In this year's Letter To Shareholders of Berkshire Hathaway (BRK.A) BRK.B), Warren Buffett wrote probably his most pointed commentary on how he feels about dividends in the scheme of things. In defending the decision for Berkshire not to pay a dividend, Buffett offered a mechanical explanation that explains why he believes the "sell shares to generate income" approach is better than having a company choose to pay out dividends along the way.
This is the first half of Buffett's argument, taken directly from the publicly available letter to shareholders, accessible here.
We'll start by assuming that you and I are the equal owners of a business with $2 million of net worth. The business earns 12% on tangible net worth - $240,000 - and can reasonably expect to earn the same 12% on reinvested earnings. Furthermore, there are outsiders who always wish to buy into our business at 125% of net worth. Therefore, the value of what we each own is now $1.25 million.
You would like to have the two of us shareholders receive one-third of our company's annual earnings and have two-thirds be reinvested. That plan, you feel, will nicely balance your needs for both current income and capital growth. So you suggest that we pay out $80,000 of current earnings and retain $160,000 to increase the future earnings of the business. In the first year, your dividend would be $40,000, and as earnings grew and the one third payout was maintained, so too would your dividend. In total, dividends and stock value would increase 8% each year (12% earned on net worth less 4% of net worth paid out).
After ten years our company would have a net worth of $4,317,850 (the original $2 million compounded at 8%) and your dividend in the upcoming year would be $86,357. Each of us would have shares worth $2,698,656 (125% of our half of the company's net worth). And we would live happily ever after - with dividends and the value of our stock continuing to grow at 8% annually.
My concern is the fact that I believe Buffett employs two premises that are uncharacteristic and often unrealistic for many investments. First of all, Buffett assumes that the company is able to generate 12% on all reinvested earnings. And second of all, he assumes that the stock constantly trades at a fair value at 125% of net worth.
The ability to generate 12% on reinvested earnings can be a tall order for many companies. Let's look at Big Oil for example. Exxon (XOM) is going to generate over $40 billion in earnings this year. Royal Dutch Shell (RDS.B) is going to generate over $30 billion in earnings this year. And Chevron (CVX) will probably generate in the vicinity of $30 billion this year. My concern is that it is an unrealistic assumption to think that all of that money can be deployed in such a way that it could return 12%. Maybe Exxon can achieve that with a portion of its earnings, but it's pretty hard to find ways to get 12% every year.
When explaining why companies pay dividends, Warren Buffett's own mentor Benjamin Graham stated that companies pay them because they can't find as good of opportunities with their "last dollar" of profit as they do with their "first dollar" of profit. In fact, it can be easier for a company to earn that 12% when management knows: (1) that a third of the money is going out the door as a dividend, so the company can be more selective with its growth pursuits by focusing on its best ideas, and (2) a dividend culture is a real thing. When a company knows that a chunk of earnings are going to be returned to shareholders with the expectation that it will grow in the future, it will put some pressure on management to avoid a strategy that might perhaps be more wasteful if the company could retain all of its earnings.
Buffett goes on to explain why he believes the sell-off approach is superior:
There is an alternative approach, however, that would leave us even happier. Under this scenario, we would leave all earnings in the company and each sell 3.2% of our shares annually. Since the shares would be sold at 125% of book value, this approach would produce the same $40,000 of cash initially, a sum that would grow annually. Call this option the "sell-off" approach.
Under this "sell-off" scenario, the net worth of our company increases to $6,211,696 after ten years ($2 million compounded at 12%). Because we would be selling shares each year, our percentage ownership would have declined, and, after ten years, we would each own 36.12% of the business. Even so, your share of the net worth of the company at that time would be $2,243,540. And, remember, every dollar of net worth attributable to each of us can be sold for $1.25. Therefore, the market value of your remaining shares would be $2,804,425, about 4% greater than the value of your shares if we had followed the dividend approach.
Moreover, your annual cash receipts from the sell-off policy would now be running 4% more than you would have received under the dividend scenario. Voila! - you would have both more cash to spend annually and more capital value.
Again, I have no problem with Buffett's argument. He defines his parameters, and then clearly outlines the logic and math for the reader to see. It is the premise of the argument that bothers me. He makes the assumption that the shares can always be sold at 125% of book value. The fact that this premise does not exist in real world investing is precisely one of the chief reasons why some investors pursue dividend-focused investing.
To cut to the chase, my response to Buffett's argument would be: How well does the "sell a few shares" strategy work in a stock market like we saw in March 2009? Or, what if we had an extended down market? How would the "sell a few shares" approach work in such a scenario? In 2008, Procter & Gamble (PG) traded for $73.80 per share. By 2009, it hit a low of $43.90. Johnson & Johnson (JNJ) fell from $72.80 to $46.30. Pepsi (PEP) fell from $79.80 to $43.80. And these are the bluest of the blue-chips. I would hate selling out my ownership stake at a moment when a company is trading for much less than it is worth. I'd rather keep my ownership stake, take my share of the profits that the company sends me in dividends, and then sell on my terms once the stocks became fairly valued or overvalued. Buffett's example works out well when you apply the 125% of book value assumption that he employs, but it offers little solace to investors in the real world that cannot enjoy the benefit of always being able to sell at 125% of book value.
For the purposes of Berkshire, Buffett's argument makes a lot of sense. He has been able to get that 12% (in fact, a lot more than that 12%) by retaining the capital of Berkshire shareholders. But with most megacap investments, it's going to be hard to find a company that can deploy all of its earnings at such a high rate, and thus, a dividend becomes an effective allocation of capital. And secondly, the art of income investing has a particular appeal to conservative investors that want to find a strategy that works in severe recessions, and Buffett's assumption that investors can always sell at 125% of book value is of little use during such an investment climate. For these reasons, investors are probably best off nodding their head in recognition that Buffett's explanation works well for Berkshire, but does not generally extend to other blue-chip investments.