Nothing is more unpredictable than the fortunes of the euro

In 2004 a tsunami hit southeast and south Asia. Unimaginable destruction and the deaths of nearly 30,000 people followed. In most areas the extent of the loss was somewhat discernible. In others however it was unknown. And nowhere was this more evident than among the Sentinelese, a Paleolithic community, who live off the coasts of Andaman and Nicobar islands.

There are no clear estimates how many survived. In an effort to get an idea of what happened the government of India sent out a reconnaissance chopper. In a little-known photo from those days, we see a young Sentinelese hunter threatening the helicopter with his arrows, we may laugh, or pity him at his seeming incomprehension regarding the technological gulf between the helicopter and the bow and arrow. Strangely, today's market participants are no different, on some levels. The most interesting force in the market today is correlations amongst asset classes.

What does one really mean by correlation? It is easy to provide a mathematical definition. Correlation between two asset classes is often defined as what is called Pearson's correlation. It is roughly the average of how the product of deviations from the mean of each asset moves when normalised by the product of their respective standard deviation. Reading this a layperson might say: "In English, please!" While an intuitive understanding might be difficult, the numerical results are easier to interpret. Simply, a correlation of one means a perfect correlation in the same direction and a correlation of negative one means a correlation in the opposite direction. This definition simplifies the world as linear. Rarely do life and the markets move with such precision, such linearity. Yet, because it is easy to comprehend and compute such ad hoc ideas of correlations continue to inform our perspective.

Continued chaos

Most importantly, correlations aren't stable themselves. If two assets are correlated at, say +0.2, these numbers can change depending on how volatile the markets are, how long a time series of data you use to estimate the correlation. Today's markets are deeply non-linear, volatile and calibrating the "correct" correlation such that it can be reliably used is more a matter of judgment rather than number crunching. So, while more quantitative tools might lull us into a sense of safety with quoted correlation numbers, it is the market prices we must watch. For prices in a liquid market contain within them estimates of what is known and more importantly the risk premium from unknown and unknowable events. And nothing is more unknowable in the markets these days than the fortunes of the euro in the coming weeks. To some the demise of the euro is a foregone conclusion, while for others it is here to stay.

For markets however, the events have allowed for correlated assets to move in manners unanticipated. The perils of such highly correlated moves are obvious. Investors who seek to hedge away risks by investing across asset classes are increasingly scampering to find assets that move independently, are liquid and reliably stable. As the prospect of continued chaos rises there is increasing clamour to depreciate the euro. There are large-scale asset investment strategies that have pegged themselves onto the fate of the euro — and subsequently, there is a rise in herd behaviour and correlation. The dark side of deleveraging seems to be a rise in correlation.

The columnist works for a major European investment bank in New York City. All opinions are personal and don't reflect any institutional perspectives.