Should You Be Fully Invested in Stocks?

"You pay a very high price for a cheery consensus."  --Warren Buffett

One of the biggest myths in the investment world is that you have to be "fully invested" at all times. This myth, which borders on dogma, is the reason why so many money managers were fully invested in stocks when the market peaked in March 2000. It's also why many of them will be fully invested in stocks again the next time a bubble pops. For the most part, they have no choice; their clients demand it.

Here's a common statement you'll hear people recite: "Don't try to time the market."

The data I have in front of me says different.

Testing Two Strategies
I downloaded the closing prices of the Dow Jones Industrial Average for every month going back to 1930. My goal was to compare the historical long-term returns (defined as 20 years) of two strategies:

Strategy 1: Buy a basket of stocks that includes all the Dow components, but only when the index has pulled back significantly from its all-time highs. Hold onto the basket for 20 years. In between, hold cash and wait for stocks to pull back again before buying more. Don't sell until the end of the 20-year period. I call this the "buy low and hold" strategy.

Strategy 2: Buy a basket of stocks that includes all the Dow components, but only when the index is within striking distance of its all-time highs. Hold onto the basket for 20 years. When stocks go down, do nothing--wait for them to go back up before buying more. Don't sell until the end of the 20-year period. I call this the "buy high and hold" strategy.

To compare these strategies, I did the following:

1. At the close of each month in the series, calculate the Dow's highest previous month-end closing price. In other words, find the highest all-time monthly level of the Dow that investors would have seen by that date in history.

2. Calculate the percentage difference between the all-time high closing price and the price at the end of each month. For example, in September 2003, the Dow closed at 9,275. The highest all-time monthly closing price to date is 11,497, reached in December 1999. That means the September 2003 closing price was 19.3% below its all-time highest monthly closing price. I repeated this exercise for all the months going back to January 1930.

3. Sort the months into two categories: Those that were more than 15% below the previous all-time high, and those that weren't. I also threw out the decade of the 1930s because the whole decade was more than 15% below its previous high and I didn't want the Great Depression to skew the data--I don't think something so catastrophic will be repeated in my lifetime. So, I started with January 1940. (I used a 15% pullback as my cutoff because declines of this magnitude occur frequently enough to be a realistic target for investors to focus on. I could have used a 20% decline, or more, as my cutoff instead; the results would be similar, but not as statistically significant because of fewer observations.)

The Results
In roughly 24% of the months over this 766-month (63-year) period, the Dow closed more than 15% below its previous all-time high. These "low Dow" months are not evenly distributed, though; from November 1990 through August 1998, for example, stocks never pulled back 15% from their all-time high.

There are 12 "waiting periods" of more than a year within this 63-year study when you would not have been a buyer of stocks if you had waited for the Dow to drop 15% below its all-time high price. The longest of these was eight years (September 1949 through November 1957). Most of these waiting periods were very good years for stocks, and by following the "buy low and hold" strategy, you would have missed out on buying stocks while they were hitting new highs nearly every month.

Two Strategies for Long-Term Purchase of Dow  Stocks
	Average of All Months ( % )	Buy Low/ Hold Strategy ( % )	Buy High/ Hold Strategy ( % )	Buy Low & Hold Outperformance ( % )
Average 20-year return	6.75	9.1	5.9	3.2
Maximum 20-year return	14.1	14.1	13.5	0.6
Minimum 20-year return	0.8	1.5	0.8	0.7
Data through Sept. 30, 2003.

So the "buy low and hold" strategy has, over the past 63 years, produced long-term stock market returns far superior to the "buy high and hold" strategy. There's nothing too controversial in that; everyone agrees that buying low is better than buying high if you're planning to hold onto your stocks for a long time.

The controversial part is this: The common belief is that you can't identify the times when stock prices are low versus the times when they're high, so you shouldn't even try. But I've just given you one incredibly simple method of doing so. There are other, more sophisticated methods for determining whether stocks are cheap on an absolute basis (more about that later), but even if you know nothing at all about stocks, you can outperform the Dow Jones Industrial Average over long periods simply by using this "buy low and hold" strategy.

Critics will point out that, by following this strategy, there would have been significant periods when you would not have been putting new money into stocks. As a result, your overall returns would have been lower because you would have been holding a lot of cash. One could argue that staying out of the market while stocks were rising would have caused you to miss the party. That criticism seems logical, but it misses a crucial point: By following the "buy low and hold" strategy, you would not have missed out on holding stocks during periods when stocks were rising. You simply would not have put new money into the market until a 15% pullback came along. In other words, after making your first investment, you would not ever have been completely out of the market, though you would almost always have some cash sitting on the sidelines waiting for a rainy day. 

And you would have avoided buying stocks in November 1951, when the average annual return over the next 20 years was 5.9%. You also would have avoided buying in April 1971, when the index rose just 5.8% over the next two decades. Same thing for March 1962, when your return would have been 0.76%. And, of course, you would not have been buying any stocks from September 1998 through August 2001. By contrast, you would have been a heavy buyer of stocks during periods when people were spooked by the thought of owning stocks: 1940-1942, 1949, 1957-1958, 1962, 1966, 1974-1982, 1987-1988, 1990, August 1998, and every month since June 2002.

On the surface, these results appear to support a strategy of buying when stocks are well off their highs, holding on, and doing nothing at any other time. Of course, this is nothing new; Warren Buffett and Charlie Munger have been following a more sophisticated version of this strategy for decades.

Dividends and Opportunity Costs
It sure looks like a "buy low and hold" strategy performs significantly better than a "buy high and hold" strategy over long periods of time, but my results are far from conclusive.

For one thing, I ignored dividends in this study. Before the 1990s, dividends were a huge component of total returns on stocks. However, I believe the inclusion of dividends would only increase the outperformance of the "buy low" strategy because when stocks fall, dividend yields go up, so an investor would be buying at times when the dividend yield was above average. If I were to include dividends in this study, the results would most likely be skewed even more in favor of the "buy low" strategy.

Second, I didn't study how either of these strategies stacks up to dollar-cost averaging, the method whereby an investor puts a set dollar amount into stocks each month no matter what. Dollar-cost averaging has merit because it compels an investor to buy more shares of a stock (or index) when the price is low, and fewer shares when the price is high. In effect, it's a noncontroversial version of the "buy low" strategy, with a dash of the "buy high" strategy mixed in.

Third, and most important, I didn't study the opportunity cost of holding cash. An investor following the "buy low" strategy would build up a lot of cash during periods when stocks are close to their all-time highs. Since 1940, the average yield on 90-day Treasuries, a proxy for money market yields, has been about 4% (with some wide variations around that average). The average 20-year return on stocks purchased when they were close to their highs was 5.87%, roughly 2% higher than the 90-day Treasury yield.

But there's an offsetting benefit to this opportunity cost: Cash gives you the option to purchase stocks when they're down. Like all options, this one has monetary value. To put it another way, buying high means your hands are tied when stocks get cheap. As Buffett says, if you want to shoot rare, fast-moving elephants, you must always carry a loaded gun. Is this "loaded gun" worth 2% a year, compounded? Maybe, maybe not--that's a subject that seems worthy of further contemplation.

Pecking Order
If I were to rank the relative attractiveness of different long-term investing strategies from lowest to highest, it would go something like this:

1. Holding no stocks at all. Since the period from June 1930 to June 1950, over any rolling 20-year period, the Dow Index has never suffered a decline. Including dividends, stocks have nearly always outperformed bonds and inflation over a long-term period. The chance that you'll do best being out of the market altogether seems very, very small.

2. Buying high and holding. As I mentioned, the average annual return from this type of strategy has been lower than the average return of the Dow. If you factor in dividends, the gap gets even wider.

3. Dollar-cost averaging. This is a combination of the "buy low and hold" and "buy high and hold" strategies. There's nothing wrong with dollar-cost averaging, but you can do better if you're willing to spend some time watching and thinking about the market--and waiting.

4. Buying low and holding. The long-term returns are good, your chances of outperforming inflation are higher, and you will always have some cash around in the event of a sharp sell-off in stocks.

Relative vs. Absolute
As of Tuesday's close, the Dow is almost exactly 15% below its all-time high monthly closing price. So is this a good time to buy stocks? 

My answer: Yes, relative to three years ago. But the key word is "relative." Relative returns are very different from absolute returns, and this article is not meant to be a prediction about the absolute returns on stocks. I have only found evidence that a "buy low and hold" strategy will outperform the overall market, as well as other strategies, over time. But you could still do poorly owning stocks, even if you buy low and hold. You'll just do better than you would have done buying near the peak.

To get a feel for what kinds of absolute returns investors can expect in coming years, we have to look at a different type of analysis, one that focuses more on valuation instead of historical stock prices. I'll explore that subject in my next column.