Risk modelling, for example, Value-at-Risk, measures and quantifies the level of financial risk within a firm or investment portfolio over a specific time frame. We use Value at risk to measure and control the level of risk which our firms undertake. But consistently the answer we get is wrong, and near impossible events happen quite frequently. Other risk measures also underestimate the real risk of ruin. Why are there so many crashes which take us by surprise – again and again?
Black Swans are the problem. Established risk modelling looks at past data to determine the future’s danger. We might calculate the likelihood of ruin as 20-Sigma and therefore, nearly impossible but yet unexpected event seem to happen quite frequently – even every few years.
Some risk is acceptable, indeed necessary. We know that without risk there can be no expectation of reward. The problem is financial theory assumes all the risk is behind you. Why is the worst necessarily behind you? Isn’t it much more likely to be ahead of you? In reality, the maximum risk – say drawdown – is most likely to be in the future, not the past. It is difficult or impossible to estimate statistically using the small sample of history and assume it is representative of the future. We actually have to calculate the future. Added to that is the likelihood is the crash event is something we have never seen before and therefore not in our data at all – Talab’s Black Swan event.
For a moment, let us agree that it is a fast fall of 20% or more. How many of these have we had in my lifetime? I was born in 1954. There have been 57 such crashes in four of the world’s major indexes (the US’ S&P 500, the UK’s FTSE All Share, Germany’s DAX, and Japan’s TOPIX) according to Christian Mueller-Glissmann, a strategist at Goldman Sachs. Does this mean there is a crash almost every year? Well no. Many crashes occur in many major markets at the same time. The pattern is not consistent. Mueller-Glissmann notes that the drawdowns are beginning to happen in concert. Crashes in the 1960s, 1970s, and early 1980s were relatively disconnected. In recent instances, such as the tech bubble burst and the financial crisis, the precipitous drops have been more coordinated.
Crashes occur more frequently than looking at past data indicates. There is more and more contagion. They are tending to happen globally.