Many authors on SA write about buying and holding blue chip stocks, especially those that pay dividends and raise them every single year. I have found them persuasive enough that I use this approach in my own investing. However, I try my best to identify pitfalls or risks in my approach. One concern I have is that the perceived benefits of blue-chip dividend growth investing could be subject to a sort of selection bias. The past results of successful companies like JNJ or MCD are very impressive. But could we have identified these companies decades ago, without the benefit of hindsight? Were there other companies that looked just as promising, but would go on to fizzle?
Inspired mostly by this article, I decided to conduct a more in-depth test of stock-picking. During the 1990's, an author named Gene Walden published several editions of The 100 Best Stocks to Own in America. I found the third edition from 1993 on Amazon and when I read it, I was impressed. I decided to see how investing in these companies would have turned out by tracking down the outcome and performance of each of these 100 stocks.
The list included 31 companies in the consumer staples sector, 20 industrials, 18 consumer discretionary companies, 17 healthcare companies, 8 tech companies and 6 financials. There were no utilities, telecoms, or energy companies. I don't know if that was by design or it just worked out that way. Walden also choose many dividend growth companies, and usually mentioned the length of a company's dividend growth streak. In some cases he only gave a lower end of the growth streak. For example, he wrote that Coca-Cola (NYSE:KO) "has raised its dividend for more than 17 consecutive years." In 1993 Coca-Cola's streak would have been at 30 or 31. To be consistent I used the numbers that he listed for the streaks, since many of them are much harder to check than Coke's is. According to those numbers, the list included 81 challengers, 64 contenders, and 21 champions, which represent streaks of 5, 10, and 25 consecutive years of increasing dividends, respectively.
The book was published sometime in 1993, and the latest price data included was from June of 1993. To be safe, I used December 31, 1993 as my start date, well after the book would have printed. The strategy I tested was totally passive. If there were mergers, the portfolio just held the new company. For example when Rubbermaid was acquired by Newell (NYSE:NWL) in 1999, I simply converted each RBD share into .7883 NWL shares (per the merger terms) and continued to hold them. Likewise when there were spin-offs the portfolio just held onto the spin-off, so in 1995 I added one Darden Restaurants (NYSE:DRI) share for each General Mills (NYSE:GIS) share the portfolio held.
I tracked capital returns, which included any cash from acquisitions or spin-offs, and dividends. I did not simulate reinvested dividends due to the work involved (checking the newspaper on microfiche, in some cases) to find historical price data for those stocks that are no longer traded. To help make up for this, I measured the annual rate of return using the IRR function in Excel for each company. I treated the initial purchase as a negative cash flow during Q4 of 1993, and all dividends and capital returned as positive cash flows during the quarter they were paid. I also included the ending value of stock still held as a positive cash flow during the last quarter of the period (Q3 of 2013). This is equivalent to calculating the rate of return with dividends re-invested under the assumption that the capital appreciated at a constant rate during the entire period.
I used portfolios that consisted of an equal dollar amount invested in each stock on the start date. I measured the performance of portfolios based on several different criteria:
# of stocks
Ann. Rate of return
Walden's top 25 stocks:
Walden's top 50 stocks:
Walden's top 75 stocks:
All 100 stocks:
All 100 stocks except the 5 best & 5 worst performers
Stocks with 30+ years of dividend increases
Stocks with 25+ years of dividend increases
Stocks with 20+ years of dividend increases
Stocks with 15+ years of dividend increases
Stocks with 10+ years of dividend increases
Stocks with 5+ years of dividend increases
Stocks with a 1993 current yield of 3% or more
Stocks with a 1993 current yield of 2.5% or more
Stocks with a 1993 current yield of 2% or more
Stocks with a 1993 current yield of 1.5% or more
Stocks with a 1993 current yield of 1% or more
SPY (S&P 500 ETF)
Walden's list of 100 stocks had a rate of return of 10.39% compared with 9.07% for the index fund, which over 19.7 years would amount to a 26.7% increase in total wealth. His results are not reliant on a few outliers, as the portfolio still outperformed the index even with the top and bottom 5 performers thrown out. Within the list of 100, his rankings did not appear to offer value - the top 25, top 50 and top 75 did not perform better than the top 100.
I did not find much additional value in the breakdown of the portfolios by dividend streaks or yields. The underperformance of the group with 30+ years of dividend increases is eye-opening but probably a fluke. The 15 stocks with yields of 3% or more did noticeably better than the entire 100, but it's worth noting the sample size was small and the industry concentration was high. Three of the stocks yielding more than 3% - UST, Philip Morris (NYSE:MO) and American Brands - were tobacco companies, and another four - Bristol-Myers Squibb (NYSE:BMY), American Home Products (later renamed Wyeth and now part of Pfizer), Warner-Lambert (also now part of Pfizer) and Merck (NYSE:MRK) were pharmaceuticals.
A more in-depth analysis would evaluate the portfolio's value over time to generate statistics on things like draw-down and volatility, but I only tracked the beginning and ending values. Based on the portfolio's composition I think its volatility profile would likely have been similar to the large-cap index.
The best and worst performers out of the 100, based on IRR, were:
Acquired in 1997
Now MO, PMI, KFT, MDLZ
Acquired in 2008
American Brands Inc
Now BEAM, FBHS
RR Donnelley & Sons
Merged in 1999
Acquired by BRK in 2006
Shaw Industries Inc
Acquired by BRK in 2001
Toys R Us
Acquired by Bain in 2005
Great Lakes Chemical
Merged - Chemtura in 2006
Acquired in 2011
Crompton & Knowles
Merged - Chemtura in 2006
Bankrupt in 2007
'Streak' means consecutive years of increased dividends as of 1993. The Capital and Dividends columns represent the total capital and dividends respectively that would have been generated by each dollar that was invested in the company on December 31, 1993. So, one dollar invested in Nike (NYSE:NKE) would have produced about $2 in dividends over the period, and been worth nearly $24 in capital at the end of the period. The table shows how a company that paid dividends earlier in the period (Philip Morris) can have a higher rate of return than a company that only paid dividend toward the end of the period (Microsoft) even though Microsoft (NASDAQ:MSFT) had greater capital appreciation and total dividends.
The timeframes used are not always equal, because many of the stocks were acquired for cash during the period. Loctite's appreciation looks modest next to some others but it was acquired in early 1997. In three years it gained nearly 74% with dividends. The bulk of American Brands - its tobacco business - was spun off as Gallaher, plc in 1997 which was acquired by Japan Tobacco in 2007. Only two companies - Stanhome and Crompton & Knowles - went totally bankrupt. Great Lakes Chemical did as well, after an ill-fated merger into Chemtura (bankrupt in 2010) with Crompton & Knowles, but they do have a spin-off, Innospec (NASDAQ:IOSP), that is still traded. Rubbermaid merged with Newell in 1999, another disaster for the shareholders of both companies. Many companies that underperformed ended up being acquired for less than the amount of the initial investment. Novell was acquired for $6.10 per share over 17 years after the portfolio bought it for $20.75 per share. It never paid a single dividend.
I was also interested in the fate of the champions:
Cut in 1999
American Home Products Corp
Froze in 2001
Yes, as PFE
Froze in 2001
Yes, as NWL
Procter & Gamble
Genuine Parts Co
Froze in 2002
Froze in 2002
Cincinnati Financial Corp
Froze in 1996
International Flavors & Fragrances
Cut in 2000
Johnson & Johnson
Froze in 2000
Froze in 1996
Yes, as PEP
Yes, as PMI, MO, MDLZ, KFT
American Brands Inc
Froze in 1996
Yes, as BEAM, FBHS
Cut in 2009
Cut in 2001
Cut in 2009
Of the 21 original champions, 7 are still investable and have their streak intact. 13 cut or froze their dividend at some point. Tambrands was acquired for cash in 1997 by Procter & Gamble (NYSE:PG).
I thought people would be interested in which companies had the highest yield on cost. Some do not like this metric but I think it is OK as long as all companies involved have the same timeframe. Each of the top 10 lasted the entire twenty-year period, although Warner-Lambert is now part of Pfizer (NYSE:PFE), and Philip Morris has been split into four parts.
End # of shares
Qtrly dividends - end of period
32 * $0.23 = $7.36
3 * $0.48 = $1.44
2.076072 * $0.13 = $0.27
3 * $0.85 = $2.55
.692024 * $0.5 = $0.35
8 * $.265 = $2.12
16 * $0.21 = $3.36
8 * $.315 = $2.52
Johnson & Johnson
4 * $0.66 = $2.64
16.5 * $0.24 = $3.96
16 * $0.225 = $3.60
4 * $0.495 = $1.98
16 * $0.28 = $4.48
1 *$0.84 = $0.84
I did not try to estimate the impact of reinvested dividends on yield on cost. The differing share counts are only due to splits, mergers, and spin-offs. I only counted dividends that had gone ex as of September 20, which means Microsoft's YOC will get a boost from their most recent increase to $0.28. Philip Morris's 32% includes all four of their child companies. Warner-Lambert (now as Pfizer) was number 7 even after Pfizer's 2009 dividend cut. WLA shareholders seem to have gotten a pretty good deal in the 2000 merger with Pfizer. I know of no meaningful way to report the yield on cost for the entire portfolio as many of the companies were acquired and so no longer pay dividends.
I was impressed with Walden's results. Consider:
- The stocks used were all well-known, very liquid, large cap stocks
- 100 stocks could be considered a 'closet index' but the portfolio was still able to outperform
- The portfolio did NO trading or rebalancing for the entire period. Nineteen+ years is a long time.
- Walden picked the stocks some six months ahead of the prices the portfolio bought at
- The book originally sold for $22.95 - a pretty good deal compared to 2-and-20.
It's also worth mentioning that I did not go through dozens of books and only report the results from the best one. This is the only book that I tested. I know that may seem like an obscure point but one of my concerns about back-tests is that people can tweak them to get the most impressive outcomes and this can result in 'overfitting' the data. The problem is discussed (very technically) here.
So how did Walden pick his stocks? The book devotes surprisingly little attention to this question, but here's one paragraph from the introduction:
In selecting the 100 companies… I looked at a wide range of financial factors, the most important of which was earnings performance. I wanted companies with a long history of annual increases in earnings per share because if a company is able to raise its earnings year after year, the stock price will ultimately follow. Other factors - such as revenue growth, stock price performance and dividend yield - also played into the screening process but none carried the same weight as earnings growth.
He does develop and explain a fairly elaborate ranking system that rates each company's revenue growth, price appreciation, earnings growth, earnings consistency, dividend growth, and dividend yield. But since his highly ranked stocks did no better than his lowly ranked stocks I mostly ignored the ranking system.
The results of this work strengthened my inclination toward stock-picking vs. index funds. There are certainly stocks that were probably considered 'rock-solid' in 1993 like Rubbermaid or IFF that disappointed. But many stocks that current SA authors point to as success stories were also on the list. Many are included in the above tables; others were Wal-Mart (NYSE:WMT), Abbott Labs (NYSE:ABT), Kimberly-Clark (NYSE:KMB), and Colgate-Palmolive (NYSE:CL).
It also shaped my thinking about dividends. Although the 81 dividend growth-stocks in the group failed to perform better than the non-growers, I do think it is telling that 81% of the stocks Walden selected had growth streaks of 5 or more years. I am less concerned about how many decades the streak has lasted or even the current yield. I believe it is more important to look at the underlying fundamentals - mainly earnings growth - that are driving the streak, as that is what seems to drive the returns. Of course I am currently young enough that I don't have a preference for income vs. capital gains. That may change later in life.
Many missing pieces remain. Valuation is important, but it's not clear that it was considered. Walden gives five-year P/E ranges for each stock, but he makes no mention of using the ratio in his selections or excluding stocks that were too expensive. And the totally passive approach is the easiest to test, but probably not the most representative of actual results. Some of the stocks were clear losers and probably would have been dumped early in the period. In other cases like Microsoft and Intel, you would have done far better if you'd sold at their 1999-2000 peaks than if you'd held them for the whole period. But that's easy to say with the benefit of hindsight.
If you are interested in this sort of thing, and don't already have one, I recommend a copy of The 100 Best Stocks to Own in America. (If you are not interested in some guy's stock picks from 20 years ago, I guess I can understand that too). My edition included more data not cited here including five-year revenue, earnings and dividend growth rates for each company, often displayed in 3-D bar charts that I'm sure were very state-of-the-art in 1993.
Past performance of course does not guarantee future results. Every market is different and you don't usually know how it's different until after it's over. But I do feel this analysis offers some credible evidence that buying and holding stocks of large, well-established, consistently profitable companies is a sound investment strategy.