NEW YORK, NY (Dec. 20, 1999) -- After last week's digression on tax matters, I'd like to pick up where I left off two weeks ago, when I argued that being fully invested is appropriate for certain people despite the current market's high valuation levels. This argument rests largely on historical data -- in particular, data from Jeremy Siegel's book, Stocks for the Long Run. Let me reprint the data I cited in the conclusion of my column on November 22:
  1. Since 1870, stocks have had positive returns in more than 90% of 123 rolling five-year periods.

  2. The after-inflation annual returns from stocks were at least 2.8% in 90% of 10-year periods, and at least 5.3% in 90% of 20-year stretches.

  3. Over 20-year periods back to 1802, stocks have never generated less than 1.0% annual returns above inflation, and stocks have never failed to beat inflation over any period of 17 years or longer.

  4. Over five-year periods back to 1802, stocks outperformed bonds and T-bills more than 70% of the time; over 10 years, more than 82% of the time; over 20 years, more than 94% of the time, and over 30 years, 99.4% of the time (and 100% of the time since 1870).
Given the widespread fear that we are currently at or near a market peak, Siegel provides one more important piece of data: Had one been unfortunate enough to invest $1,000 in stocks just before the six worst market crashes this century, one's long-term returns would still have prevailed strongly over bonds and T-bills, as the following chart shows:

30-Year Real Returns of a $1000 Investment After Market Peaks

* 1901, 1906, 1915, 1929, 1937, 1966

And the returns from stocks were even better in the 30 years after August 1929, just before stocks fell 80% during the Great Depression:

30-Year Real Returns of a $1000 Investment in August 1929

Siegel concludes from his data that "The upward movement of stock values over time overwhelms the short-term fluctuations in the market. There is no compelling reason for long-term investors to significantly reduce their stock holdings, no matter how high the market seems."

The Argument Against Using History as a Guide

Siegel makes a powerful argument, but it should not be accepted blindly. While history can be a valuable guide, one must be careful not to misunderstand and misapply its lessons.

Consider those who reduced their stock holdings after the S&P rose 37% and 22% in 1995 and 1996, respectively. This appeared to be a very reasonable strategy, given that the market had never risen more than 20% for three consecutive years. In fact, going back to 1950, in the four occasions when the S&P had risen more than 20% for two consecutive years, the average return the next year was 6.0%, and in the 11 occasions when the S&P had risen more than 10% for two consecutive years, the average return the next year was 1.0%. In addition, going into 1997 the S&P's high price-to-earnings and price-to-book ratios and low dividend yield all showed that it was at dangerously high levels. Yet those who withdrew from the market made an expensive mistake, since the S&P has doubled since then.

Another classic example of what can happen when the lessons of history are misapplied was the collapse of the junk bond market in the late 1980s. Warren Buffett provided an excellent description in the 1990 Berkshire Hathaway shareholder letter:

"At the height of the debt mania, capital structures were concocted that guaranteed failure: In some cases, so much debt was issued that even highly favorable business results could not produce the funds to service it.... When these misdeeds were done, however,... investment bankers pointed to the 'scholarly' research of academics, which reported that over the years the higher interest rates received from low-grade bonds had more than compensated for their higher rate of default. Thus, said the friendly salesmen, a diversified portfolio of junk bonds would produce greater net returns than would a portfolio of high-grade bonds. (Beware of past-performance 'proofs' in finance: If history books were the key to riches, the Forbes 400 would consist of librarians.)

"There was a flaw in the salesmen's logic -- one that a first-year student in statistics is taught to recognize. An assumption was being made that the universe of newly minted junk bonds was identical to the universe of low-grade fallen angels and that, therefore, the default experience of the latter group was meaningful in predicting the default experience of the new issues. (That was an error similar to checking the historical death rate from Kool-Aid before drinking the version served at Jonestown.)

The universes were of course dissimilar in several vital respects...."

A similar argument could be made to cast doubt on the relevance of historical data to predict the future returns of today's stock market. A reader of my earlier column, Wayne Crimi, summed it up well:

"In my view it's an error to use Jeremy Siegel's statistics to make the case that long-term investors should remain fully invested at this time. Those statistics assume that the average conditions of the past are equal to the conditions of the present. That isn't the case.

"For much of the historical record, interest rates were much lower than those that prevail today. Treasury yields in the 2%-4% range were very common. Simultaneously, P/E ratios averaged between 10-15 times normalized earnings for much of stock market history. Given an earnings yield in the 7%-10% range, a dividend yield in the 4%-5% range, similar GDP growth, and a very low bond yield, is there any wonder that stocks destroyed bonds over the long term? (And even the relatively short term.)

"The situation is reversed today. Bond yields are in the 6%-7% range and the earnings yield is in the 3% range. There is no precedent for this set of conditions. The closest we can get is the 1987 peak. Even in 1929 the earnings yield was significantly higher than the bond yield (and ROE was also around 20%). The late-'60s mania had similar interest rates, but the P/E of the S&P 500 was around 18 vs. 30+ today. Right now the P/E on the Nasdaq is well over 100 and there are literally hundreds of stocks with market caps of many billions that are reporting losses (with no end in sight). The '90s is the greatest mania in history by far.

"In order to come to the conclusion that the correct course of action for long-term investors is to remain fully invested based on historical experience, we would have to see the results of a number of similar situations in the past. We simply do not have the data to do so. It's a risk-reward call.... To be fully invested in the S&P 500 or some other such mutual fund is an extraordinarily poor bet in my view. A mean reversion or even close at any time will wipe out well over a decade's worth of growth and dividends. Even the best-case scenario will probably only marginally beat bonds. The public is simply making a horrible bet."

Warren Buffett himself expresses similar sentiments in a recent Fortune magazine article -- which is one reason he's sitting on $36 billion of cash and bonds (for perspective, that's greater than the market capitalization of all but 68 companies in the S&P 500).

My Opinion on the Debate

I think Buffett and Crimi could be absolutely right, which is a major reason why Berkshire Hathaway is my largest holding. However, I've been concerned about a stagnant market -- or much worse -- for a number of years now, but fortunately I've remained fully invested, and expect that I will continue to be so. Why?

First, while today's market is different in many ways from historical bull markets, I believe there are enough similarities such that the lessons of history still apply. For example, I would argue that the Nifty 50 of 1972 was as richly valued as a comparable group of stocks today. Click here to see the data from both periods. Though the original Nifty 50 initially suffered a huge decline, these stocks over 25 years generated robust 12.7% annual returns.

Furthermore, as the charts above show, it's not as if stocks' long-term returns -- even after market peaks -- barely edge out those of other asset classes. Stocks' returns have been dramatically higher. Thus, I believe that even if today's market proves to be more overvalued than those in the past, equities will still be the best investment over long periods of time.

Second, I am much less concerned about the market than I am about particular stocks. As long as I can find a few outstanding companies priced attractively, I will be fully invested. Buffett said it best in his 1994 Berkshire Hathaway shareholder letter:

"We try to price, rather than time, purchases. In our view, it is folly to forego buying shares in an outstanding business whose long-term future is predictable, because of short-term worries about an economy or a stock market that we know to be unpredictable. Why scrap an informed decision because of an uninformed guess?"

By refusing to take money off the table in anticipation of a market downturn, I accept that my portfolio will decline during the many corrections that are certain to occur during my investment lifetime. But I don't see a better alternative. As Winston Churchill wrote, "It has been said that democracy is the worst form of government except all the others that have been tried." I feel similarly about being fully invested. I am simply not willing to try to predict the top of this remarkable bull market (or any other for that matter).

My competitive advantage as an investor is not timing the market -- the game many others are playing. Rather, it's being patient and focused on the long run, doing good analyses, and making a few big bets when I believe the odds are heavily in my favor.

-- Whitney Tilson

P.S. A quick follow-up to last week's column, which outlined the "double-swap" and "swap-and-hold" strategies for realizing a capital loss on a stock an investor still wants to hold. One reader wrote:

"You mention only selling a loss first, then buying back after 30 days. If the investor has a cash position, or cares to use margin, he can buy the stock he intends to sell first, then wait out the 30 days and sell his old position to restore his cash position or cover his margin. If the stock moves during that period, both lots will be moving, insulating him from the fluctuations you recite as a downside. Unless the position is large, the interest on a 30-day margin loan or interest loss on a 30-day withdrawal from a money market would not be large."

Good point, though it's too late to use this strategy for this year. Also, I'd only suggest using it with excess cash, as margin is risky. Finally, while the investor enjoys a double benefit if the stock rises, there's a double loss on the downside as well.

Whitney Tilson is Managing Partner of Tilson Capital Partners, LLC, a New York City-based money management firm. Mr. Tilson appreciates your feedback at To read his previous guest columns in the Boring Port and other writings, click here.