Second Edition

by Ray Schumacher

Author and Originator of Style Index Investing

Copyright May 1, 2010 

The goal of most investors is to maximize their return within a level of risk they can comfortably tolerate.  This usually involves some combination of stocks, bonds, and cash equivalents.  While the long range return of stocks is twice that of bonds and more so of cash equivalents, over shorter periods stock returns can be far less, and are sometimes negative for months on end.  Because of this they are regarded as being far more risky than bonds or cash equivalents.

These periods of lower or negative stock returns are ones that test one's tolerance for risk, and unfortunately this level often turns out to be far lower than envisioned back in "good" times.  These "bad" times can lead the inexperienced investor to panic, often selling out on the eve of a turnaround in stock prices.  The trauma of having turned large paper losses into real losses often results in the investor swearing off of stocks for good, and missing their long range advantage over the safer alternatives.

Knowing one's risk tolerance in advance is an important prerequisite to successful investing and worth the serious thought needed to answer: "How much am I willing to see my portfolio shrink on paper before I start losing sleep?"  The smaller this allowed shrinkage on paper, the smaller should be one's inclusion of stocks in the portfolio.

Experienced investors who have been through these cycles before know their tolerance for risk, and understand that periods of negative stock returns can last for many months.  It will likely be a year or more before prices return to their prior level.  But recover they will, and his or her job is to be patient during the decline and try to be fully invested in stock to take advantage of the recovery which, when it comes, is often quite steep.

Risk is usually defined as the standard deviation of monthly returns around a mean of the latest 36 months.  While this number may have some use in comparing stocks or mutual funds, the real risk is in the emotions of the investor and how well he or she understands the nature of the stock market.  It must be understood that in the long run (decades) the market is rational and influenced by basic economics, but in the short run (months and even years) it can be quite irrational and be influenced by mass psychology.

Getting through "bad" periods can be eased somewhat by first understanding that they are a natural result of human behavior, have happened before, and will happen again, and secondly, preparing for them by determining one's withdrawal needs in advance.  Withdrawal needs vary from routine living expenses for an individual or family, to annual pay-outs for endowed non-profit organizations, to the schedule of distribution of major foundations to the causes they support.

Whatever the case, the important point is to know what these needs are so that when stocks in general appear to be getting overpriced, enough of the portfolio can be shifted from stocks to bonds or cash equivalents to cover these withdrawals.  In this way one is assured of not having to sell devalued stocks to meet these needs, and as a by-product perhaps increasing one's tolerance for risk.

The only remaining factor is how many years worth of withdrawal needs should be protected with bonds and cash equivalents.  This can never be estimated with any precision but a helpful tool is the level of over valuation of stocks, as measured by the ratio of the earning yield on stocks (S&P 500) to the yield on short term money (money markets or 90 day treasury bills).  The lower the ratio the higher the level of overvaluation of stocks.  A ratio significantly under 1.00 signals the time to protect a year's worth of withdrawals.  A second year's worth might be protected at 0.75 or lower.  A few, but not many, declines have taken more than 2 years to recover to the pre-decline level, so a third year of protection may be called for but very rarely.

The need to estimate duration with any precision goes down with the ratio of withdrawal needs to the size of the total portfolio.  If the withdrawal rate is under 5 percent the investor can add another year of protection without seriously hindering the portfolio's benefit from the stock recovery when it eventually starts.

One might logically ask why, when it is judged the market has become overpriced, not protect the entire portfolio from the decline by going 100 percent to bonds and cash equivalents, and then switch back to 100 percent stocks when it becomes fairly or underpriced.  This is called "market timing" and seems like an ideal strategy.  If the timing of the switches is right it is a great way to increase the return of the total portfolio.

There is no shortage of people willing to advise you when the market is over or under priced.  They appear in Barron's and other financial media regularly and seldom agree, but occasionally one will predict a decline or recovery with uncanny accuracy.  They will be hailed as heroes but there is little or no follow-up to determine how they perform on the next change in market direction.  The fact that stock market newsletters that employ market timing underperform those that don't leads one to believe that these acclaimed forecasting heroes are just lucky, and no more a hero that a lottery winner (see ratings of the Hulbert Financial Digest).

The fact is that the market is so influenced by psychology in addition to basic economics that to call a change in market direction with any accuracy is pure luck, and to expect that luck to hold out on both getting out and getting back in is foolhardy.  As one student of the market observed:  "It will do whatever it has to do to confound the maximum number of people.".  Today's hero who accurately predicts a decline but completely misses the following recovery goes scot free, while the media focuses on the new hero who did accurately predict the recovery.

The high odds against the success of market timing make it impractical to apply to the whole portfolio, but perhaps only the lesser of two evils to that part of it needed to cover estimated withdrawals.  Some investors who know their risk tolerance well are so skeptical of market timing that they opt to forego any bond or cash equivalent protection, and remain 100 percent invested in stocks throughout.

This does not mean that they take no defensive measures.  Many companies sell products or services that people continue to buy even during bad economic times.  They provide consumer staples, electric utilities, and health care services, and include such companies as Proctor and Gamble, Duke Energy, Exelon, Johnson and Johnson, and many drug companies.  Stocks in such companies are not immune to losing value during times when fear pervades the entire market, but their loss is far less than for companies in the more aggressive categories such as consumer discretionary, materials, and industrials.  These include such companies as General Electric, 3M, Monsanto, and Walt Disney.

During the decline and partial recovery over the 24 months from 11/07 to 11/09, these MSC I categories performed as shown in Chart A.

CHART A

Note that during the UP period the aggressive sectors outperformed the defensive ones by 121%  (68.6 / 31.1 = 2.21), while during the down period they under performed by only 71.7% (57.9 / 33.7 = 1.718).  Even so, for the total of the two periods the defensive sectors won by 15.9% (29.0 - 13.1 = 15.9).  This is supported by the old example of starting with $100 and losing 50% and having to gain 100 percent to get back to $100.

How nice it would be if one knew about when to switch between the defensive and aggressive categories to minimize losses during down markets and maximize gains during up periods.

How to accomplish this using these defensive and aggressive MSCI sector indexes is described in what follows.  Again, as repeatedly stated above, there is no attempt to time when to be in or out of stocks, but only to use the relative price movements of these sectors to time when to make shifts between them within a portfolio the remains 100 percent in stocks.

The key to accomplishing this is to create a good measure of the relative strength of each of the MSCI sectors.  The term "relative strength" refers to the total return of the sector compared to the total return of the broad market.  Popular broad market indexes are the S&P 500 and the Wilshire 5000.  Both are weighted by the capitalization (stock price times the number of shares outstanding) of the companies that make up the index.  Whether there are 500 or 5000 companies in the index doesn't make much difference, as the capitalizations of the top 50 are so huge that their returns dominate the total.  These two are large company indexes and are not representative of the choices available to the individual investor (i.e.: both large and small company stocks).

As a result we measure the broad market by the un-weighted average of the MSCI large and small company growth and value indexes.  The arithmetic averages of these 4 styles of stocks is judged to be a fair and accurate measure of the menu of stocks available to the individual investor.

The first step to calculate relative strength is to determine the ratio of the total return index of the sector in question to the total return index of the 4 style average.  This tends to have some meaningless ups and downs and so is smoothed by a 3 month moving average.  The slope of a plot of this moving average is the relative strength of the sector in question.  Figures 1 through 6 show this smoothed line (at bottom) for our six sectors.

Now we need to decide over how long a period the slope should be measured to get a reliable estimate of relative strength.  The longer the period, the more accurate the result, but also the more out of date it tends to be.  The 3 month smoothed line is still a bit bumpy, but the slope of it over 6 months is judged to be a fair comparison.

To increase the sensitivity of this measure, we calculate the percent change in the 3 month moving average from its 6 month low and from its 6 month high and add the two together.  It is the sum of these two percentages that is our final measure of relative strength.  Table 1 Shows the calculation of "Sums" for Consumer Staples and Healthcare.

Having calculated these "Sums" for all six of our sectors, we are now ready to use them to decide which sector to be invested in at any one time, first determining the category (defensive or aggressive) and then the best sector within the selected category.

Determining the category involves comparing the average "Sums" of the three sectors in each category as in Table 2.  Figure 7 (top line) tracks the difference between the aggressive average and the defensive average.  The lower line is the ratio of the average total returns of the 3 aggressive sectors to the average of the 3 defensive sectors.  This line is more irregular than the top line, but the shapes of the two lines are quite similar meaning that the top, smoother line has value in signaling changes in the slope of the lower line.

If one defines a switch signal as a change of 4.0 or more from a high or low, the switch signals at the right of table 2 result.  Table 3 summarizes them and shows that for the 6 years for which data is available, switching between the categories would have yielded a return of 77.9%.  This compares to a return of 24.2% if all six sectors had been held simultaneously.  The advantage of switching categories then is 53.7% (77.9 - 24.2 = 53.7), or about 9% per year.

How may one implement this strategy?  Must one buy all three sectors within the category at each switch?  If so this could become a burden given that switches average one every 7 months.  Might one further improve the advantage and reduce brokerage expenses by buying just one of the three sectors?

Tables 4 (Aggressive Sectors) and 4A (Defensive Sectors) show the total returns of each sector between switches (group of columns on the right).  The left and center columns show the returns between highs and lows in the "Sums" difference (Table 2) during the periods preceding the switch signals.

Bear in mind that the period preceding a switch signal was one in which the category other than the one just signaled was favored.  So the returns of sectors in the just selected category may be inversely related to their returns during the preceding period.  Figure 8 graphs the return differences calculated in Tables 4 and 4A and shows the negative correlation albeit a rather small one.  But it does say that on average, buying the single worst performer during the preceding period will adequately represent the whole 3 sector category, and on average will slightly improve the return compared to holding all three sectors.

Following this selection process results in table 5, which tracks the switches between individual sectors during the same category favored periods as in table 3.  Now the advantage increases from 9% per year to 11% per year, which is perhaps less important than being able to achieve it by investing in only 1 instead of 3 sectors.  Note also that the advantage is only slightly less during the years leading up to the large decline and partial recovery of the past 2 years.  This is reassuring in that the success of the system doesn't depend on the unusual conditions that have prevailed over these 2 years.

Despite its advantages, holding one aggressive or defensive sector at a time failed to avoid a 12% loss during the 2+ year decline and partial recovery from November 07 to January 10.  However, during the same period the six sectors held simultaneously lost 20% (see table 6).  Over the same period the 4 style average also lost 20% while the S&P 500 lost 24%.  These losses are probably typical of those of the average individual investor, making a loss of only 12% over the 2+ years quite an improvement.

Limiting the loss to only 12% probably does not convince the very low risk tolerant investor to refrain from shifting some assets from stock to bonds or cash equivalents for a "rainy day".  Those with a higher risk tolerance and a realization that "rainy day" forecasts are often premature or in error may opt to remain 100% invested in stocks using the single sector plan outlined above.

Ray Schumacher

April 25, 2010