A recent research piece by Oppenheimer talks about how the correlation within the stock market has reached a low after being elevated for the past two years. The piece argues that the recent decline in correlation coupled with their forecast that correlation is likely to remain low in the future means that it is a good time to invest in active strategies because that is the environment in which they can create value.
We think this is not a good way to think about the issue and we use some analysis from the SymmetricInfo analytics tool to delve into the subject.
First, we look at a measure of correlation in the US equity markets going back to 1925. We use a time series provided in the SymmetricInfo tool (12 month rolling correlation between sector returns and the overall market). The chart below shows that the correlation is typically elevated during significant macro events. One thing that pops out is that while the correlation over this financial crisis has been higher than any period until 1987, it appears that comparably high equity correlation was not as unusual before 1987 (infact it is the last 20 years of relatively low correlation that seems unusual). Therefore, it is difficult to argue emphatically that the period of low correlation from 1988 to 2007 is necessarily the norm and it is entirely possible that investors might have to deal with more frequent spikes in correlation going forward.
click to enlarge
The periods of higher correlation typically occur when there are worries about growth in the economy. If we look at all the periods where the average correlation has increased over a 12 month period to over 80%, it is not surprising to find that those periods have coincided with poor stock market returns.
There are exceptions to this generalization though. For example, during the rally of April 2009 or the Iranian oil embargo in 1979, correlations were very high, but the equity market rallied sharply.
Given that for most part higher correlations are often caused by concerns about growth (which tends to drive poor equity market returns) it is not surprising that when correlations eventually decline the stock market has tended to rally at the same time.
In this context, what has happened to equities recently is relatively straightforward. There were concerns about economic growth because of linkages between the US economy and Europe which caused the equity market to decline/become highly correlated. Upon the release of better than expected economic numbers in the US this year, the stock market has rallied and correlation has declined (for the time being at least).
The key argument being made by Oppenheimer is that it is hard for active equity managers to make money when correlation between stocks is high and that investors should take this opportunity of low correlation to invest in active managers. However, active managers did quite well after April 2009, when correlations were also very high, but the equity market rallied sharply. They then did poorly in 2011 when equity markets moved sideways but correlations remained high. Therefore the more plausible and simpler explanation of what drives active manager returns is not that they benefit from lower correlation, but that they are simply LONG the market in one way or another.
Even some hedge fund strategies that are arguably "market neutral" are not immune from actually being long the market in some form. A classic example is merger arbitrage. One can very closely replicate the risk characteristics of merger arbitrage strategies by simulating the PNL of going long the company being acquired and short the acquirer (for a stock deal) for every merger announcement after the deal is announced and holding onto this trade until the deal closes or collapses. When the market tends to drop by a large amount acquirers tend to rethink their offers, which tends to coincide with a larger than usual number of deals collapsing, which causes the strategy described to lose money. Therefore being invested in a merger arb strategy has similar risk characteristics to being short an out of the money put option on the equity market. The point is that if investors have to count upon low correlation for active managers to create value, then that is the same as saying that they generally have to count upon rising equity markets for an active manager to create value. It is easy to beat the market when it is rising in value (just use more leverage, buy small caps, sell some out of the money puts or use the Merger Arb strategy we just described). The fact that most managers can't perform when markets are declining (i.e correlation is high) means they are basically just long the market in one way or the other and aren't actually creating much diversification or value for investors.