Efficient Frontier
Efficient Frontiers Logo
William J. Bernstein

Of Earnings, Dividends, and Agency

In one important regard it really is different this time, and that's dividends. For most of the past two hundred years the market yield has fluctuated between 3% and 6%. Forays above/below this range have usually proved a brief but reliable harbinger of high/low subsequent returns.

But in 1994 yields fell below 3% and haven't looked back since. The new conventional wisdom is that dividends no longer matter; stock returns come not from payouts but from capital gains, which in turn rely on the ability of companies to grow their businesses through retained earnings. Implicit in this belief is the assumption that companies can invest these earnings as well as, if not better than, the shareholders can invest their dividends. A universe yielding 1.2% requires an unprecedented degree of faith in the ability of large corporations, owned almost exclusively by outsiders, to properly dispose of these massive cash flows. How well placed is this faith?

In order to answer this question we first have to examine what academicians call the "dividend puzzle;" what determines how much payout shareholders receive, or even whether they get any at all?

Consider a company that earns 10% of its capitalization each year. If it pays out all its earnings as dividends—i.e., a 10% yield—then it will not have capital to grow, its stock price will not increase, and its return is just the 10% dividend. If it pays one half of its earnings out—a 5% dividend—it will reinvest the remaining 5% in the company, which will then grow at 5% per year, producing a 5% capital gain. The 5% dividend and 5% capital growth add up to the same 10% total return. (This, by the way, was the source of the mistake made by Glassman and Hassett in their early writings on Dow 36,000 that drew howls of laughter from economists; they "double counted" earnings as both payout and growth.) And finally, if the company never pays out a dividend it will grow at 10% per year, which will accrue to the shareholder as a 10% annual capital gain.

Thus, in a taxless world a company’s dividend policy should matter not at all to the shareholder. Inside academia, this is known as the "Modigliani-Miller theorem." In the taxable world, of course, shareholders prefer capital gains to dividends. So why do companies pay them?

Because, to put it bluntly, corporate officers are often scoundrels and theives. They lie. They cheat. They steal. They invest in projects more on the basis of turf, prestige, and politics than cash flow. They run around in Learjets and eat fois gras on your nickel. Shareholders intuitively know this and insist on spiriting their cash away from these bad actors as fast as they can.

This dismal view of corporate finance falls under the rubric of "agency costs." In other words, the shareholder’s priorities (e.g., maximizing return) are not at all the same as the officers’ priorites (turf, prestige, high salaries, and luxurious surroundings, all removed from bothersome outside scrutiny).

Agency costs are also old news. Listen to Ben Graham in the 1934 edition of Security Analysis:

The typical investor would most certainly prefer to have his dividend today and let tomorrow take care of itself . . . Given two companies in the same general position and with the same earning power, the one paying the larger dividend will always sell at the higher price. (Graham’s emphasis)

This wisdom usually gets forgotten in speculative bubbles. Surely, says the modern investor, the lack of a payout is of no consequence; my return will accrue from capital gains. We’re in a new era, after all. Well, Mr. Graham spent several chapters in Security Analysis describing another "new era"—the one from 1921 to 1929. Listen again:

The customary reasoning on this point may be stated in the form of a syllogism, as follows:

Major premise—Whatever benefits the company benefits the shareholders.

Minor premise—A company is benefited if its earnings are retained rather than paid out.

Conclusion—Stockholders are benefited from the withholding of corporate earnings.

The weakness of the above reasoning rests of course in the major premise. Whatever benefits a business benefits its owners, provided the benefit is not conferred upon the corporation at the expense of the shareholders . . . An inductive study would undoubtedly show that the earnings power of corporations does not in general expand proportionately with increases in accumulated surplus.

The Dean’s elegant prose needs some grammatical and historical translation for the modern audience: "I can guarantee you that retained earnings are bad news. You say you want hard data? Well, I don’t have it. Remember, this is 1934—prying the requisite numbers from companies is back-breaking work. Maybe when the 1933 Securities Act kicks in and we get some sunshine I can prove my thesis. Maybe some day modern technology will allow us to extract this information with the push of a few buttons. But not just yet."

Over the past several decades the sun has indeed shone in on corporate finance. In fact, one of the great ironies of global finance is that our markets have become the world’s most transparent and liquid precisely because they are the most heavily regulated. But I digress. A large body of data has indeed accumulated proving Graham’s hypothesis, much of it published in the past several months in the Journal of Finance. (Note: all of the articles I’ll refer to are available online, but you’ll need the Acrobat reader to view them.)

In the December issue Jarred Harford found that cash-rich firms destroyed 7 cents of corporate value for every dollar of cash reserves held. How does this happen? Let’s take two firms, both of which are considering a project or acquisition of marginal value. The first firm is cash-poor, and must obtain the capital from a bank, or a stock or bond issuance. This necessitates scrutiny of the project from the outside. The second firm is cash-rich, and thus requires no outside scrutiny—they can simply cut a check. Clearly, the cash-rich company is much more likely to make this potentially unprofitable investment. Or, as Harford puts it, "Large cash balances remove an important monitoring component from the investment process."

In short, retaining earnings provides management with excess cash, which in turn frees them from outside scrutiny. Rather counterintuitively, excess cash turns out not to be such a good thing.

Three more pieces from the February issue dovetail with this theme. Rajan, Servaes, and Zingales look at the performance of large conglomerates, and find that investment capital tends to flow most readily to its least productive divisions. The more highly diversified the company (i.e., the less related its component businesses) the more dramatic the effect. What is most interesting is that Harford's research found that cash-rich companies are more likely to make diversifying acquisitions—in other words, to turn them into the same companies that this paper shows are the least efficient. Like, say, when a large cash-rich internet service provider acquires an even larger company know for its glossy magazines and Looney Tunes.

In the same issue, Ang, Cole, and Lin find that agency costs are highest when the managers own little or no stock, when there is a large number of nonmanager shareholders (i.e., ordinary investors), and when there is no bank scrutiny. In short, the most efficient business is one run by the sole owner. This should resonate with health-care practitioners, as it is a commonplace that small practices are the most efficient and only rarely fail; large specialty clinics go belly up with alarming frequency, and investor-owned HMOs have become a long-running disaster movie playing at a mutual fund near you.

Last and not least, the February JoF contains an absolute gem from La Porta, Lopez-de-Silanes, Shleifer, and Vishny on dividend policy around the globe. Their primary finding is that in so-called "civil law" countries, such as most of Latin America, Scandinavia, and southern Europe, where investor protection is the weakest, dividend payouts are low. In so-called "common law" countries—basically the world’s English-speaking nations, where investor protection is excellent—payouts are high. Which gets back to Graham’s basic premise; investors prefer dividends and take them whenever the law and culture allow. The authors reinforce the points made by Graham and the other pieces; "failure to disgorge cash leads to its diversion or waste, which is detrimental to outside shareholders’ interest."

But what is most remarkable about this piece is its tone, which is almost Menckenesque in its description of modern corporate ethics. They describe a Hobbesian world in the kind of plain English rarely seen in academic finance; "Firms appear to pay out cash to investors because the opportunity to steal or misinvest it are in part limited by law, and because minority shareholders have enough power to extract it."

So we are faced with a startling paradox. The enormous recent success of American corporations has left them with elephantine cash flows that free them from outside scrutiny, which is in turn likely to result in future capital inefficiencies. A nearly identical scenario played out when low dividend payouts led to the conglomerate mania of the '60s, which in turn generated the industrial malaise of the 70s. Plus ça change.

To Efficient Frontier Homepage E-mail to William Bernstein

Copyright © 2000, William J. Bernstein. All rights reserved.