Stock repurchases are replacing dividends for a new generation of companies.
This article first appeared in the August 2007 edition of Capital Ideas, which summarizes of research conducted by the faculty of the University of Chicago Graduate School of Business. Reprinted with permission.
Over the last 30 years, corporate payout policy — a firm’s policy for paying dividends and making stock repurchases — has changed significantly as earnings have become more volatile. A new study links the increased variation in reported earnings to changes in corporate payout policy, finding that stock repurchases are now being used as substitutes for dividends even for firms that continue to pay dividends.
Publicly traded firms typically distribute cash to stockholders as dividends or stock repurchases. Dividends are cash distributions to stockholders, while repurchases involve companies buying back their shares from stockholders.
Previous research by Eugene F. Fama, Robert R. McCormick Distinguished Service Professor of Finance at the University of Chicago Graduate School of Business, and Kenneth French of Dartmouth University’s Tuck School of Business shows that far fewer U.S. companies are paying dividends to their stockholders today than in the late 1970s. Repurchases emerged as an economically significant phenomenon in the early 1980s, and by 1998, the amount of stock repurchases exceeded the amount of dividends.
A new study, “The Evolving Relation between Earnings, Dividends, and Stock Repurchases” by University of Chicago Graduate School of Business professor Douglas J. Skinner, suggests that companies now view dividends and stock repurchases as essentially equivalent methods of paying out earnings to stockholders and analyzes whether the change in payout policy is linked to changes in reported earnings.
“I wanted to investigate whether the increase in the variability of earnings made it more difficult for companies to pay dividends, since dividends are usually paid based on earnings,” explains Skinner.
Over the last 20 to 30 years, the nature of corporate earnings has changed, with greater variation across firms: losses are much more common; earnings are increasingly concentrated in a small group of very large firms; and earnings tend to be more volatile over time for a given firm.
Skinner shows that the increased volatility of earnings helps explain changes in payout policy over the last 30 years.
First, Skinner shows that since 1980 there have been three main groups of payers: 1) established firms that have always paid dividends and now also make repurchases on a regular basis; 2) firms that make regular repurchases but do not pay dividends; and 3) firms that make occasional repurchases.
— Douglas J. Skinner is John P. and Lillian A. Gould Professor of Accounting at the University of Chicago Graduate School of Business.
Skinner finds that firms in the first group have been paying dividends for decades and continue to do so largely because of their dividend history. Over time, managers of these firms increasingly use repurchases to pay out earnings increases, with dividend policy becoming more conservative.
This suggests the reluctance to reduce or omit dividends has become stronger in recent years. Skinner also finds that these firms dominate the set of publicly held firms, accounting for more than half of all total earnings and payouts.
These results suggest repurchases are increasingly used in place of regular dividends. The result holds not only for those large, mature and profitable firms that continue to pay dividends, but also for firms that only make repurchases, most of which have never paid dividends.
“Financial economists have always been puzzled that companies pay dividends at all, given their tax disadvantaged status relative to stock repurchases,” says Skinner.
Furthermore, dividends represent a significant ongoing commitment by the firm. Once companies have established a dividend, they only reduce or omit them in exceptional circumstances, because dividend reductions are interpreted by the market as a negative signal about the company’s future earnings prospects. This explains managers’ tendency to increase dividends conservatively, and why firms that do not already pay dividends prefer to make repurchases rather than initiating a dividend. For example, Oracle and Dell have never paid dividends, but make substantial repurchases.
“Dividend policies are incredibly sticky. Once firms have established a dividend amount, they feel as if they have to maintain that level of dividends going forward,” notes Skinner.
One problem with increasingly variable earnings is it becomes difficult for companies to be sure they can maintain their level of dividend payments. “There’s total flexibility with stock repurchases, whereas you’re committed forever with a dividend,” says Skinner. “With repurchases, you could pay out $3 billion in total repurchases one period and zero the next.”
The availability of repurchases means that newer firms without a dividend history are unlikely to initiate a dividend. For the majority of firms, dividends do not offer benefits over repurchases. Although some argue that dividends are useful barometers of “earnings quality,” there is not much evidence to support this idea.
“The main contribution of this study is the substitution hypothesis, determining if stock repurchases are actually used as substitutes for dividends,” says Skinner. “If this is the case, stock repurchases should be linked to earnings, because prior research has shown that managers set dividends based on current and past earnings.”
Skinner finds that the relationship between dividends and earnings has become weaker in recent decades, largely because managers now set dividends in a mechanical fashion with small, predictable increases. At the same time, the relation between earnings and repurchases has become stronger, suggesting that managers are increasingly using repurchases to pay out earnings increases.
To test the substitution hypothesis, Skinner used a panel of data on earnings, dividends and net repurchases for publicly held U.S. industrial firms from 1980 to 2005. Skinner finds that earnings become increasingly volatile over time, partly due to the increasing frequency and magnitude of losses, which helps explain the growth of repurchases as a substitute for dividends; for legal and institutional reasons, it is often difficult to pay dividends when the firm reports losses.
Due to a change in regulation, stock repurchases emerged as a viable form of payout around 1983 and increased rapidly after that time. Aggregate net repurchases were consistently in the $30 billion range from 1985 to 1990 before declining briefly in the early 1990s. After this, repurchases increased dramatically, and exceeded aggregate dividends for the first time in 1998. Net repurchases also exceeded dividends in 1999, 2000, 2004 and 2005. Aggregate repurchases grew to $233 billion in 2004, nearly twice the level in 1998. This amount is substantially larger than dividends, which totaled $197 billion in 2004.
Skinner's research shows that dividends and repurchases are loosely tied to earnings. However, based on an analysis of 345 firms that paid dividends and made regular repurchases between 1980 and 2005, he finds firms have a conservative dividend policy and use repurchases to supplement shareholder distributions in years with strong earnings. The combined payout (dividends plus repurchase) tracks earnings closely.
Consistent with the previous research by Fama and French, Skinner finds that the proportion of dividend payers, which peaked in the late 1970s at around 66%, declined steadily over this period, from 42% in 1980-89 to 28% in 1995-2004. In spite of the decline in the number of dividend-paying firms, aggregate dividends increase over this period, reflecting a massive increase in the concentration of dividend payments in a relatively small group of large, established firms.
Skinner finds that those firms that both pay dividends and make repurchases have a conservative dividend policy and use repurchases to supplement dividends in years with strong earnings.
“If a firm has unusually high earnings one year, they’re not going to increase dividends; instead, they’ll make a large stock repurchase,” says Skinner.
Skinner predicts that firms that continue to pay dividends do so largely because of history. To test this prediction, Skinner quantified the dividend history for these companies as the number of years before 1980 in which they paid dividends. Consistent with the idea that these firms only continue to pay dividends because of their history, Skinner finds the following: most firms that continue to pay dividends today have paid dividends for many decades; firms that make regular repurchases but do not pay dividends have little or no dividend history; and few newer companies initiate dividends.
For firms that have never paid dividends but make regular repurchases, Skinner also finds an increasingly strong relation between earnings and repurchases, suggesting these firms use repurchases in place of dividends.
Interestingly, Skinner finds that earnings also do a good job of explaining payouts for firms that only make occasional repurchases, consistent with the substitution hypothesis.
Other factors also explain repurchases. Previous research finds that managers repurchase more when their companies’ stock prices are low to offset the dilutive effects of stock options programs and increase reported earnings-per-share. While Skinner finds that these factors continue to explain repurchases, he also finds that they largely explain the timing of repurchases; the level of repurchases is driven by earnings.
Apart from those firms that continue to pay dividends and make regular repurchases, dividend payers are no longer economically important. From 1980 to 2005, the percentage of firms that pay dividends but do not make repurchases declined to 7% from 13%. Pure dividend-payers are disappearing.
Dividends are now the domain of a relatively small group of large well-established “blue chip” firms like GE, Exxon Mobil and Altria. Collectively, these companies now account for over half of the dividends and earnings of all public firms listed in the United States.
In addition, firms that only make repurchases have grown considerably. This group will likely become more important as relatively young technology firms continue to mature and distribute more cash to stockholders.
“If a firm has unusually high earnings one year, they're not going to increase dividends; instead, they'll make a large stock repurchase.”
— Douglas J. Skinner, professor, University of Chicago Graduate School of Business
“Some technology companies like Dell and Oracle have never paid dividends, and my bet is they never will,” says Skinner. “Companies that traditionally would have paid dividends are going straight for stock repurchases.”
The results also suggest that industrial firms now display a largely two-tiered structure, with a smaller number of large firms that collectively dominate the distribution of both earnings and payout, and a larger number of smaller, often unprofitable firms with high growth opportunities. Some of the top earners now eschew dividends in favor of repurchases, which means that while the inertia in dividend payments is considerable, there may come a time when dividends completely disappear.
“The Evolving Relation between Earnings, Dividends, and Stock Repurchases.” Douglas J. Skinner. Forthcoming in the Journal of Financial Economics.
Northern Trust is proud to be a corporate sponsor of the University of Chicago Graduate School of Business Initiative on Global Markets (IGM). IGM explores business, financial markets and public policy issues. In addition to research projects, IGM promotes a greater exchange of ideas between GSB researchers and the private and public sectors. For more information, visit the IGM Web site at research.chicagogsb.edu/igm.