This is based on our experience that risk and return form a continuous spectrum, rather than a choice between alternatives, as conventionally offered by other advisers to investors. While others believe that high return = high risk, our experience show that high return = low risk.
The more one is invested in likely winners, the less one has left to invest in likely losers. The faster one runs forward, the less time is lost overall when one suffers the inevitable occasional stumble.
Others seek comfort in the spurious precision of hindsight performance and market-weighting statistics, using technology to match their asset allocation ever more closely to the centre of the investing herd.
This crisis discredited such lemming-like behaviour as never before, because it caused more damage to the average portfolio than any previous setback in a generation. Between October 2007 and February 2009 equities fell -55% worldwide while 10 year government bonds rose only 15% and other bonds also suffered.
However, our strategies, which we call RouteMaps, use technology to quantify the odds of predicting winners and losers on as wide as possible a range of different investing strategies.
During the crisis diversification did not work. Equity markets re-coupled as the world economy suffered an unprecedented inventory correction. Most hedge fund strategies failed to hedge as leverage maximised their downside exposure to liquidity pressures and short positions rose while long positions fell.
Commodities fell with equities, rather than rising, as in the last great commodities boom a generation ago. Corporate and mortgage bonds were exposed as equity proxies. A year later private equity markets have still to re-establish clearing prices.
Virtually the only assets to hold or increase in value were gold, Treasury Bills and bonds, yen and Swiss francs.
These are all are low-yield, no-yield investments. Diversification into such safety-first alternatives would have been an enormous opportunity cost for anyone who did it on a long-term basis.
Worse still, because this happened in the most sudden stock market downturn since the Depression, this meant that most of the popular statistical risk analysis tools let their users down, including correlation, Sharpe ratios, value at risk and volatility.
Short-term risk
Conventional analysis views risk as a short-term danger, and defines it as the volatility of monthly or annual returns. However this is inappropriate for the long-term nature of most portfolios.
There are three reasons for holding liquid assets such as cash, namely transactions, emergencies and speculation on market setbacks. The first is minimal for most investors, the second is a balance between severity and frequency and the third is mainly short term.
While others believe that high return = high risk, our RouteMaps show that high return = low risk.
The last bear market was an imploding real estate bubble climaxing in a systemic liquidity crisis when the authorities failed to nationalise Lehman.
Liquidity crises are essentially short-term phenomena, but that was hard to see because this one was part of a bigger bear market. To illustrate the point the chart below compares developments since L-Day with the two other global liquidity crises within living memory.
The comparable crises are the Wall Street Crash of October 1987, where so-called portfolio insurance programmes generated a cascade of stop-loss selling and the LTCM crisis of September 1998, where the implosion of this highly-leveraged hedge fund threatened a global crisis in investment banking.
Global markets had fully recouped their losses within a year. Now two-thirds are recouped in a year.
Therefore the appropriate risk management strategy for unanticipated temporary setbacks is to hold a small proportion in cash to meet transactional and precautionary needs – and do so in a separate account for clarity of purpose.
For anticipated setbacks the right course of action is to temporarily increase safety-first assets substantially – if one is smart enough.
New statistical stupidity
Since the crisis, risk modellers have been busy recalibrating their algorithms. Some have applied advanced mathematical techniques to incorporate the greater likelihood of extreme events. Others do so by default, merely by advancing rolling three, five or even ten year monthly time-series to include more post-crisis data and less pre-crisis data.
This would improve the predictive value of the statistics if the extraordinary event was likely to reoccur, but that is not likely. There have only been three systematic banking crises in the United States over a century, namely 1907, 1930-34 and 2008.
In real life, actual risk is often inversely related to perceived risk, so the odds of another such severe crisis have probably fallen – until the next generation.
This is indeed likely because democratic capitalism is an excellent system for responding to change.
After failing so badly in their duty to maintain the monetary system, the authorities can be relied upon to give top priority to fixing the problem they caused and preventing its reoccurrence.
Simultaneously, business, consumers and investors can be relied upon to rearrange their affairs to respond to the recent past – at least till it fades from their memories.
The timescale can be measured not in months, as those who fear another crisis now do, nor even in years, but a decade or a generation.
Slow response
Politicians will be busy for years. It took till 1911 for the Federal Reserve Act to be passed in response to the 1907 Bankers Panic. It took from 1933 to 1940 before the all the laws to stop another 1929 were passed.
Similarly, it will be a generation before the lessons are forgotten. The intervals between speculative euphoria on Wall Street were 22 years after 1907, 37 years after 1929 and 34 years after 1966. So far, not even a decade has passed since the Millennium bubble.
Some things even take two generations. It took Congress 66 years to repeal the Glass-Steagall Act, which separated commercial and investment banking in 1934, before they were completely free to repeat the same mistakes.
While other statistical systems are adapting to include the recent extraordinary past, our RouteMaps are ‘cheating’ by largely ignoring it. This is most important for the sentiment dimension, which measures human behaviour directly. Deliberately, our back-testing ends in August 2008, just before the Lehman crisis.
Subsequently, those who tend to bet right, got it wrong and vice versa. The sentiment strategy had performed brilliantly up to that point but was atrocious for the year afterwards.
Essentially, the shrewd took it for granted that the monetary authorities would do their job of maintaining the stability of the payments system by nationalising bust banks, as they had done ever since 1934.
Since the shrewd collectively got it wrong just this once, it seems stupid to re-engineer risk control systems.
It is unlikely that the shrewd will get it so badly wrong again. Investors RouteMap will not be re-engineered.
HIGH REWARD = LOW RISK
Our strategy, the Investors’ Routemap, is an odds-based betting system, in principle similar to the form book for horse racing.
It assesses a large range of investment strategies. None of these can have high odds of working consistently and successfully for ever, otherwise their comparative advantage would be arbitraged away.
However, many strategies can be modestly successful. By combining these, Investors RouteMap has been able to achieve an exponential improvement in performance.
Actual risk is often inversely related to perceived risk, so the odds of another such severe crisis have probably fallen.
It seems stupid to re-engineer risk management systems to bet that the shrewd will get it so badly wrong again.
