The survivorship bias argument against stocks for the long run is that U.S. stocks have done well, but many other stock markets have done terribly. Consider an investor in East Germany who patiently invested his or her money into stocks.
Our East German stock investor would have lost every penny when the communists took over East Germany after World War II. For an East German, stocks for the long run would have meant a complete loss. Consequently, there are no Jeremy Siegels promoting East German stock investments.
U.S. stocks are the Michael Jordans of the investing world. No sane person would decide how much time to practice basketball with the assumption that she or he could be the next Michael Jordan. Similarly, no one should decide how much to invest in stocks based solely on the history of U.S. stocks.
Thus, in addition to the U.S. stock market performance, we should consider the outcome of other markets. Consider some of the debacles that faced equity investors in major markets around the world just in the last century. Russia, China, and much of Eastern Europe became communist with state ownership of much or all business. Germany suffered through hyperinflation and destruction in World War II. The allies in World War II similarly destroyed Japan. By some measures Argentina was richer than France a century ago yet has had severe economic struggles since then.
Some people argue that the poor performance of stocks in countries like East Germany is irrelevant. The argument goes something like, “Equity investments in a place and time where property rights are not respected (or are so heavily taxed) will always be a poor bet. Communist markets of any kind by their nature have to be inherently fictional—a government made-up number.”
So is it fair to include East Germany? The answer depends on the time period you analyze. No sane investor would have wanted to own East German assets after the country was taken over by the Soviet Union. The investor back in 1802, however, had no idea that part of the world would become communist (Karl Marx was born in 1818) and part of the communist world would be called East Germany. So the part of the world that ended up becoming East Germany belongs in the analysis of the stock returns from 1802 (the beginning of Professor Siegel’s U.S. data), and excluding it would be committing an error of survivorship bias.
Interestingly, one doesn’t have to go all the way to communist countries to see that just looking at the United States is a form of survivorship bias. In Triumph of the Optimists: 101 Years of Global Investment Returns, Elroy Dimson, Paul Marsh, and Mike Staunton argue that most analyses of stock market returns are overly optimistic.3 Most studies, the authors argue, ignore places or time periods where stocks did very poorly. The authors study 16 major countries (the United States plus 15 others including the United Kingdom) and write that standard analysis “has provided investors with a misleadingly favorable impression of long-term equity performance.”
The argument of survivorship bias has also been performed rigorously in academic papers. These results suggest that part, but not all, of the advantage of U.S. equities is a result of survivorship bias.4 U.S. stocks have been both lucky and good. It would be naïve to expect the future of U.S. stocks to be as bright as the past.
Why Jeremy Siegel Does Not Live in Nuclear Winter without Electricity
Our examination of survivorship bias is worthwhile but incomplete. Students of Eastern philosophy are told, “The Zen that can be taught cannot be real Zen.” Similarly, the survivorship bias that can be measured cannot be the real survivorship bias. To understand the profound problem in measuring survivorship bias, we can gain perspective from the field of cosmology. This field studies such deep issues as the formation of the universe.
The analysis of survivorship bias in cosmology is motivated by the following observation: Our universe is built with certain basic properties including the speed of light and other parameters that most of us have never heard of such as Planck’s constant. Everything in the universe is affected by these parameters. The amazing thing is that scientists have determined that if these basic aspects were even slightly different, then life would be impossible.
Stephen Hawking summarizes this in A Brief History of Time: “The remarkable fact is that the values of these numbers seem to have been finely adjusted to make possible the development of life.”5 One obvious explanation for our good fortune is that God or another intelligent force designed the universe. Scientists have developed the “anthropic principle” either as an alternative to intelligent design or in order to understand the details of intelligent design. The anthropic principle comes in a weak form and a strong form. Stephen Hawking describes the two versions of the anthropic principle:
The weak anthropic principle states that in a universe that is large or infinite in space and/or time, the conditions necessary for the development of intelligent life will be met only in certain regions that are limited in space and time. The intelligent beings in these regions should therefore not be surprised if they observe that their locality in the universe satisfies the conditions that are necessary for their existence. It is a bit like a rich person living in a wealthy neighborhood not seeing any poverty.6
According to the strong anthropic principle,
there are either many different universes or many different regions of a single universe, each with its own initial configuration and, perhaps, with its own set of laws of science. In most of these universes the conditions would not be right for the development of complicated organisms; only in the few universes that are like ours would intelligent beings develop and ask the question, “why is the universe the way we see it?” The answer is then simple: if it had been different, we would not be here!7
Most efforts to estimate survivorship bias in stocks ask the equivalent of the weak form of the anthropic principle. They test the outcome of an average stock investment around the world, not just the United States. In the case of Hoop Dreams this weak form of survivorship bias requires us to look at all 7,600 high school basketball players for each one who makes it to the NBA.
The strong form of survivorship bias is much more fundamental. It asks us to consider not just stock markets in other countries, but also other possible global histories. This sort of question is most often investigated in movies, particularly those that involve time travel. In Back to the Future, Marty McFly (played by Michael J. Fox) travels back to the time when his parents were in high school.
In his effort to get back to the future, Marty changes the outcome and jeopardizes his family. As Marty looks at a family snapshot that he carries in his wallet, he sees the image of his siblings—and then of himself—begin to dissolve as his actions in the past change the future. By the end of the movie, of course, the happy outcome includes a different, improved outcome for the modern McFlys.
With regard to U.S. stock returns, the strong form of survivorship bias suggests that not only has the United States been the luckiest of all countries, but also that the whole world has been extremely lucky. Just as small actions by the time-traveling Marty McFly had big effects on the future, “small” events that did not happen in the history of this world could have destroyed the value of stocks.
Among the most obvious bad thing that didn’t happen is nuclear war between the United States and the Soviet Union. This avoidance of war seems to have happened with a large dose of luck. Recently released documents from both sides during the Cuban Missile Crisis reveal just how close we came to global nuclear war. During one confrontation between U.S. and Soviet ships, two Soviet commanders gave orders to launch nuclear weapons.
This near nuclear miss is described by Noam Chomsky as follows:
We learned that the world was saved from nuclear devastation by one Russian submarine captain, Vasily Arkhipov, who blocked an order to fire nuclear missiles when Russian submarines were attacked by US destroyers near Kennedy’s “quarantine” line. Had Arkhipov agreed, the nuclear launch would have almost certainly set off an interchange that could have “destroyed the Northern hemisphere,” as Eisenhower had warned.8
How much have U.S. stocks benefited by global luck? This is extremely difficult to ascertain because it requires assigning probabilities to events that didn’t happen. Furthermore, such analysis should include all possible alternative scenarios including those that would have exceeded the actual outcome.
Presumably, humans could have had an even better outcome than we realized. For example, we could live in a world without holes in the ozone layer, where Princess Diana is still alive and married to Prince Charles, and where there are no nuclear weapons that can be hijacked and used to hold the world hostage.
How can we quantify these alternate possible histories? The mathematical tool of Monte Carlo simulations allows one to run the calculation, but still depends on the assumptions. Some of the best efforts to generalize this stronger form of survivorship bias have been done by Nassim Nicholas Taleb and are contained in academic papers, his first book Fooled by Randomness, and his forthcoming book The Black Swan.9
Taleb concludes it is likely that all of the excess return of U.S. stocks is due to luck, not skill. While we can’t know the answer for sure, I lean toward believing that things have turned out better than we could have expected.
Putting both weak and strong forms of survivorship bias into the analysis of U.S. stocks requires adding some lines to Professor Siegel’s table. The amended data might look something like what we see in Table 8.2.
So where do we stand on assessing the elements of luck and skill in historical U.S. stock market performance? U.S. stocks have done better than stocks in other countries. Furthermore, it seems likely that the world has been lucky in avoiding any global wars, nuclear or otherwise, since 1945. Both of these suggest that luck played a significant, and perhaps dominant, role in the fantastic performance of U.S. stocks.
TABLE 8.2 U.S. Stocks Have Benefited from Good Luck
Source: Stocks for the Long Run, second edition
U.S. Stocks Have Survived. Are They Expensive Today?
The historical survivorship analysis suggests that the amazing performance of U.S. stocks is unlikely to be repeated. Buying stocks today is not the easy choice that it would be if we had a time machine and could go back into U.S. history.
My father is a physician of the old school. He likes to mock modern-day physicians and their reliance on technology. He jokes, “If all else fails, let’s look at the patient.” His meaning is, of course, that we ought to start by looking at the patient. Similarly, we are several pages into a discussion of stocks without actually looking at any stocks. Let’s look at the financial patient and analyze some stocks.
We begin with Microsoft and then move to the entire S&P 500. Microsoft is one of the most profitable companies in the world and consequently one among companies with the highest stock market value. It is also part of all the major financial indices including the Dow Jones Industrial Average, the S&P 500, and the NASDAQ 100. Beyond financial strength, Microsoft generates strong emotions in many people, ranging from frustration with Windows glitches to the joys of a beautiful Excel spreadsheet.
When it comes to investing, we put aside our feelings and look at the numbers. Perhaps the most common analysis uses the so-called “Fed model” that compares a stock to the 10-year Treasury bond. The 10-year Treasury bond is a safe alternative to risky stock investments. Figure 8.1 shows the projected return on a $100 investment in the next year for Microsoft stock and for the 10-year Treasury bond.
Investing $100 in a Treasury bond earns $4.40 in interest per year (at current rates of 4.4%). The person who puts $100 into Microsoft stock (at $28) will receive a dividend of $1.14. In addition, $100 of Microsoft stock buys ownership in additional profits that will be retained by the company. This figure for 2005 at Microsoft is expected (by Wall Street analysts) to be $3.43 for every $100 investment.
So is Microsoft stock a good investment? The answer is that it depends on your optimism about the future. The T-bond is going to pay $4.40 a year for 10 years and then you will get back the $100 investment. Microsoft stock could be much better or much worse. The payoff to Microsoft is the fact that it could grow substantially. Microsoft is far riskier, however, than the government bond and even big companies can go bankrupt. While the bond investor can be pretty confident of getting back the original $100, the stock investment contains an element of speculation.
FIGURE 8.1 Microsoft vs. 10-Year T-Bond (1-Year Return on $100 Investment)
Sources: Federal Reserve, Microsoft, The Wall Street Journal
FIGURE 8.2 S&P 500 vs. 10-Year T-Bond (1-Year Return on $100 Investment)
Sources: Federal Reserve, Standard & Poor’s, Goldman Sachs & Co.
Is the risk of stock investing worth the reward? Before addressing this question, Figure 8.2 shows the same calculation for the S&P 500.
A $100 investment in the S&P 500 is estimated to return $5.92 in expected 2005 profits versus $4.57 in Microsoft profits. So the stock market places a premium valuation on Microsoft profits, presumably reflecting the superior value of the company’s franchise. Beyond this difference, an investment in the S&P 500 also yields more money today in dividends than a similar investment in Microsoft.
This Fed model framework provides an easy summary of optimistic and pessimistic views on U.S. stocks.
The pessimistic view of stocks (Figure 8.3) is that earnings will be (or actually are) lower than projected. The negative view on earnings includes both a pessimistic view on the economy as well as the way in which earnings are calculated. Accounting rules still allow for earnings games related to issues including stock options and pensions. In addition, most pessimistic stock analyses include an expectation of higher interest rates. Thus, the pessimist contrasts lower stock returns to higher bond interest rates and concludes that stock prices are too high.
FIGURE 8.3 A Pessimistic View of Stock Prices (Return on $100)
Sources: Federal Reserve, Standard & Poor’s
The optimistic view centers on the growth in earnings. The optimistic chart (Figure 8.4) assumes that the economy will grow modestly and that corporate profits will grow at the same rate as the economy. As we’ll see, profits have had a good run where they have done better than the economy, but even with the more conservative assumptions, it is easy to build a positive case for stocks. (Figure 8.4 assumes 7% annual growth in the economy, which could come from 3% growth in productivity, 3% inflation, and a 1% growth in the population.)
Even though this Fed model view of stocks avoids many of the details, we can already draw some conclusions about stock valuations.
Stock Prices Do Not Look Ridiculously High
Stock and bond returns are about equal. Good arguments can be made to suggest that stocks are expensive or cheap. This balance of possible reward and risk is characteristic of fair value.
FIGURE 8.4 An Optimistic View of Stock Prices (Return on $100)
Sources: Federal Reserve, Standard & Poor’s
Stock Prices Are Not Cheap
While stocks do not look terribly expensive, by most measures they are priced above historical averages and far above the valuation levels that existed at true market bottoms. So stocks appear to be either fairly valued from a current perspective, or richly valued from a historical perspective. It is hard to characterize stocks as really cheap.
Growth Rate Is Crucial to Determining Stock Valuations
The expected growth rate is enormously important to valuations. In the case of Microsoft, a $100 investment earns less today than the same investment in a 10-year Treasury bond. Furthermore, investors in Microsoft stock risk losing part or all of their money. To make the risk worth the reward, Microsoft earnings must rise over time to be substantially above the payoff to the ultra-safe Treasury bond. This question of earnings growth is so crucial that we will examine it further.
Beyond simple valuation calculations, emotional mood swings play a major role in stock prices. In some periods, investors are willing to take risk and predict good futures. In other periods, investors are skeptical and want cash no
w to compensate for risk. In The Great Crash, John Kenneth Galbraith uses the word “bezzle” to describe an aspect of this mood swing. Bezzle is derived from embezzle, and Galbraith describes the ability of people to siphon funds out of companies and the world more generally. In good times, investors (and perhaps regulators) are optimistic and lax. Therefore, the bezzle increases. In contrast, Galbraith writes, “In [economic] depression, all this is reversed. Money is watched with a narrow, suspicious eye. The man who handles it is assumed to be dishonest until he proves himself otherwise. Audits are penetrating and meticulous. Commercial morality is enormously improved. The bezzle shrinks.”10
The overall pattern of optimism and pessimism is a recurring theme in investments. When everyone is optimistic about stocks, they tend to be worse investments. Doom and gloom generally predict good stock market returns. These emotions get reflected in stock prices. Consider that the S&P 500 investor today accepts about $1.81 in dividends and for-goes $4.40 in sure interest. Why accept a lower return? The lower return today is accepted because investors are optimistic about the growth of dividends. In other periods, stock investors have been so skeptical as to require that dividends exceed bond interest.
Mean Markets and Lizard Brains Page 19