Ideas can become so entrenched that they are not questioned. One such idea is that investors should care only about total return (dividends plus price appreciation) rather than about the fraction of return delivered by dividends or price gains. This belief is pervasive in the financial services industry and I believe that very few people would argue with the proposition that the 'total return' paradigm is the standard of practice.
The idea that investors should be indifferent as to whether their returns come from dividends or prices gains originates with the Modigliani-Miller Hypothesis. The basic idea, as far as dividends are concerned, is that companies it should be no more or less profitable to use your own retained earnings to fund future growth than it is to borrow money to fund future growth. As such, returning earnings to shareholders in the form of dividends should not matter to long-term profitability and, ultimately, to investor returns. The the MM Hypothesis depends on an assumption that capital markets are both rational and efficient. In this world view, there is no value- or small-cap effect. There is also no reason to invest in anything other than a market-cap weighted market index.
If one believes in totally rational markets, in which the prices of every asset reflect all of the available information, it is perfectly reasonable to ignore dividends. The problem is that the real markets are demonstrably irrational. The existence of bubbles, for example, is a clear violation of efficient markets. There are many more examples, of course. A key idea that is ignored in so much of the rational markets literature is the uncertainty that people have when estimating future returns on different assets. Investors must explicitly or implicitly come up with expectations for future risk and return on the various investment alternatives available to them. How great is our uncertainty with regard to these estimates? How might our uncertainty with regard to our estimates of future return vary across different types of investments? This is where dividends differ from price appreciation. Our estimation of future returns from dividends is relatively more certain than our estimates of future price gains. I recently wrote a long article exploring this issue.
The key issue that comes out of that analysis is that dividends matter if you care about being able to predict the consistency with which your portfolio generates returns. Clearly, this is a primary concern for people in or entering retirement. Even if you know with certainty that your portfolio will generate an average return of 8% per year, you may still end up depleting your portfolio entirely with only a 5% income draw if the portfolio has a bad run of returns at the start of retirement. If your portfolio has a dividend yield of 5% and expected price appreciation of 3%, for a total return of 8% per year, your situation is clearly different from a situation in which you have a dividend yield of 2% and are expecting price appreciation of 6%.
David Blanchett, head of retirement research at Morningstar, published an important paper this summer that lays out some of the differences implicit in income-oriented portfolios and total return portfolios. This paper does not, however, deal with the issue of the differences in estimation risk for income-generating assets and non-income assets. In email communications, however, he has acknowledged the importance of this issue. His treatment assumes an investor preference for income vs. price appreciation via a new utility function, which is a mathematical description of how an investor balance risk and return.
Receiving a dividend means that a company is paying cash to investors rather than investing this money in future growth opportunities. Investors are giving up future (uncertain) growth in preference for a payment today. This is similar to investors who sell call options against their holdings. Certainly your upside potential is limited while you still hold all the downside risk in both cases. In the worst case, in which a company defaults, neither dividends nor income from a covered call will substantially reduce your pain.
I believe that the issue of dividends in equity returns is going to be enormously important as the U.S. ages and we transition from a population of investors who are fairly insensitive to the consistency of their returns to one in which investors have amplified sensitivity to variations in return from year to year. Once you start drawing income, consistency matters a great deal. This challenge is further magnified because bond yields are so low. While a portfolio entirely dominated by bonds was once a viable option for some fraction of the population, the percentage of people with sufficient assets to live entirely off of bond yield must surely be becoming vanishingly small.