Benefits of Using the Historical Return Data for Predicting Stock Returns

Benefit #1 of using historical return data for predicting stock returns — it permits you to know whether stocks are worth buying at the price they are being offered or not.

Historical Return Data

When purchased at a P/E10 valuation of 14 (a moderate valuation level), the most likely 10-year return on an investment in S&P stocks is 6.3 percent. When purchased at a P/E10 valuation of 27 (the valuation level that applied in March 2007, when this article was posted), the most likely 10-year return on an investment in S&P stocks is 1.1 percent. Those are greatly differing value propositions.

For most investors, there are some valuation levels at which it makes sense to buy more stocks and some valuation levels at which it does not make sense to buy more stocks. Predicting stock returns lets you know whether buying more stocks at the price level being offered to you at a particular time is a good idea or not.

Benefit #2 of using historical return data for predicting stock returns — it confirms the conventional wisdom of recent decades that stocks offer an outstanding value proposition when held for the long run.

Let’s make another effort at predicting stock returns for purchases made at a moderate stock valuation level (a P/E10 of 14) and at a high stock valuation level (a P/E10 of 27), this time looking at the most likely 30-year return rather than the most likely 10-year return. Go 30 years out, and there is not nearly as great a difference between the results obtained at the two very different valuation levels. The most likely 30-year return at the moderate valuation level is 6.7 percent. The most likely 30-year return at the high valuation level is 5.3 percent.

Both of those returns are darn juicy returns, no? You can probably beat the most likely 10-year return for the investment made at high valuation levels with an investment is something safe like Treasury Inflation-Protected Securities (TIPS). But it is not going to be easy to beat the most likely 30-year return of 5.4 percent.

The moral? Stocks really are an outstanding investment class for the truly long run. It’s just important that you understand how long the long run can go at times of high valuation. Holding stocks for 10 or 20 years often is not good enough at such times. Holding stocks for 30 years will in almost all circumstances offer you a likely return that is at least acceptable. Even the worst possible 30-year annualized real return (presuming that stocks perform in the future as they have in the past) for stocks purchased at a P/E10 of 27 is 3.3 percent.

Benefit #3 of using historical return data for predicting stock returns — it reveals the benefits that can be obtained by holding off on stock purchases when stocks are at high valuation levels.

Historical Stock Return Data

S&P stocks reached their highest valuation in the history of the U.S. market in January 2000, when the P/E10 level went to 44. By March 2007, the P/E10 level had fallen to 27. That’s still very high, but not quite as absurdly high. Did that drop in prices make a material difference in the return expectations of valuation-informed indexers who held off making stock purchases until recently? It sure did.

For purchases made at a P/E10 of 44, the most likely 20-year annualized real return is 1 percent. Not good. The most likely 20-year return for purchases made at a P/E10 of 27 is 2.6 percent. That’s nothing to write home about, but it’s a big improvement on the earlier number all the same. Those who were telling you that you would surely “miss out” if you did not immediately buy stocks when they were being sold at the absurdly high prices that applied at the top of the recent price super-bubble were offering guidance that leaves something to be desired. The sort of return provided by stocks purchased at a P/E10 level of 44 is something that most of us can afford to “miss out” on.

Benefit #4 of using historical return data for predicting stock returns — it reveals what factors influence stock returns at different sorts of holding periods.

At time-periods of less than 10 years, stock prices are unpredictable. There are so many factors affecting the result that no one factor—even the all-important stock valuation factor–controls.

Meaningful predictions are possible for holding periods of 10 years. But look at how much shorter the Color Bar of The Stock-Return Predictor is at 20-years than it is at 10 years. That shortening of the Color Bar represents an increase in the precision of the prediction being made. Stocks are only somewhat predictable at 10 years. The 20-year predictions possess more clarity.

At 30 years, the Color Bar grows shorter still. There’s a good degree of precision to stock predictions made for holding periods of 30 years. When a P/E10 of 14 applies, the range of predictions at 10 years goes from 0.34 percent to 12.34 percent. That range of 12 percentage points covers a lot of territory. At 30 years, the range is reduced to 4 percentage points (from 4.73 percent to 8.73 percent). A prediction extending over a range of only 4 percentage points is providing highly meaningful information.

The precision possible in predicting stock returns does not get much better, even when the holding period extends to 60 years. At 60 years, the range is down only another 1.5 percentage points, with possible returns ranging from 5.3 percent to 7.8 percent.

Benefit #5 of using historical return data for predicting stock returns — it identifies for you the time-periods in which the starting-point stock valuation level matters most.

The starting-point stock valuation level makes a big difference at 10 years. It is my view, though, that the predictions that can be made at 10 years are sufficiently imprecise that it is for 20-year holding periods that valuations matter most.

The range of possible outcomes for 10-year holding periods are sufficiently broad that it remains possible to obtain a good return for a stock purchase made at a high valuation level. Look again at the results obtained from The Stock-Return Predictor when the P/E10 value entered is 44, the level that applied in January 2000. The most-likely return is a negative number, as noted above. But the “Lucky” return (a return that has about a 1 in five chance of turning up) is 1.9. That’s not good. But it’s not so terrible for a 10-year time-period.

How Stocks Work

Now look at the “Lucky” return for a 20-year holding period. The number is higher — 3.0 percent. But is that a better result than the 1.9 percent “Lucky” return that applied at 10 years? In my eyes, it is not. Stocks are a volatile asset class, so volatile that the long-term returns provided by them are generally much higher than either of those numbers. I think that a good case can be made that it is worse to have to hold stocks 20 years to obtain a return of 3.0 percent than it is to hold them for 10 years to obtain a return of 1.9 percent. The 10-year return is better on its face. But the 20-year return is not enough of an improvement to justify holding a highly volatile asset class for an additional 10 years.

There are two competing forces coming into play in the move from a 10-year holding period to a 20-year holding period. The likelihood that the starting-point valuation will have had time to have had an effect is becoming stronger, asserting a negative effect on the 20-year number. And the likelihood that the effect of the starting-point valuation will have begun to diminish is also becoming stronger, asserting a positive effect on the 20-year number.

The timing of the price drop cannot be ascertained in advance. So we cannot know how far on the way to recovery we will be at 20 years. But the calculator tells us the most-likely scenario. It allows us to compare the most-likely scenario at Year 10 with the most-likely scenario at Year 20 and thereby to draw a conclusion as to which time-period is of greater significance.

My view is that the 20-year number is more troubling than the 10-year number. I don’t like the idea of earning only a 1.9 percent annualized real return from an investment for 10 years. I like even less the idea of earning only a 3.0 annualized real return for 20 years. Why? Because stocks purchased at fair prices can be realistically expected to earn a 10-year return of over 6 percent. There’s generally not much appeal in having your money tied up in a high-risk asset class for 20 years and obtaining a return of only 3 percent real for your trouble.

Benefit #6 of using historical return data for predicting stock returns — it permits you to develop informed assessments of alternate asset classes.

Treasury Inflation-Protected Securities (TIPS) were paying a real return of 4.0 percent at the top of the bubble. How attractive is that return? Compare it to the long-term return offered by stocks at times of normal valuations (which is what most investors do), and it does not appear that attractive. Stocks are of course riskier. But stock investors are compensated well for taking on the extra risk at times of moderate valuations. At a P/E10 of 14, the most likely 10-year return is 6.3. That’s more than 2 points of real return greater than what could be obtained from TIPS at the top of the bubble. My guess as to why many investors did not snap up TIPS at the time is that they were thinking that it was realistic to expect returns of 6.3 real even at those times of super-bubble-level prices.

What We Know About Stock Investing
The Stock-Return Predictor tells us that it is not so. The reality is that stocks purchased in January 2000 were not even likely to provide a positive return in 10 years. TIPS were by far the better choice. TIPS offered a far higher return at far less risk.

Benefit #7 of using historical return data for predicting stock returns — it disproves the Efficient Market Theory.

The Efficient Market Theory is the most dangerous of today’s investing myths. According to this “theory” (it is more properly termed an “assumption,” one very much at odds with both common sense and the historical record), stocks always must offer a better long-term return than a risk-free asset class like TIPS. Why? Because the theory assumes that investors are completely rational in their decisions as to what to do with their investing dollars. Rational investors would not be willing to take on the risks associated with stock investing without being compensated for it. So the theory assumes that there must always be a risk premium, that stocks must always offer a better return than risk-free investment choices.

Why, then, does a regression analysis of the historical data tell us that stocks were likely to provide a negative 10-year return at a time when risk-free TIPS were offering a return of 4 percent real? Because the root assumption of the Efficient Market Theory is nonsense. Investors do of course employ reason when selecting investments. Emotion is a more important influence, however. At times of high prices, most investors cannot stand the idea of passing up stocks because they have been providing good returns in the recent past. It is generally only when stock prices are low (and the value proposition for stocks is high) that investors are willing to give serious consideration to “boring” asset classes like TIPS. I doubt very much that TIPS will ever be providing returns of 4 percent real at times when stocks are low-priced (and, thus, unappealing to most investors).

The risk “premium” remains negative to this day. A better 10-year return is available today from TIPS than is available from stocks, according to the numbers generated by The Stock-Return Predictor.

Benefit #8 of using historical return data for predicting stock returns — it teaches you that small differences in valuations don’t matter much.

The difference in 10-year returns when we go from a P/E10 of 14 to a P/E10 of 16 is 1.3 percent. That’s not insignificant. But the precision of the 10-year numbers is not high. Given the lack of precision, the 1.3 percent difference really does come close to being insignificant, in my assessment. The calculator makes a case that it does not make sense to change stock allocations except when valuations change significantly.

Benefit #9 of using historical return data for predicting stock returns — it permits you to put temporary upward and downward price movements into context.

Past Performance Is No Guarantee of Future Results

Stocks prices were coming down in 2002 and 2003. Since then, they have headed upward. Many investors have thus come to believe that it was a mistake to lower one’s stock allocation in 2002 or 2003.

The Stock-Return Predictor argues otherwise. The lowest P/E10 value we saw during the downturn was a P/E10 of 21. That’s not low enough to justify a belief that stocks offer a strong long-term value proposition. Yes, we’ve gone upward since then — for a time. To the long-term investor, however, the value proposition of stocks has remained poor for 10 years now.

Using The Stock Return Predictor for predicting stock returns permits you to distinguish price drops that are worth taking advantage of from price drops that are traps for long-term stock investors. Just because the price of stocks has dropped significantly since 2000 does not mean that stocks are now a good buy. The price had gone so high in 2000 that it would take a huge price drop even from today’s levels to get to prices where the long-term value proposition would be strong.

Benefit #10 of using historical return data for predicting stock returns — it provides you the confidence needed to become a true long-term buy-and-hold investor.

There’s a song that argues that: “Everyone’s crying mercy, but they don’t know the meaning of the word.” I think that applies to stock investing, except that in that context the way that it needs to be said is that: “Everyone’s crying ‘buy-and-hold’ but they don’t know the meaning of the phrase.”

Buy-and-hold is good stuff. Buy-and-hold works. You want to practice buy-and-hold investing.

It’s hard, though. Real hard. Few are willing to tell you how hard.

Do you know what separates those who merely talk the buy-and-hold talk from those who truly walk the buy-and-hold walk? It’s confidence. To stick with your investing strategies when they are under fire, you need to believe in them in a deep way.

Dow Jones Industrial Average

Confidence follows from preparation. There are going to be times when your investing strategies are not going to be working out so well. Those times are going to be the test of whether you successfully pull off buy-and-hold or not. If you have properly used the The Stock-Return Predictor to inform your asset-allocation decisions, you will be prepared for what comes down the line, presuming that what comes down the line is no worse than what we have seen come down the line in the past (and the price crashes of the past have been bad enough that it is none too likely that we are going to see anything much worse).

Use the historical stock-return data to inform your investing strategies, and you probably will be fine in the long run. Allow yourself to get caught up in the euphoria of wild bull markets, and you stand a darn good chance of getting burned and needing to sell at the worst possible time for doing so.

The Stock-Return Predictor generates results that are hard medicine for some to swallow. That’s just what it is supposed to do. Those of us who are serious about wanting to succeed as long-term investors need to swallow some hard medicine from time to time. Today (we are now at price levels rarely before seen in the historical record) is very much such a time.