Efficient Frontier
Efficient Frontiers Logo
William J. Bernstein

The Two-Percent Dilution

Over the past two centuries, common stocks have provided a sizeable risk premium to U.S. investors. For the 200 years from 1802 until 2001, inclusive, the returns for stocks, bonds, and bills were 8.42%, 4.88%, and 4.21%, respectively. In the most simplistic terms, the reason is obvious: a bill or a bond is a promise to pay interest and principal and, as such, its upside is sharply limited. Shares of common stock, on the other hand, are claims on the future dividend streams of the nation’s businesses. Instead of a fixed, paltry trickle from low-risk fixed-income securities, the ever-increasing fruits of technologically-driven economic growth fall to the shareholders.

Viewed over the decades, the national economy grows with remarkable uniformity. Figure 1 plots the real GDP of the U.S since 1820:

 

During this period, real GDP has grown fairly evenly at about 3.6% per year. The long-term uniformity of economic growth is both a blessing and a curse. It is reassuring to know that real U.S. GDP doubles once every 20 years (and real per-capita GDP once per generation). But it is also a dire warning to those predicting a rapid acceleration of economic growth from the computer and Internet revolutions. Such extrapolations of technologically-driven increased growth are painfully oblivious to the broad sweep of scientific and financial history. The impact of recent advances in computer science pales in comparison to the technological explosion that occurred between 1820 and 1850. This earlier era saw the most deep, far-reaching technology-driven changes in everyday existence throughout human history. They profoundly affected the lives of those from the top to the bottom of the social fabric in ways that can hardly be imagined today. At a stroke, the speed of transportation increased tenfold, and communications became almost instantaneous. Before 1820, people, goods, and information could not move faster than the speed of the horse. Within a generation, journeys that had previously taken weeks and months now involved an order of magnitude less time, expense, danger, and discomfort. Important information could be instantaneously transmitted. Put another way, the average inhabitant of 1800 would have found the world of 50 years later incomprehensible, whereas a person transported from 1950 to 2000 would have little trouble understanding the relatively small intervening changes in everyday life.

The comparatively uniform increase in GDP also implies a similar uniformity in the growth of corporate profits, which is, in fact, the case. Figure 2 demonstrates that, except for the Great Depression when net corporate profits disappeared, aggregate company earnings have remained fairly constant at about 10% of nominal GDP:

Should it not follow that stock prices also grow at the same rate? After all, there has been a direct relationship between corporate profits and GDP since 1929. Unfortunately for the shareholder, earnings and dividends will keep up with GDP only if no new shares are created. Since 1871, real stock prices have grown at 2.48% per annum versus 3.45% for the GDP (the slightly slower growth rate for the more recent period reflecting the slowdown in population growth). There has thus been about 1% per year of "slippage" between stock prices and GDP. Further, as we shall see, the true degree of slippage is quite a bit higher, since much of the 2.48% rise in real stock prices after 1871 was due to an upward revaluation, as the highly illiquid industrial stocks of the post-Civil War period, selling at three to four times earnings, gave way to instantly and cheaply tradable common shares selling that much more dearly.

This slippage is the result of the net creation of shares, as existing and new companies capitalize their businesses with equity. It suggests a very simple paradigm for measuring the degree of slippage: the ratio of the proportionate increase in market capitalization to the proportionate increase in price. For example, if over a given period, the market cap increases by a factor of ten, and the cap-weighted price index increases by a factor of five, then there has been 100% net share issuance in the interim. More formally,

Net Dilution = {(1+c)/(1+r)} – 1

where c = capitalization increase, and r = price return

This relationship has the advantage of factoring out valuation changes, as they are embedded in both the numerator and denominator. Further, it holds only for universal market indexes such as the CRSP 1-10 or the Wilshire 5000, since less inclusive indexes can vary the above ratio simply by adding or dropping securities. In Figure 3, we plot the total market cap and price index of the CRSP 1-10, with 1926 equal to 100:

Note how market cap slowly and gradually pulls away from market price. By the end of 2001, the cap index has grown 4.97 times larger than the price index, suggesting that for every share of stock extant in 1926, there are now 4.97 shares! To give a better idea of how this has proceeded over the past 75 years, in Figure 4 we plot this dilution index, defined as the cumulative net creation of new shares:

These data are consistent with a nearly continuous net dilution of common shares. The process is seen to have been more rapid during the late 1920s, quickly decelerating after the crash of 1929. As capital costs rose in the 1970s, it slowed yet further, and during the late 1980s, there was even a brief net contraction as companies responded to peak capital costs with stock buybacks. However, in the 1990s, shares again began to dilute. The overall rate of dilution since 1926 is 2.15% per year.

The slippage between aggregate economic data and per-share performance can be independently examined by comparing the rise in per-share corporate dividends versus GDP growth around the globe. Recently Dimson, Marsh, and Staunton, in their wonderful monograph, The Triumph of the Optimists, have examined the real dividend-growth rates in 16 nations over the entire 20th century. These can then be compared to the growth of real GDP and real per-capita GDP growth rates.

We divide these nations into two categories according to the degree of devastation visited upon them by the calamities of the 20th century: Group 1, which suffered no substantial destruction of their productive physical capital during World Wars I and II and the Spanish Civil War, and Group 2, which did.

Group 1

Real GDP

Real Per

Capita GDP

Div Growth

Growth

Dilution

Growth

Dilution

Ireland

-0.80%

2.21%

3.01%

2.05%

2.85%

Switzerland

0.10%

2.66%

2.56%

1.85%

1.75%

Canada

0.30%

3.87%

3.57%

2.07%

1.77%

UK

0.40%

1.89%

1.49%

1.44%

1.04%

US

0.60%

3.28%

2.68%

1.96%

1.36%

Australia

0.90%

3.29%

2.39%

1.60%

0.70%

S. Africa

1.50%

3.49%

1.99%

1.16%

-0.34%

Sweden

2.30%

2.62%

0.32%

2.05%

-0.25%

Average

0.66%

2.91%

2.25%

1.77%

1.11%

Group 2

Real GDP

Real Per

Capita GDP

Div Growth

Growth

Dilution

Growth

Dilution

Japan

-3.30%

4.11%

7.41%

2.99%

6.29%

Italy

-2.20%

2.96%

5.16%

2.40%

4.60%

Denmark

-1.90%

2.86%

4.76%

2.09%

3.99%

Belgium

-1.70%

2.15%

3.85%

1.72%

3.42%

Germany

-1.30%

2.78%

4.08%

1.79%

3.09%

France

-1.10%

2.37%

3.47%

1.99%

3.09%

Spain

-0.80%

2.79%

3.59%

2.00%

2.80%

Netherlands

-0.50%

2.96%

3.46%

1.80%

2.30%

Average

-1.60%

2.87%

4.47%

2.10%

3.70%

The first column in each table tabulates the growth of real per-share dividends in each nation between 1900 and 1998, and the second, the growth of real aggregate GDP for the full century. The third column tabulates the difference between the two. It is noted that in all nations per-share dividends grow more slowly than aggregate GDP. The gap is lowest in Sweden at 0.32% per year, and more than 2% per year in five of the eight Group 1 nations, including the U.S, where it was 2.68%. This is close to the 2.15% value obtained by the market-cap/market-price model. It is even closer to the 2.25% average for the Group 1 nations. The fourth and fifth columns do the same for per-capita GDP, where gaps of 1.11% and 3.70% are found for Groups 1 and 2, respectively.

The data for Group 2 are striking: Amazingly their economies repaired the devastation wrought by the 20th century, with overall GDP and per-capita GDP growth rates equivalent to Group 1. The bad news is that the same cannot be said for per-share equity performance; there was almost 4.5% slippage between the growth of their economies and per-share corporate payouts.

It thus seems that in "normal nations" of Group 1—those untroubled by war, political instability, and government confiscation of the economic commanding heights—the ongoing capitalization of new technologies produces a net dilution of outstanding shares of about 2% per year. (Did I hear anybody say "stock buybacks?" Ah, then I’ve some wonderful vacation plots in the Everglades to show you.) The Group 2 nations represent a more fascinating phenomenon. These can be thought of as experiments of nature in which physical capital is devastated and must be rebuilt. Fortunately, it is much harder to destroy a nation’s intellectual, cultural, and human capital; within little more than a generation, GDP and per capita GDP catches up with, and in some cases surpasses, the Group 1 averages. Unfortunately, this requires a high rate of equity recapitalization, reflected in the large dilutions seen in columns 3 and 5, and which mulcts existing shareholders.

This analysis has disturbing implications for paradigmistas convinced of the revolutionary nature of biotech, the Internet, and personal computers. It may very well be that a rapid rate of technological change could, in effect, turn a Group 1 nation into a Group 2 nation, as an increased rate of obsolescence destroys the economic value of plant and equipment as surely as bombs and bullets. The resultant recapitalization would then dilute per-share payouts much faster than the technology-driven acceleration of economic growth, if such acceleration exists at all.

But whatever the true nature of the interaction of technologic progress and per-share earnings, dividends, and prices, it will come as an unpleasant surprise to many that even in the Group 1 nations, average real per-share dividend growth was only 0.66% per year; for the Group 2 nations, it was strongly negative.

Thus, at the dawn of the new millennium, the equity investor cannot expect a real return greatly in excess of a generally derisory dividend yield. Nor will he be rescued by more rapid economic growth, which is unlikely to occur. But even if it does, its benefits will undoubtedly be more than offset by the dilution of his ownership interest necessitated by technologically-driven increased capital needs.

To Efficient Frontier Homepage E-mail to William Bernstein

Copyright © 2002, William J. Bernstein. All rights reserved.

The right to download, store and/or output any material on this Web site is granted for viewing use only. Material may not be reproduced in any form without the express written permission of William J. Bernstein. Reproduction or editing by any means, mechanical or electronic, in whole or in part, without the express written permission of William J. Bernstein is strictly prohibited. Please read the disclaimer.