We have all been taught about the merits of diversification in investments. It is a variation of the old adage, "Don't put all your eggs in one basket."
Indeed, professional investment managers are trained to develop portfolios according to the tenets of Modern Portfolio Theory (MPT). MPT traces its roots to the work of Harry Markowitz and his seminal writings on "Portfolio Selection." In his pioneering research, Markowitz was able to demonstrate the mathematical basis for diversification.
Essentially, Markowitz showed that selecting assets that have a positive expected return but exhibit low or (preferably) negative correlation to one another produces a combined portfolio that retains the positive expected return properties, but with lowered risk (as defined by variance).
Theoretically, this result arises due to the presence of at least two major sources of risk: nonsystematic (or unique) risk and systematic (or market) risk. While it is very difficult to eliminate market risk, it is possible to reduce the risks associated with unique investment assets. By combining investment assets that are subject to certain specific, unique risks with other investment assets that are subject to other unique risks, it may be possible to reduce the overall risk of the combined portfolio.
For the past several decades, this has been the mantra to which all investment managers adhered. Unfortunately, recent experiences in the capital markets have led both academics and professional investment practitioners to rethink portfolio construction. With the increasing interconnectedness of global markets and investment pools, we have seen that correlation structures among various investment assets are not always stable.
In fact, assets that typically exhibit low correlation with one another can dramatically change direction and begin exhibiting increased correlation during periods of market distress. The increased correlation leads to a reduction in the power of diversification and thus to increased risk in the overall portfolio. Unfortunately, this upward shift in correlation happens at exactly the time when an investor needs correlation the most: market distress.
As a result, investment managers need to be exceedingly careful in constructing portfolios that are able to withstand the dynamic nature of correlations, especially as the market experiences large disturbances. These "disturbances" are becoming much more commonplace: the Asian currency crisis of 1997, failure of the major hedge fund "Long Term Capital Management" in 1998, the burst of the "dot-com" bubble in 2000/2001, the terrorist attacks of 2001, the burst of the real estate bubble in 2007/2008, and the credit crisis of 2008/2009. In nearly every case, correlation structures among various assets increased at precisely the time when investors needed protection the most.
The best portfolio construction techniques have an appreciation for the fact that correlation structures may change during different "states of the world" or regimes. By incorporating these state-dependent correlation structures into portfolio design and optimization, investment managers can move to better protect portfolios during times of market distress.