska explosion in Russia in 1908, a cosmic event that released, according to one estimate, the equivalent of roughly 1,000 WWII atomic bombs, is a case in point in that regard. In the event of such an event hitting, say, London, Sharpe ratios do not matter that much. Cosmic impacts of the Tunguska variety are expected roughly every 100 years. In sporting parlance, we’re in “overtime”.) Another reason to think a bit out-of-the-box when contemplating risk is that sometimes Murphy’s Law applies. Sometimes it happens that you have a weak economy and are hit by an earthquake and by a tsunami and have a nuclear disaster all at the same time. Accidents happen and sometimes Murphy’s Law does indeed apply.
Below we summarise some aspects of diversification, or, more precisely, the sensitivity of absolute return strategies to accidents. The table shows how some of the hedge fund strategies perform and respond to financial accidents. We focus herein on equities, as it is equities that have been the largest contributing factor to the risk of institutional portfolios by a wide margin in the past. (All this builds on our research effort over the past 20+ years which — in the tinniest of nutshells — states that large losses kill the rate at which capital compounds and compounding capital negatively is not good for one’s financial and mental health.) The table below is based on occurrences where the MSCI World lost more than 7% of its value within one, two, three, or four months on a month-end basis. The worst return from the four returns was chosen. The negative equities event can then be compared to another asset or strategy.
Source: Ineichen Research & Management
Recent hedge fund performance is close to a low point. It is fair to say that hedge funds have not shot the lights out recently. There is most likely an element of disappointment, especially if expectations were calibrated towards hedge funds shooting the lights out. However, hedge funds have done well in relative return terms; especially in difficult as well as in low-real-return environments. For investors who bought into the absolute returns idea around 2006, the first five-year experience is a sobering one. They didn’t get what they had signed up for. They got superior relative returns instead, but that was not what they had signed up for. However, it is this superior relative performance that is causing more and more institutional investors to switch their long-only mandates for mandates in which risk is controlled actively. One of the ironies in this regard is that to a certain extent, relative returns matter to the absolute returns industry as well.
Back to Listing