Statisticians spend a great deal of time debunking the notion that “correlation implies causation.” The idea being that just because two data series are correlated does not necessarily mean one causes the other, or vice versa. This is not a new topic to readers of Abnormal Returns. There has been a great deal of chatter in the financial media and blogosphere about the long march towards higher correlations.
Josh Brown at The Reformed Broker had a nice post yesterday noting the “pandemic” that is global correlations. He writes:
This doesn’t mean to ignore geography when planning a portfolio. No, the takeaway is not to rely on geography – because when the going gets rough, it’s all one market. On top of that, even in periods of calm we are headed there anyway – globalization of trade means globalization of stocks in the long run.
In regards to the financial markets we should also recognize that correlation is not destiny. For instance everyone recognizes that small cap and large cap US stocks are highly correlated. Since January 2001 through August 2011 the Russell 2000 and S&P have a correlation of 0.89 based on monthly returns. Pretty high by most standards. Over this same time period $IWM returned 71% versus 12% for $SPY. Although these two asset classes are clearly correlated their returns can (and will) diverge over time. In short correlation is not destiny.
Another way to think about this is via a thought experiment. Let’s take the daily returns of the S&P 500 for the past year and simply add 1 basis point to each day’s returns. $SPY over this time period returned 7.7%. Our modified S&P 500 would have returned 10.4%. However the correlation between these two return series is by definition 1.00! Two highly correlated assets can see their returns diverge over time.
The bottom line is that correlation is an important concept to understand when it comes to asset class returns. There is no doubt that stocks and sectors with the US and global markets have been getting more correlated over time. This does not mean that certain stocks, sectors and markets cannot outperform (or underperform). Higher correlations simply mean that the relative performance along the way is going to get masked by overall market movements. Correlation does not imply causation, nor does correlation imply destiny either.