Worldwide diversification is one of the most powerful factors that contributes to successful investing, and I am increasingly convinced that a combination of worldwide diversification and market timing is the closest thing anywhere to an ideal investment strategy for anybody who is retired or preparing to retire.

This article is full of numbers, mostly in tables. But it’s well worth your time to carefully and methodically go through this analysis, and it’s not as daunting as it might seem.

THE FIFTH DIMENSION

I have often said my goal in investing is to achieve the highest return with the lowest risk, the lowest cost and the highest liquidity. We believe that you can achieve this with no-load mutual funds. I now want to add a fifth dimension to this goal: "with the lowest emotional impact."

I don’t have anything against emotions, but they are rarely an investor's best friend. Almost always, emotions lead investors astray. They’ll tempt you to choose the most popular investments, most likely ones that have already risen greatly in price by the time you learn about them.

Emotions will tempt you to sell when instead you should be thinking about adding to your portfolio. And emotions will tempt you to think you can predict short-term market trends and outguess the market. In the long term, a successful investor is one who makes a good plan, then follows it regardless of the herd instinct, regardless of what's in the media — and regardless of his or her emotions.

THE CASE FOR WORLDWIDE BOND INVESTING

In the first table below, we present seven years of data, mostly total returns expressed in quarterly percentage changes, on two mutual funds, the Kemper High Yield Fund and the T. Rowe Price International Bond Fund.

We could have chosen funds that illustrate our points more dramatically. But these two are representative of U.S. and international bond funds. (In fact, we don't recommend the Kemper Fund because it charges a load. We cite it because we tracked actual accounts in this fund that used this timing discipline throughout these seven years.)

In the second and third columns of the table, we show results for the Kemper Fund with and without market timing. The fourth and fifth columns do the same for the Price International Bond Fund. The last two columns are labeled "Worldwide Bond." They show the results of combining equal amounts of these two funds with and without market timing. This 50-50 split, combined with market timing, produced amazingly good results.

Now let's start by giving the critics of market timing some ammunition. Notice that in the majority of quarters, the buy-and-hold strategy outperformed market timing. If you study the figures for the Kemper Fund, you'll see that the last time market timing beat buy-and-hold was the fourth quarter of 1990! One reason is that in many quarters we are fully in the market, just as buy-and-hold investors are. But in those quarters, our management fee of 0.375 (bond funds) to 0.5 (stock funds) percent per quarter gives our timed accounts a lower return. Investors who used our timing systems on their own would have matched the buy-and-hold results in those quarters.

As you will see in the rows marked "Total Return," and "Annual Return", over the seven years market timing outperformed buy-and-hold, and by a substantial margin — so substantial that you would not hesitate to choose which strategy you would have preferred to use. And you will see the same pattern throughout this table: with the T. Rowe Price International Bond Fund and the Worldwide Bond Portfolio, too.

Yet it didn't seem that way for much of the time in the middle of this process. When you look at the Kemper columns, you'll see that the only reason we outperformed a buy-and-hold strategy over this period is the results of fourth quarter 1989 through fourth quarter 1990. Again, you'll find the same pattern everywhere in the tables in this issue: a relatively few quarters (unfortunately, impossible to predict in advance) make all the difference.

However, high performance is not our only objective with market timing. We are also seeking lower risk. At the bottom of the tables we have included several ways of measuring risk.

THE UGLY UNDERBELLY OF INVESTING

Most of the risk measurements focus on losing quarters. This is the ugly underbelly of investing, the part that most people don't want to discuss, think about or talk about. And that's one reason most investors are ultimately unsatisfied — because they want to look only at the good parts and ignore the losses in the hope they will vanish.

Look at the line labeled "Worst 4 Qtrs." This line looks for the worst rolling four-quarter period in this seven years — in other words, the worst period of that length you would have had to live through. Notice that in each case, the worst four quarters were substantially worse for buy-and-hold investors than for those who used market timing. By the same token, in each case, the average losing quarter was significantly worse with buy-and-hold than with market timing.

You can also see that buy-and-hold investors suffered significantly more losing quarters. And notice in the last two columns that the combination of market timing and the Worldwide Bond Portfolio reduced the number of losing quarters to zero. The elimination of all losing quarters with timing was because each of the two losing quarters with Kemper happened during different quarters than the five quarters T. Rowe Price lost money.

Also note the worst quarter with buy-and-hold was Kemper's 1990 third quarter loss of 11.8 percent vs. timing's same quarter gain of 1.0 percent. The worst single quarter with timing was the third quarter 1989 loss of 1.7 percent.

Finally, turn away from losing quarters to the average of all quarters. Again you will see that in each case, the average quarter was not only better but substantially better with market timing than with buy-and-hold.

Notes for tables

Results for Kemper High Yield and Safeco Growth are based on actual timed accounts from January 1987 through December 1993. Results for T. Rowe Price International Bond are based on actual timed accounts since July 1992 and T. Rowe Price International Stock since December 1987. Prior quarters are hypothetical. All results are net of all management fees. All results include reinvestment of all dividends and capital gains. PM&A has been managing accounts since 1983, however, for the purpose of illustration, this period was selected to include at least one complete market cycle for all four funds included in the study. The study is for a select group of clients only. As we time many mutual funds with many different timing systems some clients experienced better and some clients experienced worse results for this period. These accounts were selected for illustration purposes only. Past results are no guarantee of future profitability.

THE CASE FOR WORLDWIDE STOCK INVESTING

I am going to perform magic right in front of your very eyes. Only it's not magic. By combining a U.S. equity fund and an international equity fund to form a Worldwide Equity Portfolio, we can lower your risk below that of owning either the Kemper or T. Rowe Price bond funds, and potentially raise the returns significantly.

The following table is much like the one we just saw, except it focuses on two no-load equity (stock) mutual funds. We chose T. Rowe Price International Stock Fund because of its broad international diversification and long track record. We selected Safeco Growth as a representation of U.S. growth funds. We didn't pick Safeco Growth because it is the best equity fund or the best illustration of the points we want to make. We use it because I have been applying market timing to that fund since 1983. Even if it's not the very best fund, Safeco Growth is a good fund, and we think it represents the results you are most likely to achieve if you pick an average growth fund, but not necessarily the very best.

There are some downright ugly numbers in this table. Look at third quarter 1990, the quarter when the Middle East crisis started to heat up. That was a terrible quarter to be in the market, as you can see straight across the page. In each case, market timing improved the results substantially, but not enough to eliminate losses.

Here's another point to note: there are periods when market timing will significantly underperform a buy-and-hold approach. In the case of Safeco the second and third quarters of 1987, first and second quarters of 1988, first and second quarters of 1990, first and second quarters of 1991 and the first and fourth quarters of 1993 were all good examples of times when buy-and-hold investors were better off than those who used market timing.

Now notice what happened to Safeco Growth on the buy-and-hold side in the last quarter of 1989 and the first three of 1990. Those quarters were awful for Safeco Growth Fund investors, who lost 23.7 percent of their money. Market timing reduced the loss to just under 14 percent — but that's still a significant loss.

This is a great example of why I don’t trust my money to only one asset, or even one timing system. If you lean very heavily on one type of asset — in this case the U.S. stock market — you are almost always setting yourself up for a string of bad quarters and maybe bad years. Buy-and-hold investors should know this. Those who use market timing may think they are protected from such ravages, but their protection is still flimsy if they use only one asset class and one timing system. That is why we use many timing systems and many assets. In fact, the other three equity timing systems had our clients out of the stock market and safely in money market funds prior to the outbreak of the Iraq/Kuwait problem. The average of all four timers for the third quarter in 1990 was a loss of 2.4 percent compared to the 10.1 percent loss with the single timing system.

Now let's look at another line on the table: the one showing the standard deviation. This is a very good measure of investment risk, with lower numbers indicating lower risks. (The concept is explained just below.) In every case in every table, you'll see that market timing significantly reduced the standard deviation, and thus the risk.

For just a moment, flip back to the previous table focusing on bonds, where you'll notice the same pattern: market timing reduced volatility, and worldwide diversification reduced it more.

Standard Deviation

Standard Deviation is a statistical measurement that allows you to compare the volatility of entirely different classes of things ... for our purposes, different investment assets. It is different from Beta, which measures the changes in a given item compared to a standard reference, such as the Standard and Poor's 500 Index.

Standard deviation always applies to a series of numbers. It is the range up and down from the average of the individual numbers in the series within which 67 percent of the individual numbers will fall. You can think of it as a band of probability. The lower the standard deviation, the narrower the band and the lower the volatility of numbers in the series.

For example, think of a series of numbers, the average of which is 40. If the series has a standard deviation of 10, that means two thirds of the numbers in the series will fall within 10 of the average ... or between 30 and 50. If the standard deviation were 18, then two-thirds of the numbers would fall between 22 (40 minus 18) and 58 (40 plus 18). The second series would be much more volatile than the first.

For our purposes, standard deviation gives us a good idea of where the majority of our quarters will fall. For instance in the chart, the first column shows that the Kemper High Yield fund, without timing, had an average quarterly gain of 3.1 percent with a standard deviation of 6.1 percent. The standard deviation figure means that two-thirds of the quarters fell between 9.2 percent (the average plus the standard deviation) and minus 3.0 (the average minus the standard deviation.) You will see in the next column that the same fund, with market timing, had an average quarter of 3.8 percent (a significant improvement) and a standard deviation of 4.6 percent (another significant improvement if you like low volatility).

THE CASE FOR WORLDWIDE DIVERSIFICATION

Here's where we put the puzzle together. You will notice that the first four columns of the following table are similar to the earlier tables. All show results with market timing. Each column represents one general class of asset: either bonds or equities, either domestic or international.

Our Worldwide Balanced Portfolio consists of equal parts of each of those asset classes. In other words, domestic bonds, domestic stocks, international bonds and international stocks — each gets 25 percent.

Of the four classes only Kemper High Yield Bond Fund achieved higher returns than the Worldwide Balanced Portfolio over these seven years. Furthermore, the standard deviation of the Worldwide Balanced Portfolio with timing, 3.4 percent, was lower than that of any one of its four components. In fact, that 3.4 percent was lower than all but one of the standard deviation figures you saw in the previous two tables,  the lone exception being the Worldwide Bond Portfolio with timing.

Note also that the Worldwide Balanced Portfolio had fewer losing quarters than any timed component except the high-yield bond fund. And every component, on a buy-and-hold basis, had more losing quarters than the Worldwide Balanced Portfolio with timing.

Here's another note for those who may like the worldwide diversification theme but are still skeptics regarding market timing: You can see that in the Worldwide column without timing, the combination of the four asset classes was less risky (5.4 percent standard deviation) than any class by itself measured on a buy-and-hold basis.

And even without timing, I believe the 13.2 percent annual return from the Worldwide Balanced approach is a very reasonable expectation of what is possible over extended periods of time. You can chase the 25 percent and 40 percent returns in some small-cap stock funds, and you may have a great ride for a few months, a few quarters or even a few years. But the Worldwide Balanced Portfolio is not based on fads. It's not based on recent market performance. It's not based on forecasts of where the "action" will be in the next quarter or the next year.

Instead, it is based on a diversification principal so simple that a child of 12 could understand it. Over time, we believe this combination of asset allocation and timing should be at least competitive with the Standard & Poors 500 index, and perhaps outperform it by up to 2 percent, at one-third the risk of the S&P 500.

THE MAGIC OF ASSET ALLOCATION AND TIMING

When everything seems to be falling apart, worldwide diversification can insure that you get a share of any islands of investing prosperity. In the fourth quarter of 1987, domestic and international stock markets were reeling from the October crash. Safeco Growth and the Price International Stock Fund were each off by more than 17 percent. But in the scramble to quality in that same quarter, Price's International Bond Fund was up almost 17 percent. Who would have predicted that? And of those who predicted it, how many would have transferred a significant part of their assets into international bonds at that time?

The Worldwide Balanced Portfolio, with timing, was up 5.2 percent that quarter. To appreciate that figure, think of it this way: you started with four equally weighted components, two of which were down by more than 17 percent, one of which was up by 2.1 percent and the fourth of which was up by 16.7 percent. Diversification without timing tamed the loss that quarter to 4.1 percent. That alone would have been amply appreciated as investors licked their wounds at the end of 1987. But when it was applied to those same four components, market timing produced a gain of 5.2 percent. If that's not magic, it's about the closest thing I know.

FOR THOSE WHO NEED INCOME

Many of our clients are retired and use our discipline either through the Fund Exchange or through our managed accounts, to meet their need for current income. To illustrate a typical situation, suppose you had $100,000 and you needed a return of 8 percent. If you followed the Worldwide Balanced Portfolio — without market timing — and reduced your portfolio by 2 percent at the end of each quarter, you would have withdrawn a total of $64,508 over the past seven years. At the end of that time, your holdings would be worth $135,273.

The addition of market timing would have given you income over the same seven years of $70,198 and left you with a portfolio worth $147,850 at the end of the period. In other words, if you can afford the slight fluctuations in income, taking out a percentage instead of a fixed amount can give you income while your portfolio grows at the same time. You can do this with or without timing — but timing made it better!

THE TIMING ADVANTAGE

One of the common criticisms of timing is that the strategy is commonly late getting into bull markets and late getting out after market tops. The criticism is absolutely true but there's more to the story.

Let's look at the big quarterly buy-and-hold gains and losses -- gains of more than 5 percent and losses of more than 3 percent. The next table reflects all of those "emotional" quarters for the four funds on a buy-and-hold basis and the results with timing for the same quarters.

Due to the late entry into advancing markets you will note the timing results averaged 2.1 percentage points less than the buy-and-hold in the profitable quarters. On the other hand, the timing results outperformed buy-and-hold by an average of 7.6 percentage points in the losing quarters.

My conclusion: Timing substantially reduced losses in the worst times and nearly matched the big gains of buy-and-hold in the best times. That is exactly what market timing is supposed to do.

THE ULTIMATE INVESTMENT

Below we present all the information from the earlier tables in three meaningful charts. For those of you who get bored with a bunch of numbers, these charts make it easy to visualize the differences between buy-and-hold and market timing.

The first chart represents the four funds without the attempt to defend against market declines. This buy-and-hold strategy produces a relatively high level of volatility and investor anxiety.

The second chart  represents a "kinder and more gentle" approach to investing. You will note the peaks of market timing may not be as high as those of buy-and-hold but the valleys are much less severe.

For me, the ultimate investment, when plotted on a graph, would show a straight line, always going upward — and going up fast enough to achieve a double-digit return or whatever your long term goal might be. The only investment that doesn't fluctuate in real life is a fixed-income instrument that is really fixed, such as a U.S. Treasury Bill. The problem, of course, is that Treasuries simply don't produce the kind of returns most of us want and need. However, as you can see from the charts, the combination of the Worldwide Balanced Portfolio and market timing is extremely close to a straight line going up at a double-digit rate. If you're looking for a 10-15 percent return over a long time span, I haven't found a better way to achieve it than this. The numbers here are our evidence.

JUST SHOW ME THE BOTTOM LINE

How many times have we heard someone say "Just show me the bottom line"? The second table early on in this article (The Case for Stock Fund Timing) showed that Safeco Growth without timing outperformed Safeco Growth with timing on a total return basis. However, the difference was minuscule (149.3% vs. 155.2% or .4% per year on a compound basis). Comparing only the "bottom line" it may be difficult to determine which is the better strategy. Let's recreate the experience most investors would have faced, and their emotional response, and see which strategy you prefer.

Assume you decided to give timing a try with Safeco Growth starting with the first quarter of 1987 (second and third columns in the chart). By the end of the first quarter you realize, "I didn't need timing to get the 20% return." At the end of the second quarter you're really dumbfounded. On a buy-and-hold basis you would have made 1.3 percent but with the "magic" of timing, due to an unprofitable trade, you actually lost 3.6 percent. So far you don't have much good to say about timing.

By the end of the third quarter you're really questioning your decision to use timing because buy-and-hold outproduces timing by 5 times, 5.9 percent vs. 1.1 percent. By this time you might have gone back to the traditional buy-and-hold approach just in time to lose 17.9% in the stock market crash of 1987, while timing actually gained 1.2% during the same period. All of a sudden timing has justified itself by accomplishing what it is designed to do during a declining market.

But timing's success is short-lived as the following quarter buy-and-hold clobbers timing by almost 10 times, 10.4 percent vs. 1.1 percent.

As you continue to compare the two disciplines over the entire period you will find more of the same. Buy-and-hold wins in most of the profitable quarters while timing tends to add value in the unprofitable quarters.

In the end, timing produced virtually the same return as buy-and-hold at one-third less risk.