Current Research B

Keeping In Tune with the Human Element

Updated August 21, 2005.

I am opening up this new Current Research section in response to these letters from Rob Bennett.

John:

Here is another question that I have wondered about from time to time.

As you know, I have never been one to go too far in using numbers as a guide to investing decisions. For me, emotions trump numbers. The primary value that I see in using historical data is as a means of better informing one's emotions, so that one is using emotions in an effective way rather than in a self-destructive way. To be sure, that is no small thing. It is because SWR analysis helps investors inform their emotions that I find SWR analysis so exciting a breakthrough in the field of investment analysis.

What this means in practical terms is that I would generally not advise investors to get out of stocks at today's valuations. Valuations are high, and a statistical argument can be made that it is better to stay on the sidelines for a time and then return to stocks when they offer a stronger long-term value proposition. The flaw in this argument, in my view, is that it ignores the emotional effect that large short-term stock gains may have on the investor following a get-out-of-stocks-for-now strategy. Short-term price changes are unpredictable. Stocks could go up dramatically over the course of the next year, or the next three years, or the next five years. If they do, the investor may lose confidence in his number-backed investment strategy and buy into stocks at the worst possible time.

My view is that it is best to maintain a moderate position in stocks at times of high valuation and that it is also best not to go too extreme on the high side in one's stock allocation at times of low valuation (because in the short-term stocks may drop sharply even from a starting point at which valuations are low). I view the TIPS position as being a grounding device. I believe it makes sense to make shifts in one's stock position to take advantage of long-term timing opportunities. But I don't believe in making dramatic shifts because to do so sacrifices the benefits of effective grounding.

It would be nice if there were a way to use the historical stock-return data to demonstrate the benefits (presuming there really are any) of this stay-grounded-at-all times strategy. My thought is that the way to do it would be to go back to earlier points in history, determine the SWRs that applied for various portfolio allocation strategies at those times, and then look forward to see how the strategies fared in the years that followed .Then you would need some means of assessing the length and intensity of regret that would be suffered by investors employing various sorts of portfolio allocations. By speculating as to how the regret experienced by the various types of investors would influence their decisions to alter their portfolios (or not alter them), you could show whether it was a good idea to rely primarily on statistics to form one's strategies or to balance what the statistics say with a concern for what might happen in a new sort of worst-case scenario (one that causes maximum regret over one's earlier investment choices).

If you have any questions or thoughts you would like to run by me re this, please do.

I sent this as a regular e-mail because I was not sure that it was suitable for use in your Letters to the Editor section (as the thinking here is so unformed). However, if you would like to use it as a Letter to the Editor to document the exchange-of-ideas process that generates many of our findings, I am fine with that.

Rob

I made some comments and Rob replied.

John:

Your last comment is the focus of my concern. [Rob is referring to an earlier email.]

If history repeats, we will see in the not-too-distant future a time when "everybody will be too scared to invest anything in stocks." That's an overstatement, of course, but it is a reasonable approximation of what has happened in the past in the years following the sort of overvaluation that applies today. We should be planning today for the conditions that will apply then, just as we (I mean the entire community, not just you and me) should have been planning in January 2000 for the conditions that apply today. It is because we (the community) are too influenced by current-day concerns that so many missed out on the opportunity to invest in 4.1 percent TIPS that was available in an earlier day.

My goal is to use the historical data to develop an approach to investing that avoids the negatives at both extremes of valuation: (1) being too heavy in stocks at times of overvaluation; and (2) being too light in stocks at times of undervaluation. I see the two goals as being complementary ones. Those who do not suffer pangs of regret from being too heavy in stocks at times like today will feel more comfortable investing a good bit in stocks when values hit low points. Thus, they obtain a double hit of good stuff--they avoid portfolio-destroying losses AND they tap into the amazing portfolio building returns generally available only in the wake of times of overvaluation. [I believe that Rob means undervaluation.--JWR]

The key to realizing all these benefits in the real world is using an approach that is BOTH data-based AND in tune with investor emotions as they exist in the real world. I don't agree with Pete [Pete posts under the username peteyperson] that it is "not smart" to go with an allocation other than the one supported by the numbers. I understand his point, but I believe that he is making a mirror-image mistake of the mistake that Intercst [Intercst founded www.retireearlyhomepage.com] supporters [i.e., supporters of the conventional methodology] make. Intercst supporters kid themselves into thinking that stocks today are a better value proposition than they really are by ignoring the long-term possibilities and Pete is (in my view) kidding himself that stocks today are a worse value proposition than they really are by ignoring the short-term possibilities. I'm seeking a middle-ground strategy.

It is my view that it is a lot harder to follow the numbers than many seem to think. I believe that the key to following the numbers IN THE REAL WORLD is making use of moderation in one's employment of the numbers. Second-best statistical choices that are in tune with investor emotions are likely in most cases to deliver better long-term returns that first-best statistical choices that are not, in my view.

The "regret" that I am concerned about is not the loss of returns that could have been had by going with first-best statistical choices. The question that I am trying to answer is--If someone lowered his stock allocation in 1996, as advised by Shiller in his congressional testimony of July 1996, what are the chances that the regret he would have experienced when stocks went up dramatically in the late 90s would have caused him to jump ship on a theoretically appealing investing approach at the worst possible time to do so? My guess is that the investor who stuck with a 30 percent stock allocation had better chances of sticking with the statistics-backed approach than the investor who got "greedy" and went with a zero percent stock allocation.

What I am suggesting is that we develop our own subjective assessments of what size "losses" would cause an investor enough regret to abandon a statistics-based investing approach, and then check the historical data to see how common it is for the conditions that would cause such losses to actually take place. My guess is that the gains from following a statistics-based approach are so great, and the losses attached to allowing regret to cause one to abandon the statistics-based approach are also so great, that a statistics-based argument can be made that it is better not to get too "greedy" in application of a statistics-based approach.

In short, I am saying that it would probably have been better for the typical investor (My particular circumstances were not typical) persuaded by Shiller's testimony in 1996 to have gone to 30 percent stocks than to 0 percent stocks. The reason is that the investor who went to 0 percent stocks took a big real-term risk of blowing all his gains by becoming so frustrated with the statistics-based approach as to abandon it before it paid off. I am saying that the short-term is so unpredictable and the intermediate-term so predictable that it is better to take half a loaf than to shoot the moon and adopt a pure statistics-based allocation strategy.

Rob

Benjamin Graham’s Constraint with Switching

As my first effort, I have examined what happens if we follow Benjamin Graham’s advice to maintain stock and bond allocations between 25% and 75%.

Professional investment advisors do well to heed such advice. If they recommend being out of the market and it does well, as during the bubble, they can expect to lose customers. If they like stocks but the market turns against them, they can point to the bond holdings as a form of insurance.

Some individuals might wish to quiet the critics with their endless recitations of short-term performance and strategies that require a bubble to be successful. Typically, it costs money. This tells you how much.

Benjamin Graham recommended constraining both stock and bond allocations to 25% to 75%.

These are the results for switching allocations with stocks and TIPS with a 2% (real) interest rate subject to Benjamin Graham’s constraints.

I used my Deluxe Calculator version 1.1A08. I set expenses to 0.20%. I selected to use the CPI for inflation adjustments.

The stock allocation is 100% whenever the value of P/E10 is below the lowest threshold. It is 0% whenever P/E10 is higher than the highest threshold. Both of these allocations are fixed. To work around this limitation, I set the lowest selectable threshold equal to 2. I set the highest adjustable threshold equal to 80. This leaves us with two intermediate thresholds and three allocations.

Taking a hint from previous optimizations, I set the lowest allocation equal to the maximum stock holding allowed (75%). I set the highest allocation equal to the smallest stock holding allowed (25%).

I incremented the withdrawal rate in increments of 0.1%. I recorded the lowest rate at which there was at least one failure. I recorded the lowest rate at which there were 6 or more failures. I recorded the lowest rate at which there were 12 or more failures.

I surveyed to find the highest withdrawal rates at which there were 1 or more, 6 or more and 12 or more initial failures. That is, I maximized what produces a minimum.

Initial Optimization

I optimized the two intermediate P/E10 thresholds and the single intermediate stock allocation.

The previous allocations and thresholds had been 100%-50%-30%-20%-0% and 9-12 (or 13)-21-24. I set the thresholds to 2-12-21-80 and the adjustable allocations to 75%-varies-25%.

The first failure occurred at a withdrawal rate of 4.9% when the intermediate allocations were 20% and 30%. The first failure occurred at a withdrawal rate of 4.8% when the intermediate allocation was 40%. At all other allocations, the first failure occurred at lower withdrawal rates.

With a 40% allocation, the withdrawal rate with 6 (or more) failures was 5.1%. It was 5.0% with allocations of 30% and 50%. At all other allocations, it was at lower withdrawal rates.

With a 40% allocation, the withdrawal rate with 12 (or more) failures was 5.4%. It was 5.3% with allocations of 30%. At all other allocations, it was at lower withdrawal rates.

I selected allocations of 30% and 40% for further optimizations.

P/E10 lower threshold

I set the allocations to 100%-75%-30%-25%-0% and the P/E10 thresholds to 2-varies-21-80. This is equivalent to allocations of 75%-30%-25% with a lower threshold that varies and a higher threshold of 21.

The selection of the P/E10 threshold was easy. There were no conditions that did better than a P/E10 threshold of 11. There were many conditions that did as well, but not as often.

With a 30% stock allocation, the first failure occurred at a 4.9% withdrawal rate for all thresholds from 9 through 13. Six failures occurred at a withdrawal rate of 5.1% for P/E10 thresholds of 9 and 11. They occurred at a withdrawal rate of 5.0% for P/E10 thresholds of 10, 12 and 13. Twelve failures occurred at a withdrawal rate of 5.3% with P/E10 thresholds of 11, 12 and 13. They occurred at a withdrawal rate of 5.2% with P/E10 thresholds of 9, 10 and 18. All other conditions had lower withdrawal rates than these.

With a 40% stock allocation, the first failure occurred at a 4.8% withdrawal rate for all thresholds from 9 through 13. Six failures occurred at a withdrawal rate of 5.1% for all P/E10 thresholds from 9 through 13. Twelve failures occurred at a withdrawal rate of 5.4% with P/E10 thresholds of 11, 12 and 13. They occurred at a withdrawal rate of 5.2% with P/E10 thresholds of 9 and 10. All other conditions had lower withdrawal rates than these.

P/E10 upper threshold

I set the allocations to 100%-75%-30%-25%-0% and the P/E10 thresholds to 2-11-varies-80. This is equivalent to allocations of 75%-30%-25% with an upper threshold that varies and a lower threshold of 11.

The selection of the upper P/E10 threshold was not critical. I selected thresholds of 20 and 21 (without preference). With a 30% stock allocation, all conditions with all P/E10 thresholds from 20 through 26 produced the best results. With a stock allocation of 40%, thresholds of 20 and 21 produced the best results in terms of six failures and twelve failures. The withdrawal rate at the first failure was 4.8% as opposed to the best rate of 4.9%.

With a 30% stock allocation, the first failure occurred at a 4.9% withdrawal rate for all thresholds from 11 through 26. Six failures occurred at a withdrawal rate of 5.1% for P/E10 thresholds of 20 and 26. They occurred at a withdrawal rate of 5.0% for P/E10 thresholds of 11 through 19. Twelve failures occurred at a withdrawal rate of 5.3% with P/E10 thresholds of 14 through 16 and 18 through 26. They occurred at a withdrawal rate of 5.2% with P/E10 thresholds of 11 through 13 and 17. All conditions with a P/E10 threshold of 20 through 26 had the highest withdrawal rates at one, six and twelve failures.

With a 40% stock allocation, the first failure occurred at a 4.9% withdrawal rate for thresholds from 11 through 13 and 18. The first failure occurred at 4.8% for withdrawal rates of 14 through 17 and 19 through 24. Six failures occurred at a withdrawal rate of 5.1% for all P/E10 thresholds of 20 and 21. They occurred at a withdrawal rate of 5.0% for P/E10 thresholds of 11 through 13, 18 and 19 and 22 through 26. Twelve failures occurred at a withdrawal rate of 5.4% with P/E10 thresholds of 20 through 26. They occurred at a withdrawal rate of 5.3% with P/E10 thresholds of 17 through 19. All other conditions had lower withdrawal rates than these.

The Final Selection

My final selections are based on intermediate stock allocations of 30% and 40%, a lower P/E10 threshold of 11 and upper P/E10 thresholds of 20 and 21.

Here are the results:

With stock allocations of 75%-30%-25% and P/E10 thresholds of 11 and 20:
1) the first failure occurs at a withdrawal rate of 4.9%.
2) the sixth failure occurs at a withdrawal rate of 5.1%.
3) the twelfth failure occurs at a withdrawal rate of 5.3%.

With stock allocations of 75%-30%-25% and P/E10 thresholds of 11 and 21:
1) the first failure occurs at a withdrawal rate of 4.9%.
2) the sixth failure occurs at a withdrawal rate of 5.1%.
3) the twelfth failure occurs at a withdrawal rate of 5.3%.

With stock allocations of 75%-40%-25% and P/E10 thresholds of 11 and 20:
1) the first failure occurs at a withdrawal rate of 4.8%.
2) the sixth failure occurs at a withdrawal rate of 5.1%.
3) the twelfth failure occurs at a withdrawal rate of 5.4%.

With stock allocations of 75%-40%-25% and P/E10 thresholds of 11 and 21:
1) the first failure occurs at a withdrawal rate of 4.8%.
2) the sixth failure occurs at a withdrawal rate of 5.1%.
3) the twelfth failure occurs at a withdrawal rate of 5.4%.

These results are similar. Differences are minor. We do not have to worry about any dangerous sensitivities.

I selected stock allocations of 75%-40%-25% to make the intermediate allocation stand out. An allocation of 30% would have been very close to the 25% constraint.

I selected P/E10 thresholds of 11 and 21 as a concession to those with a strong preference for holding stocks. In addition, I have much more data using thresholds of 11 and 21 than with thresholds of 11 and 20.

This is my final selection:

1) The stock allocations are 75%-40%-25%. (The calculator sees this as 100%-75%-40%-25%-0%.)
2) The P/E10 thresholds are 11 and 21. (I put 2-11-21-80 into the calculator so that the stock allocations of 100% and 0% never get selected.)

I refer to this portfolio as SwAT2. It uses Switching. It is subject to constraint A, which I define as keeping both stock and bond allocations between 25% and 75%. The portfolio includes TIPS with an interest rate of 2.0%.

Data Collection and Processing

I collected a complete set of 30-year Historical Surviving Withdrawal Rates from 1871-1980.

I used Excel to fit a straight line to the 1923-1980 Historical Surviving Withdrawal Rates as a function of the percentage earnings yield 100E10/P.

The equation for the Calculated Rate is:
y = 0.3776x+3.5222
R-squared is 0.6466.

My eyeball estimates of the confidence limits are minus 0.7% and plus 1.3% (when the earnings yield is less than 8%, which means that P/E10 is greater than 12.5).

At today’s valuations, the earnings yield is 3.5% (roughly).

Here are the results for SwAT2 at today’s valuations:
The Safe Withdrawal Rate is 4.1%.
The Calculated Rate is 4.84%.
The High Risk Rate is 6.1%.

This compares with an optimal unconstrained portfolio (100%-50%-30%-20%-0% with thresholds of 9-12-21-24) of stocks and 2% TIPS, which had the following results:
The Safe Withdrawal Rate is 4.4%.
The Calculated Rate is 5.12%.
The High Risk Rate is 6.4%.

The Calculated Rate is lower 0.28%. The Safe Withdrawal Rate and High Risk Rate are lower by 0.3%.

Here are the January values of P/E10 throughout the last decade.

1995 20.22
1996 24.76
1997 28.33
1998 32.86
1999 40.58
2000 43.77
2001 36.98
2002 30.28
2003 22.89
2004 27.65

Here are the Safe, Calculated and High Risk Rates of the last decade for SwAT2:

Year, Safe Withdrawal Rate, Calculated Rate, High Risk Rate

1995…..4.7…..5.39…..6.7
1996…..4.4…..5.05…..6.4
1997…..4.2…..4.86…..6.2
1998…..4.0…..4.67…..6.0
1999…..3.8…..4.45…..5.8
2000…..3.7…..4.38…..5.7
2001…..3.8…..4.54…..5.8
2002…..4.1…..4.77…..6.1
2003…..4.5…..5.17…..6.5
2004…..4.2…..4.88…..6.2
Today..4.1…..4.84…..6.1

Conclusions

Benjamin Graham’s constraints reduce today’s 30-year Safe Withdrawal Rate from 4.4% to 4.1%. This corresponds to a 95% likelihood of success.

At a 4.1% withdrawal rate, the chance of success without switching is better than 50%-50% with a fixed allocation of 50% stocks and 50% TIPS (drawing 2% interest), but the Safe Withdrawal Rate is only 3.6%. At a 4.1% withdrawal rate, the chance of success without switching is almost exactly 50%-50% with a fixed allocation of 80% stocks and 20% TIPS (drawing 2% interest). The Safe Withdrawal Rate with 80% stocks is 3.0%.

What Does This Mean?

Rob Bennett sent this follow up letter. It brings up an important point.

John:

Thanks much for putting together the "Keeping in Tune with the Human Element" article.

You say: "Benjamin Graham’s constraints reduce today’s 30-year Safe Withdrawal Rate from 4.4% to 4.1%." My read of the lesson here is that you "pay" three-tenths of a percentage point of SWR to obtain the insurance that an experience of emotional regret will not cause you to miss out on making the benefits of your data-based investing strategy pay off in the real world.

To my way of thinking, the added security is worth the price.

But, as always, it is for the investor making use of the research to determine that for himself or herself. The benefit of the Data-Based SWR Tool is that it translates vague thoughts and concerns and speculations into data-backed actionable insights.

I liked Greg's letter on "Worry-Free Investing" and Bernstein's comments on it, and your response to Greg's letter, as well.

And the hits just keep on coming!

This site is one amazing resource for today's middle-class investor.

Rob

HERE IS MY RESPONSE

You are right concerning the basic interpretation. You pay to follow Benjamin Graham’s advice by reducing your Safe Withdrawal Rate by 0.3%.

How about this as an alternative explanation of what you buy?

If the future is similar to what has happened in the past, your 0.3% buys you protection from emotional regret.

If the future differs significantly from the past, your 0.3% protects you from the experiencing the worst of the downside. We could see the bubble reappear and remain for a considerable length of time. Later, toward the end of this secular (i.e., long-term) bear market, we could see stock prices fall below bargain levels (with P/E10 below 9) to extreme bargain levels, possibly to record lows (with P/E10 below 5).

You may recall that I refuse to place confidence levels greater than 90% around Calculated Rates. [This results in one-sided confidence levels of 95% for the Safe Withdrawal Rate and for the High Risk Rate.] We know enough never to rely on numbers at a higher level of confidence.

One thing that each generation can expect is to be surprised. After the 1929 stock market crash and the Great Depression, many older investors stayed in bonds and away from stocks. They had learned their lesson. It was the wrong lesson. They were caught by a prolonged period of high inflation and stagflation. We have lived through a spectacular, prolonged bull market. Many investors have learned their lesson: always invest everything in stocks. It was the right lesson during the bull market. It has been the wrong lesson since the bubble burst in 2000.

So how could we be surprised today?

We could see something that doesn’t happen very often, but which does happen. Or, if it never has happened, it is something that could happen. Maybe there is another surprise out there.

Benjamin Graham’s insurance policy costs 0.3% (in terms of the Safe Withdrawal Rate). But it protects against more than emotional regret.

Owning stocks at today’s valuations may seem FOOLISH. The numbers tell us that. But you never know. Let us assume that the numbers are right. You still might be the last one to sell to a GREATER FOOL.

Have fun.

John Walter Russell
Updated June 22, 2005.

We Should Be Hedging Our Bets

Rob Bennett followed up with this letter (Wednesday, June 22, 2005).

John:

I appreciate your most recent response. The words you put forward in it, however, cause me to flip over to the PeteyPerson side of the table for a bit.

You say:

"Benjamin Graham’s insurance policy costs 0.3% (in terms of the Safe Withdrawal Rate). But it protects against more than emotional regret."

I think that it is this "something more” question that Petey was referring to when he made the suggestion that one should just go with the numbers because in the long run any short-term gains you experience as a result of allowing your emotions to trump the numbers will be "given back" if you really are a long-term investor. Most middle-class investors are investing for long-term goals, so I think that it is the long-term that matters for most of us. Most of us are not day traders or week traders or month traders or year traders. So I think Petey's point is a very important one.

I think Petey is right. The numbers don't tell us WHEN we will give back any cotton-candy gains we happen to enjoy as the result of valuations getting ahead of themselves. But they do tell us that we will give those gains back.

It could be that the long-term will end up being so far out that it will not matter in practical terms. For most of us who experience long life spans (and we need to plan for that in any event), however, the long run is going to catch up to us sooner or later. So, again, I see Petey’s point as being a compelling one.

My reluctance to go along with Petey altogether is due to a concern that emotions are as important as numbers.

I just have never bought into the idea that you can study stocks in a laboratory and get results that are truly meaningful. In the laboratory setting, one of the major factors at play is absent from the analysis. The missing factor is the nature of the entity who OWNS the stocks, who makes the buy and sell decisions. Ignore that factor and you have ignored so much that you discredit the remainder of the analysis in important ways, in my view. I would go so far as to say that to ignore the human element is ultimately as grave an error as to ignore the effects of valuation. It's just too significant a factor to overlook (and what I know of the historical data indicates to me that the historical data backs me up on this point).

In my mind this always circles back to my initial question--Is there a way to assess statistically the odds of experiencing emotional regret over a decision to follow a data-informed investment strategy? Do stocks generally go up and down in statistically detectable patterns, so that they can be planned for by those employing a data-based strategy? Or is there so much variability to the patterns that the typical middle-class investor cannot reasonably expect to steel himself against the dangers of the feelings of regret and anxiety he is likely to feel if he goes with a pure numbers-based approach?

I think the data lends more support to the latter description of the realities. But I am not entirely sure.

Are there patterns to how long we stay in bull markets and how long we stay in bear markets? If yes, then it is possible for the data-informed investor to prepare himself for what is to come when he ventures down the data-informed investing path. The book "Stocks Cycles" argues that there ARE patterns. And my recollection is that in your research you have found at least a little bit of evidence telling the same story.

But the length of time in which stock prices have remained high following our most recent Bull Market (assuming that it has indeed come to an end and we are not going to see another march upward in the near future) suggests that these patterns are highly unpredictable.

I often think back to the Shiller testimony of July 1996 saying that those heavily in stocks would live to regret it within 10 years. Shiller still has one year remaining for that prediction to come through for him. Will it? Perhaps. My guess is that Shiller is not nearly so confident today that his prediction will come through as he was in July 1996. There's no one in this field smarter and more experienced than Shiller, and Shiller is cautious in the predictions he puts forward. If Shiller can't get the predictions that follow from taking an informed look at the historical data right, who can?

The key reality re this matter is the distinction between the unpredictability of the short term and the predictability of the long term. The strategic question is--Should one be looking at the short-term behavior of stocks or the long-term behavior of stocks in putting together an investment strategy? I think that one needs to look at both.

The long-term behavior is the most important because it is long-term investment results that matter most. But it is a mistake to ignore the short-term altogether because it is the short term that is of primary significance in affecting human emotions, and human emotions matter big time in any complete assessment of investment strategies. If your expectations re the short term fail you, you may end up changing your allocation before the long term kicks in and you realize the benefits of your long-term strategy in the real world.

I think that the next step is to gain a sense of how great the likelihood is that data-informed investors will be fooled by the appearance of new returns patterns (or by the reappearance of old returns patterns that they did not anticipate us seeing again). Is the length of time in which the current high valuations have prevailed really unprecedented? If so, how often do unprecedented patterns pop up? If the answer is "often enough that we need to take the possibility into account," how should data-informed investors take this reality into account in development of their strategies?

My working theory is that we should be hedging our bets, not going too far in the direction in which data not accounting for the human element points us. But I would feel more assured that this theory is a good one if I knew that the historical data really does show what I sense it shows--that returns patterns are sufficiently unpredictable in the short term as to cause significant amounts of regret for most investors following data-informed strategies because of their apparent long-term appeal.

Rob

HERE IS MY RESPONSE

In terms of the S&P500, P/E10 exceeded 20 during the 1920s for the first time in 1928. The run up (beyond a P/E10 of 20) was very short. P/E10 exceeded 20 for most of the decade of the 1960s. Real price increases stalled in 1966 and then started falling as real earnings increased. P/E10 broke through 20 in 1993 and it has stayed above 20 ever since.

There is no precedent for the length of time that stocks have been at current valuations. In this sense, it really is different. At least, in the United States.

In terms of our Standard Switching Portfolio, we would have abandoned stocks entirely in late 1995 and stayed out except for a brief period from mid-2002 to mid-2003. [P/E10 exceeded 24.]

In terms of SwAT2, we would have kept our stock allocation down to 25% from the fall of 1993 through today. [P/E10 exceeded 21.]

Knowing what we now know, there would not have been a need for regret. TIPS came on the market at an interest rate of 4.0% and higher. Our calculations tell us that, if you could get replacements at year 30, you could withdraw 4.46% of your initial balance (plus inflation) for 50 years or 5.05% of your initial balance (plus inflation) for 40 years.

Those are excellent numbers, especially when we consider the effect that valuations have on Safe Withdrawal Rates. In terms of what is reasonable to expect, it was the winning strategy.

Today, a TIPS-only holding still makes sense even though the interest rate has dropped below 2.0%.

What our calculations have not told us is what will happen to portfolios with high stock allocations starting from 1993-1995. [The SWR Translator helps us address this type of question.] In terms of real prices, the S&P500 index was around 550 compared to today’s level of 1200+. This is an annualized real rate of return of 8% before reinvesting dividends. Clearly, this is an outstanding payoff for those following a high-risk strategy.

Regarding the human element: those who followed the high-risk strategy thought (and most still think) that they followed a low-risk strategy.

Regarding these questions: If so, how often do unprecedented patterns pop up? If the answer is "often enough that we need to take the possibility into account," how should data-informed investors take this reality into account in development of their strategies?

I am convinced that each new generation sees something unprecedented enough to catch it by surprise.

Data-informed investors should think through what would happen if they were wrong.

Benjamin Graham’s approach was excellent for investment professionals. They would have a chance to keep their customers when their advice was wrong, regardless of the reason.

Ken Fisher plans two portfolios based upon what he expects the market to do. He divides the possible outcomes into three parts. He excludes investments based on the consensus forecast. He assumes that the consensus is ALWAYS wrong. This leaves him with two possibilities. He puts most of his money in the outcome that he expects to happen. He puts a little bit of money on the remaining outcome as insurance.

Ken Fisher’s approach aims at keeping customers. He can exclude the consensus regardless of what actually happens because he markets himself as a contrarian.

Here is something else to ponder. Even a hedge can become extreme. You may end up abandoning your hedge against regret! A possible side effect is that you might hold on to some of those cotton-candy gains.

Portfolio Safety Insights

We look at how portfolios behaved in the past to gain insights into the future. We can learn a lot about portfolio safety. I have collected data that show us what to expect.

Here is the Summary from my most recent investigation:

"In the first 10-12 years and often sooner: more than two-thirds of the time, you will know if your portfolio is in danger."

"In the first 10-12 years and often sooner: more than two-thirds of the time, you will know if your portfolio is safe."

"There is a region of uncertainty. These are balances above 60% and below 90% of the starting balance."

This is the kind of thing that real people need to know.

You can read the full story at this link.
Portfolio Safety Insights

You Can't Count on 7% Articles

Here are a series of articles about what to do if you have a lump sum to invest for thirty years.

Many people would have you grit your teeth and buy-and-hold stocks regardless of valuations. Historically, that has been a good choice because of the length of time.

These articles lead you to a better choice.

Start with this one.
Edited: You Can't Count on 7%: Dollars

Now fill in the details.
You Can’t Count on 7%
You Can't Count on 7%: Application
You Can't Count on 7%: Dollars