The quest for effective risk diversification has been on the top of the investment community agenda for many decades. The complexity of the matter lies on the fact that, in reality, there is a very limited number of safe haven assets in the investment universe. Diversification is mostly needed during times of elevated turbulence, as investors seek extra protection. During periods of regime shifts in volatility (abrupt and violent increases) asset correlations tend to converge. This convergence makes diversification practically impossible at a time that it is mostly needed, as it difficult to find and invest in uncorrelated, or, even, negatively correlated assets.
In practice, however, there are asset classes that exhibit a lower risk profile and are, traditionally, considered as safe haven investments. Government bonds, money market instruments, gold, as well as a limited number of currencies, such as the Swiss Franc or the Japanese Yen, have long been considered as defensive asset classes that tend to outperform other asset classes during turbulent periods. Gold, in particular, is considered as the ultimate safe investment.
But has gold behaved as a safe haven investment lately? Graph 1 plots the gold price (Generic 1st Gold, 100 oz) against the S&P 500 index. Moreover, Graph 2 plots the S&P 500 index against the U.S. 10-year Treasury yield, another defensive asset. It is evident that gold’s reaction during the August global financial market turmoil has been very muted, at least relative to the reaction of U.S. yields.
The truth is that since 2008 the correlation of equities and gold has been practically zero, as opposed to the negative correlation (-45%) that was recorded during the 2008 U.S. recession, as defined and dated by the National Bureau of Economic Research (NBER). The picture is similar for pair wise correlation against U.S. 10-year Treasury yields; during the whole sample period there is nearly zero correlation, and only during a period of extreme stress the correlation becomes positive.
It is well known that implied volatility exhibits negative correlation versus equities. The CBOE VIX index is plotted against the S&P 500 index to portray this relation (Graph 3). How is gold faring against an asset that provides effective protection against a risky asset? Interestingly, the pattern remains the same; almost zero correlation to changes in the VIX index during the 2008-2015 period and positive correlation during the 2008 U.S. recession (48%). The picture is similar, if we estimate the pair wise correlation against the VIX, instead of the change.
The underlying asset correlation structure is dynamic, meaning that correlations change through time, and gold does not exhibit safe haven characteristics constantly. This can be easily seen also from the 26-week and 52-week rolling correlations (Graph 4 and 5).
It is interesting to note that for a large part of 2015, the correlation to equities was positive (26-week correlation) and only recently it has turned negative, reflecting the turmoil in financial markets. Even the 52-week correlation became less pronounced during the course of the year.
All in all, one needs to bear in mind that apart from the systematic characteristics that an asset class exhibits, its performance is also driven by idiosyncratic elements. The identification and effective differentiation of the two drivers can be very difficult at times. Gold seems to behave as a safe haven investment (systematic part) during periods of extreme stress.
Is gold an ‘all weather’ investment?