I'd like to preface this article by suggesting that everyone is different with regard to his or her investment ideology, personal income and tax consequences. As such, this commentary should in no way be perceived as guidance or a blanket recommendation. However, I would like to present an interesting construct in relation to the dividends / capital gain tax debate.
Academically, ignoring taxes and frictional expenses, if a company earns satisfactory returns, dividends and capital gains are thought of in a similar manner. Now this has been widely debated on Seeking Alpha - but I'm not here to weigh in or comment on that portion. However, once taxes are added in, reality would hold that capital gains only trigger taxes when sold (that is, at the investor's discretion) while dividends are taxed year-in and year-out. As such, an argument could be made that tax-wise capital gains have a leg-up on dividend payments.
Again, the underpinning assumptions can be widely deliberated, but that's not the point. I am simply suggesting that a common argument for a "capital gain approach" - that is, focusing on appreciating share price only rather than income - is something along these lines:
"Capital gains are only taxed when you decide to sell. As such, long-term capital gains are more tax-efficient than dividends. Thus you should just focus on accumulating the greatest amount of wealth. Plus, later on you can always transform this accumulated wealth into those "dividend paying machines" you DGI fans are always talking about."
Both mathematically and academically, there's a seemingly reasonable point there. In fact, even the dividend devotees can't escape the idea that Warren Buffett has previously demonstrated this type of logic. In his 2012 Berkshire Hathaway letter Warren details why Berkshire does not pay a dividend. I'll spare you the summary and jump to the second advantage of what he called the "sell-off approach:"
"The second disadvantage of the dividend approach is of equal importance: The tax consequences for all taxpaying shareholders are inferior - usually far inferior - to those under the sell-off program. Under the dividend program, all of the cash received by shareholders each year is taxed whereas the sell-off program results in tax on only the gain portion of the cash receipts."
However, I would caution on always accepting generalized counseling. Surely the math works out in some instances and obviously Warren makes the case for Berkshire quite well. Yet, as I'm about to illustrate, it's important to underscore the idea that investors can have highly individualized circumstances.
We'll run through two examples to show you what I mean. (Both of which will assume you won't need the funds for 20 years and thereafter will begin withdrawing.)
Let's start with the "capital appreciation only" scenario. Now this certainly isn't my investing foray, but imaginably you might own a collection of non-dividend paying companies in the rational expectation of higher share prices in the future. For instance, perhaps you hold Cognizant (NASDAQ:CTSH), Google (NASDAQ:GOOGL) and Berkshire Hathaway (NYSE:BRK.A) along with 20 or 30 other companies. Further, perchance a sensible assumption for share price growth (not business results) would be about 9% annually over the coming 20 years. Finally, we'll assume you invest $10,000 each year for a total nominal investment of $200,000.
What does this mean? Well, a $10,000 yearly contribution compounding at 9% annually turns in to a little more than $557,000 over 2 decades - which we'll call $550,000 for simplicity. So the first option would be to have $550,000 in paper net worth, having paid zero taxes along the way, with $350,000 in capital gains ready to be taxed when or if you so choose to "cash out."
On the other hand, you could construct a dividend growth portfolio with the classic dividend growth examples: Coca-Cola (NYSE:KO), Johnson & Johnson (NYSE:JNJ), Procter & Gamble (NYSE:PG), PepsiCo (NYSE:PEP) and McDonald's (NYSE:MCD) for instance (plus a couple dozen more). In effect, any group of holdings with an expected 3% aggregate yield and 6% dividend growth expectation (and constant payout ratio) would work. Now it should be underscored that in both this example and the "capital appreciation only" example there are a variety of simplifying assumptions being made; chief among them the lack of volatility. However, this does not take away from the lasting realization.
Arguably, one might prefer "capital appreciation only" stocks due to the possibility for higher returns. However, in this instance and for the reason of comparability, I assumed the returns would be the same. Interestingly, before taxes and frictional expenses, a 3% yield coupled with 6% growth leads to the same results as the 9% share price growth example when you reinvest the dividends. For instance, at the end of year one the initial $10,000 would be worth $10,900 with either 9% share price growth or 6% growth ($600) and a 3% ($300) dividend. If the $300 dividend is reinvested, the next year's dividend example would equal $11,881 ($11,554 + $327) the same as the share price growth example ($10,900*1.09) and so on.
So prior to taxes and frictional expenses, as academics would indicate, both methods lead to the same results. Yet that's the rub: we of course cannot forget about taxes and frictional expenses. Given that the "capital appreciation only" method requires no such payment of taxes or trading costs (other than those also borne by the dividend example in purchasing shares) it follows that the "no dividend method" is superior, correct? Expressed differently, with the dividend method you not only have to pay taxes each year but you also have trading costs associated with reinvesting dividends, right?
Not quite. Or perhaps more respectfully: "possibly, but not necessarily." Many brokerages allow for direct reinvestment plans (DRIPs) that do not charge to reinvest dividends. Better yet, companies like Scottrade allow for flexible reinvestment (FRIP) whereby you can pool your dividend funds and choose to reinvest in the same or different securities free of charge. More plainly: it's possible to reinvest dividends without transaction fees.
The point of taxes is important, but highly specific. First, in both instances you could be using a tax-advantaged account and thus not have to pay taxes as you go on either example. I'll take an even greater leap: you don't even need to have a tax-advantaged account to avoid dividend taxes. If you're in the 10% or 15% tax bracket, the tax rate is 0%. So, despite the seemingly unrealistic assumptions of no taxes or frictional expenses, it's entirely possible to avoid both. Sure things could change in the future, but as of today it's certainly possible for many.
So now you have two distinct examples - the "capital appreciation only" method or the "dividend" method - with end values that are equivalent. Call it $550,000 in paper worth or else roughly $16,500 in possible dividend payments assuming a 3% yield. So both are equal, right? Not necessarily.
With the "capital appreciation only" method you could use the "sell-off" approach and generate the same amount of payments. With the "dividend" method you don't have to do anything and you'll receive about $16,500 this year. Conceivably you could continue on either path, with both long-term capital gains and dividends being taxed at the same rate.
Yet what happens if you get tired of selling shares - subjecting yourself to the whims of the market - and want to change your portfolio out for the equivalent "dividend growth machine" that the second example provides? Here's where the script flips. If you have the "capital appreciation only" portfolio, you have $550,000 in paper worth that you would like to convert into $550,000 worth of dividend growth companies. But there's a problem. Remember that your cost basis is only $200,000. If you sell all at once, that triggers $350,000 in long-term capital gains - much greater than the 10% or 15% tax bracket and thus subject to the 15% long-term capital gains rate. When you come back to the market to purchase those DGI companies, you're now left with just under $500,000 (not to mention the additional transaction costs of selling and buying shares). In effect creating the precise two things you initially worried about with the dividend method.
Now obviously this was a bit of an extreme example. Although it should be underscored that - with the exception of the theoretically assumed perfect consistency - it is entirely possible. Moreover, this is why it's important to consider what type of method you want to start out with - it could have implications down the road.
Here's the takeaway: a common argument for a "capital appreciation only" approach is that you avoid taxes and transaction expenses. That, all else equal, you would rather have the returns in the form of capital gains than dividends such that you decide when to pay taxes. However, as I demonstrated above, it's conceivable that one wouldn't have to pay taxes or transaction costs with the dividend method either. As a consequence, it could actually be the case - given a penchant for dividend income in the future - that dividends are more tax-efficient than capital gains. So before following a blanket ideology, recognize the importance of evaluating the underlying math in your personal situation.
Disclosure: The author is long MCD, JNJ, KO, PEP, PG. The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.