A common characteristic of most investors and traders is to always be on the lookout for patterns and connections between various asset classes. Whenever one correlated asset confirms the move in another it adds a layer of confidence to an investor’s thesis. One long-held example is the Dow Transports as a leading or coincident indicator for the broader market. For decades now, many investors have followed the transports for confirmation of the broader market moves. If the transports — which move all of the physical goods in the economy — rally, it suggests that the broader market will be strong, while periods when the transports start to roll over are read as a signal that there’s an underlying weakness in the economy.
As the US economy has become more service and digital-oriented in nature, there has been a valid argument made that the transports have lost some of their importance as an indicator of the broader economy. Along these lines, we have suggested that rather than transports, semiconductors may represent this century’s ‘transports’ as they are a part of just about everything in this digital age. Whether you agree with this or not isn’t important, but the important takeaway is that just because two asset classes have been highly correlated in the past doesn’t mean that they will remain that way in the future. It’s one thing to recognize a correlation between two asset classes, but it’s much more important to understand why they are correlated and be on the lookout for factors that may change the status quo in the future.
One example of a radical change in a relationship between two asset classes is the interaction between the relative strength of growth and value stocks versus the slope of the yield curve. From 2002 through 2011, the two were closely correlated. As the curve flattened in the early part of this century, growth stocks underperformed value by a wide margin (falling blue line). Then in mid-2007, as the curve steepened and came out from inverted territory, growth stocks started to rip higher relative to value. Beginning in 2009, though, the curve stopped steepening and the relative strength of growth relative to value stalled out. The two series were so closely joined at the hip during this ten-year stretch that the correlation coefficient between the two was +0.82, which is indicative of two series moving in lockstep with each other.
If the paths of the yield curve and the relative strength of growth versus value couldn’t be separated from 2002 through 2011, the relationship soured in 2012 when the two came down with a case of the ten-year itch. At that point, they couldn’t separate fast enough. The chart below shows the same two series from the start of 2012 through the present. Now, when one goes up the other goes down and vice versa, as the paths are nearly exact opposites. In fact, in the nearly eight years since 2012, the correlation between the two is -0.90.
In the chart below we have shown the two series over the entire time period spanning 2002 through 2019. The non-shaded area represents the period covered in the first chart, while the shaded area covers the second period. Right around the time where the shaded period starts is when the positive correlation turned on a dime, and beginning in 2013 when the curve started to flatten, investors who were still hanging on to the idea that a flatter yield curve was a green light for value stocks, saw what turned out to be an extended period of misery relative to the performance of growth stocks. In the words of Intel Founder Andy Grove, “Adapt or Die.” Start a two-week free trial to Bespoke Institutional to access our interactive Chart Scanner and much more.