Written By: Bernard Abrahamsen
In 1952, American economist Harry Markowitz devised the meanvariance model of portfolio selection, which essentially defined “risk” as “historical volatility”. So, under the model, an asset with a history of high volatility is, by definition, “risky”.
Over the years, it gradually became more commonplace in academia and markets to use the words “risk” and “volatility” almost interchangeably. Therefore, a number of different models have been developed over time to measure volatility, including information ratio, tracking error and value-at-risk.
Simply put, it makes sense that if you buy a volatile asset instead of a stable one, you are at greater risk of losing your money in the short term. However, this view of risk is somewhat limited and one of these limitations was brutally exposed during the financial crisis of 2008.
“Toxic” US subprime mortgagebacked securities were very stable before the crisis, and therefore “safe” according to volatility models. However, when investors realised the credit quality of the bonds was worse than they previously thought, the assets fell substantially in price, ultimately inflicting a great deal of pain on their owners.
The essential point of this is: past volatility is no predictor of future volatility. Actual investment risks are too complicated to be summed up by a simple volatility statistic (which always ignores factors like credit quality). Investors who base their future investment decisions only on historical volatility statistics risk doing the equivalent of driving towards a cliff edge while looking in the rearview mirror.
Volatility measures tell investors to sell when the market is cheap. Worryingly, volatility measures can also actively dissuade investors from buying cheap assets. To explain, imagine Bond X is trading at 100% of its par value, and you as a portfolio manager consider it attractive.
However, before you get the opportunity to buy, the latest drama in the Eurozone debt crisis kicks off and corporate bonds sell-off across the board. Bond X now trades at 80% of its par value, but nothing concrete about that issuer or bond has actually changed.
Looking at it logically, if you liked it at 100%, you should be salivating at the chance to snag it at 80%. However, try telling that to your risk statistics. The simple effect of it falling 20% in one day has caused its historical volatility to rise. This means that it’s apparently now more risky and less attractive.
A classic example of a general market sell-off, just like the example above, occurred last summer. Those who were slaves to their volatility measures sold their corporate bonds, allowing others to buy them if they considered any to be attractive at their new prices. Since the sell-off, credit markets generally increased in price, meaning that those who purchased the most valuable securities enjoyed impressive performance figures.
It’s been demonstrated that everyday shoppers spend more when sales are on, but when the bond markets are effectively “on sale”, investors tend to hide, instead of seeking out potentially good deals. Financial markets are the only markets that operate this irrationally, but happily, rational investors are able to benefit.
At M&G, how do we look at risk? We look at risk as the probability of a permanent loss of capital (usually triggered by a bond default). Although we produce the usual array of popular volatility statistics, our fund managers have never taken much guidance from them, because we do not believe they will help us avoid defaults.
So, how do we mitigate potential defaults? We attempt to do so through diversification, employing a team of 86 credit analysts (a team we believe to be one of the largest in the city), and a dedicated team of restructuring specialists who can step in when a security becomes distressed.
But that isn’t the entire story. Our portfolio managers are very mindful of the fact that they may have got their positioning wrong and they aim to ensure that, if the market moves against them, the adverse effects on their portfolio will not be substantial. Richard Ryan, manager of our total return credit product, the M&G Alpha Opportunities Fund, does this via a process called scenario analysis – also known as “stress testing”.
The benefits of scenario analysis. Scenario analysis involves mathematically imposing various adverse market developments (or scenarios) on a portfolio and seeing how well the positions hold up. For example, the tests show the team behind the M&G Alpha Opportunities Fund how much they’d lose if credit spreads were to double tomorrow. Similarly, they run tests for changes in bond yields, currency values and many other potential developments.