Dividend Aristocrat Correlation Pairs
On June 27th, 2014, I published an article discussing pairing Dividend Aristocrats based on their correlations. The idea behind the article was to further reduce volatility in Dividend Aristocrat positions by pairing together the most uncorrelated Dividend Aristocrats.
Back in June of 2014, the most uncorrelated pair was Medtronic (MDT) and McDonald’s (MCD).The pair had a correlation of just 0.12; which is very low.
In the 2 year period from July 2014 through June 2016, the low correlation Dividend Aristocrat pair reduced volatility:
- Medtronic annualized standard deviation was 19.9%
- McDonald’s annualized standard deviation was 17.2%
- Pair annualized standard deviation was 15.8%
- S&P 500 annualized standard deviation was 14.8%
Returns were also very high:
- Pair annualized total returns (without rebalancing) were 15.5%
- S&P 500 annualized total returns were 5.2%
The returns portion can be attributed to luck; the purpose of pairing off Dividend Aristocrats is not to maximize returns; it is to reduce volatility.
What was interesting about the Medtronic/McDonald’s pair is how intuitively these businesses are ‘opposites’. Medtronic’s largest segment sells devices that treat cardiac and vascular symptoms.
I’m not a health food advocate – I eat at McDonald’s myself from time to time. But even ardent McDonald’s fans will likely agree that the company’s burger/fries/soda meal is far from ‘heart healthy’. One could even argue that McDonald’s is indirectly driving Medtronic’s sales.
Current Pairs Methodology
The 1 year correlation matrix for all 50 Dividend Aristocrats from August of 2015 through July of 2016 can be downloaded here (the spreadsheet provides a much clearer picture than the image below):
The top 10 lowest correlation pairs are listed below:
- 0560 Consolidated Edison (ED) and Franklin Resources (BEN)
- 0599 Consolidated Edison and Nucor Steel (NUE)
- 0685 Consolidated Edison and S&P Global (SPGI)
- 0696 Consolidated Edison and AbbVie (ABBV)
- 0698 Hormel (HRL) and AbbVie
- 0702 Consolidated Edison and Cardinal Health (CAH)
- 0707 Consolidated Edison and Dover (DOV)
- 0849 Consolidated Edison and Pentair (PNR)
- 0922 Consolidated Edison and V.F. Corp. (VFC)
- 0972 Consolidated Edison and T. Rowe Price Group (TROW)
Analyzing The Pairs
What immediately stands out about the 10 lowest correlation Dividend Aristocrat pairs is that 9 of them include Consolidated Edison.
What makes Consolidated Edison different from the rest of the Dividend Aristocrats?
It is a utility. Utilities are more correlated with interest rate movement than other businesses. That’s because their value is based mostly on current dividend yield; future growth plays only a minor part in the value of the business.
Consolidated Edison will likely continue to grow at around the pace of inflation, or maybe 1 or 2 percentage points faster if the company’s management does very well. This makes Consolidated Edison more like a bond than a stock for valuation purposes.
Consolidated Edison has historically been a very low volatility stock – and it has a low correlation to other Dividend Aristocrats. This makes adding Consolidated Edison at the right price a good idea for investors looking to reduce volatility in their portfolio.
The businesses that have the lowest correlation with Consolidated Edison appear to be (based on qualitative analysis) the businesses that have the most exposure to general stock market moves instead of interest rates specifically. Asset managers (like TROW and BEN) do well when the market goes up (as it causes AUM to expand and more investors to want to put money into the market), and poorly when the market goes down.
Industrial businesses and steel manufacturers (NUE, PNR, DOV) are also closely tied to the performance of the economy instead of interest rates. The global economy appears to be driving stock price movement in the health care sector right now as well.
One of These Is Not Like the Others
The lone pair that does not include Consolidated Edison is Hormel and AbbVie.
The Hormel/AbbVie pairing bears a striking resemblance to the McDonald’s/Medtronic pairing of 2 years ago.
Hormel and McDonald’s both sell food products. While Hormel does have health-centered options (Justin’s peanut butter, MuscleMilk, Applegate Organics), the company has many brands that would not appeal to health conscious consumers, like Spam and Dinty Moore.
Medtronic and AbbVie are both in the health care sector. AbbVie’s primary revenue driver is Humira. Humira is used to treat a variety of illnesses, including:Ulcerative Colitis, Rheumatoid Arthritis, Crohn’s Disease, and Plaque Psoriasis.
One of the causes of ulcerative colitis is a high intake of linoleic acid. Red meat is high in linoleic acid. While it’s a bit of a stretch, again we can see a bit of link between the meat products Hormel sells and AbbVie’s Humira profits.
Final Thoughts
The goal of rational investors is to maximize risk adjusted returns, subject to constraints like need for current income and risk tolerance.
Investing in high quality businesses like Dividend Aristocrats that have low correlations to one another can help investors to reduce volatility in their portfolios.
Analyzing the correlation of Dividend Aristocrats to each other also provides clues as to what is driving returns for specific companies.
Consolidated Edison has a low correlation with most other Dividend Aristocrats, which makes it a compelling addition to portfolios seeking lower volatility. Consolidated Edison is currently trading for a price-to-earnings ratio near 20, which is very high for a low growth company. Now may not be the best time to add to or initiate a position in the company.
The other interesting thing about analyzing correlations is seeing the low correlation between food businesses and health care businesses. A ‘food and medicine’ portfolio would provide meaningful correlation benefits while also focusing investors on the 2 best performing sectors over the long-run; consumer staples and health care.
Disclosure: None.
As always, interesting stuff here Ben. I'm digging the correlation analysis. One I definitely haven't seen before and has produced some interesting results!
Bert