There has been an ongoing discussion/debate here on SA about how to best create income from your retirement portfolio. As a dividend growth investor, I feel that the wisest strategy is to create a portfolio of dividend growth stocks that pays you enough in dividends to fund your retirement. In this scenario, it is not necessary to sell any shares of your principal because the dividend yield gives you all the cash you need.
The opposing argument is that, although there is nothing wrong with holding good quality dividend stocks, it is not the best way to create income. The opposing camp feels that if you hold good quality growth stocks, whenever you need income you simply sell some of the shares in your portfolio and withdraw that money. Those in the "growth stock" camp feel that the total return from growth stocks, in the end, will be higher than the total return from dividend stocks, so you will be better off buying growth stocks.
We in the "dividend stock" camp feel that the growth camp is playing a dangerous game with this strategy. By selling shares every time you need income, those shares will forever be removed from your portfolio and they will never again be able to contribute to the growth of your wealth. Conversely, by simply using dividends for your income you never have to sell any stock, and all the shares will still be in your portfolio, churning out dividend after dividend, quarter after quarter, while also contributing to your capital gains through price appreciation.
In an ironic twist, both sides of the debate bring up Warren Buffett to support their argument. The growth camp correctly points out that Berkshire Hathaway (NYSE:BRK.A) does not pay a dividend. And if dividends are so important, why doesn't Buffett choose to pay one? Conversely, the dividend camp points out that although Berkshire does not pay a dividend, Buffett has chosen to own many dividend-paying stocks, such as Coke (NYSE:KO), Wells Fargo (NYSE:WFC) and IBM (NYSE:IBM), and has made a lot of money by owning these stocks. BRK collects millions of dollars worth of dividends every year from the companies it owns. Why would Buffett own so many dividend-paying stocks if he didn't think that dividends were important? With these two points, both sides make compelling arguments.
In the comments section of a recent article by Geoff Considine, Larry Swedroe referred the readers to an article which quotes Buffett from one of his recent annual letters to Berkshire shareholders, in which he speak directly to the dividend issue. In the letter, he discusses the reasons why Berkshire does not pay a dividend, and in response to the critics who say that it should, he states that BRK holders who need income can simply sell some BRK shares to get the cash they need. And he states that if you only sell a small percentage of your holdings, then the increase in the price of the remaining holdings will be enough to overcome the loss of those few shares every year.
To demonstrate his argument, he uses his own holdings of BRK as an example. He points out that from 2005 through 2012, he had given away 4.25% of his holdings to charity every year. During that time period, the number of shares of Berkshire he held dropped from 712,497,000 B-equivalent shares down to 528,525,623 shares. And yet, the value of his holdings increased from $28.2 billion to $40.2 billion.
Now I'm not writing this article to make an argument about whether or not BRK should pay a dividend. Buffett is the best person to decide what BRK's dividend policy should be. I'm bringing up Buffett because in this letter to shareholders he makes the point that someone looking to create income can do so by simply selling shares in his portfolio. But as I read the article, I had many problems with the points that Buffett was making and how it relates to creating retirement income. The most obvious one, at least to me, is that not everybody will be able to invest their money as well as Buffett, so it's not fair to conclude that one's portfolio will continue to increase in value, even after selling shares, the way Buffett's did.
But there were other issues as well. 2005-2012 was a great time to be invested in BRK. Of course, someone who held a large number of shares of BRK would do well during that time period, even if he sold a few percent each year. But it is during down markets when problems develop with the selling strategy. What would have happened during a period where BRK's price was dropping? Buffett doesn't say. Also, Buffett's discussion covered just a 7-year period. Many people's retirements last 20 years or more, and they have to make their money last for all that time. How would the strategy hold up over a much longer time period?
And finally, what would have happened if Berkshire Hathaway DID pay a dividend? Maybe he could have given away 4.25% of his portfolio every year WITHOUT selling any shares, and his portfolio would be worth even more today than it actually is. Maybe he would have done even better if BRK actually had been paying a dividend.
To investigate these questions I decided to run a test, using the actual historical prices of BRK going back to 1990 (that is as far back as the prices go for BRK on Yahoo.com), and see what would happen if, in scenario A, a regular dividend was paid and withdrawn as income, or, in scenario B, a small percentage of stock was being sold during the entire 24-year period.
Obviously, BRK has not been paying a dividend for the past 24 years, so I had to make some assumptions to run the scenario. First, since the "4% rule" is the one most often used to figure out how much income people should withdraw each year for retirement, for scenario A, the "BRK with Dividend" portfolio, I assumed a 4% yield for BKH every year, or 1% each quarter. I felt it would be too much of an assumption to include a growth rate for the dividend, so I simply kept the dividend at 4% a year, year after year. So, the theoretical dividend paid each quarter moved up or down as the price moved up or down. The cash created by this dividend was removed from the portfolio. To appease those who feel that the payment of a dividend causes a permanent drop in the stock price by the same amount of the dividend, I adjusted BKH's stock price down, by the amount of the dividend, after each dividend payment. I made sure that this effect accumulated throughout the study period each time a dividend was paid. (more about this later)
For scenario B, the "BRK without Dividend" portfolio, I withdrew the same amount of cash as was withdrawn from the "BRK with Dividend" portfolio every quarter, but I "created" the cash by selling shares of BRK. To keep the amount of cash produced from each portfolio the same, I allowed for selling of fractional shares. For the scenario B portfolio, the actual historical BRK stock price was used throughout. No adjustment was made for the payment of dividends as in scenario A (since no dividends were paid).
BRK with Dividend Portfolio
- Due to the high price of BRK stock, the portfolio was started with $10 million to get a sufficient number of shares.
- 1% of the portfolio value was removed each quarter (4% a year) as a dividend payment.
- The stock price was lowered by the amount of the dividend paid, and this price adjustment was cumulative for entire 24-year study period.
- Since no shares were sold, and there was no dividend reinvestment, the number of shares of BKH held throughout the study period was the same as the number of shares bought at the beginning.
- The final portfolio value was determined by the final adjusted stock price multiplied by the number of shares owned.
BRK without Dividend Portfolio:
- The portfolio was started with the same number of shares and the same stock price as the "BRK with Dividend" portfolio.
- An amount of cash equivalent to the dividend payment from the "BRK with Dividend" portfolio was removed each quarter by selling shares.
- No dividend adjustment was made to the historical stock price, and the final value of the portfolio was the final stock price multiplied by the final number of shares.
The study ran from Jan 12, 1990 through Jan 27, 2014. During that time, BRK's stock price went from $8200 to $169,511. The adjusted stock price for the "BRK without Dividend" portfolio at the end of the study was $118,616.48 due to the payout of $50,894.52 of dividends over the 24 year period.
The "BRK with Dividend" portfolio both started and finished with 1,219 shares. The "BRK without Dividend" portfolio had shrunk down to 285.36 shares by the end of the study.
The portfolios created a total of $61,719,656.20 for the portfolio owner. (Again, this is with a starting portfolio value of $10 million).
The final portfolio value of the "BRK with Dividend" portfolio was $144,593,491.01, for a return of 11.75%, not including the dividend payouts. The final value of the "BRK without Dividend" portfolio was $48,372,089.38, for a return of 6.77%.
First of all, let me say that I don't agree that the stock price of BRK should be permanently adjusted downward by the amount of the dividend each time a dividend is paid, as I did in this scenario. Although the markets do this automatically on the ex-div date, there is no evidence that this is a lasting effect. Usually within a few days or weeks that drop in stock price is erased. At $169,511, BRK has a PE of a little over 14. If we really did adjust the price downward for each dividend, and BRK's price was only $118,616, then its PE would be about 10.25. I don't think anybody believes that BRK should be trading at such a low PE. No matter how many dividends were paid, the market would bring the price right back up to where it should be based on its earnings. But I did it in this study to cut off that criticism before it ever gets made.
The final value of the "BRK with Dividend" portfolio would have been even higher had I not adjusted the share price. And still, even with lower adjusted the price, the "BRK with Dividend" portfolio created $96 million more value than the "BRK without Dividend" portfolio. Although they both were able to give their owner a 4% yearly payment, while still growing the portfolio's value, the "BRK with Dividend" Portfolio outperformed the "BRK without Dividend" portfolio by over 4 percentage points a year, which translates into almost a 200% larger portfolio after 24 years.
To me, this shows that selling shares to create income is a dangerous strategy. BRK's stock price had an annual return of over 21% from 1990 through 2014, so both strategies were able to be profitable. But how many of us will be able to build a retirement portfolio that will return 21% a year? I'm guessing none of us. With the more reasonable 8-10% annual return we can expect during our retirement, a non-dividend paying portfolio most likely would have run out of money, or, to keep from running out of money, the owner would not have been able to take out as much income as the dividend portfolio owner was able to.
So, although I hesitate to criticize anything Warren Buffett would do or say, in this case I would have to say he is wrong. Even for someone like himself, a master of capital allocation, I feel that for the sake of his own portfolio (not necessarily as a business decision for BRK) he would have been better off had BRK paid a dividend. He would have been able to donate those dividends to charity, and still maintained ownership of all his shares of BRK.
For those of us who cannot expect to produce 21% annual returns like Buffett, using a dividend strategy is not only a more conservative strategy, one that makes it unlikely that you will run out of money, but it is also one that is likely to allow you to continue to grow your portfolio while in retirement. You get the trifecta of security, good income, and continued capital growth.
Thank you for reading my article. I welcome your comments and criticisms.