January 30, 2009 Letters to the Editor

Updated: February 3, 2009.

Observations from your P/E10 Sequences

I received this excellent letter from Michael.

Hi John - as always, thanks for the food for thought. I had two take-aways from your post today (1/30/09) on 20 year P/E10 Sequences:

1) in ALL of the 10 Runs, the P/E 10 at year 20 was lower than today's. That surprised me so I then wondered whether the data suggested we would have higher P/E10s in the interim but just fall again by year 20.

2) However, of the 40 data points (10 runs * 4 observations per run), there were only FIVE observations above today's 14, and if you exclude the three that were essentially the same as today (14.1, 14.1, and 14.6), only TWO of the 40 were substantially above today's 14 - a 17.3 and a 19.6.

Sobering data even for this perma-bear! Is this just another way of saying that secular bear markets take a long time to work themselves out? I wonder what you mean when you conclude that "Much of the randomness has returned by Year 20" - these numbers seem to me to suggest hardly any randomness at all, just a continued unwinding of crazy high p/e 10 levels - agree or no?

Thanks John

HERE IS MY RESPONSE

I am guilty. I did not discuss the issue of turning points adequately.

The market type will change to a Normal Market shortly after P/E10 falls to 10.0 or so. The exact amount of time that it will remain low is uncertain. It should only be a few years, not many.

Our ability to estimate turning points is limited. The best that I can say is that the market takes about 30 to 40 years to go from one peak to the next. Expect another market peak around 2030 to 2040. Cutting this time period in half [from 2000 to 2030 or 2040], we might expect the market to bottom between 2015 and 2020.

Alternatively, I have seen an estimate based on demographics between 2012 and 2020.

Do not become a perma-bear. In fact, if you can weather the next few years, you should expect to enjoy a long lasting (secular) Bull Market.

Rational Investors Do Not Become Permabears!

I received this letter from Rob Bennett.

John:

The exchange between Michael and you ("Observations from Your P/E10 Sequences") highlights a point which I believe is of huge importance but which I believe the newcomer to Rational Investing can easily miss (it is a bit counter-intuitive).

Say that an angel came down from the sky today and told us that the P/E10 level is going to be lower 10 years from today than it is today. We know this with absolute certainty to be so (this is an angel known for her honesty). Do we lower our stock allocations in response to this glimpse into the future?

We do not!

A lower P/E10 value might be a P/E10 level of 12 or 13 (today's number is 14). That means a healthy return for those invested heavily in stocks for that entire time-period. If stocks remained at the same P/E10 level for the entire time-period, the annualized real return would be something in the neighborhood of 6 percent. If the P/E10 drops only a little over the time-period, the return will be less than 6 percent, but still high enough to beat just about any alternative investment class. Small drops in P/E10 value are good news for those heavily invested in stocks.

This is a critical point that must be understood by all using the historical data to ascertain the probabilities for various investing strategies. We intuitively think that an increase in P/E10 value is good and a decrease is bad. But that's not quite right. Stocks on average provide such a good return that it takes a big drop in P/E10 to make stock investing unprofitable. It's only when P/E10 has gone so high that big drops become likely that stock investing becomes dangerous.

You explored this idea in an article you wrote saying that the reason why the dangers of investing in stocks at times of high valuations has not been generally accepted yet is because times of extremely high P/E10 levels are so rare that valuations risk is usually not an issue. That article offers a great explanation of why our understanding of how stock investing works remains so primitive. The reality is -- the rules are always changing. "Experts" see how stocks perform in one environment and conclude -- "There are the rules." Then the environment shifts and all the rules are stood on their heads. Paradoxically, those who are "expert" enough to have studied the rules are the last ones you want to go to for investing advice!

We need to understand the right principles before we can develop the right rules; our thinking must proceed from the general to the specific or it will fail us. We need to understand what is going on underneath the surface so that we can develop rules that remain in effect in greatly varying environments.

The rule here is -- don't worry too much whether P/E10 is rising or falling. Stocks can be a good bet even during times of falling P/E10 levels. The principle is -- Focus on risk. When the P/E10 level exceeds 20, the risk of a big drop in P/E10 value (and a devastating loss to your portfolio) is high. That's dangerous. You don't know for sure when the big drop is going to take place or how big it is going to be, but you know that an extraordinary amount of risk is present. It is that negative potential that compels a lowering in your stock allocation. The negative potential is not present during a time when all that is likely is a small drop in P/E10 level or a large drop not likely to remain in place for long.

These observations are the product of a different mindset than the mindset of the Conventional/Passive Investor. We are not really trying to "outsmart" the market; that's their language. What we are trying to do is to get the probabilities on our side (perhaps it could be said that we are adapting to the changes in the market rather than trying to outsmart it). That's a subtle but very important distinction.

We must warn people not to become perma-bears. To be a perma-bear is as bad as to be a Passive Investor. The two psychologies are two sides of the same emotional coin. The biggest danger we face today is that millions of middle-class investors are going to become perma-bears because they fell for the Passive Investing gibberish and are now disillusioned with how it works in the real world. Stocks cannot be blamed for these losses. The market cannot be blamed for these losses. The blame lies with the Passive Investing model, which encourages reliance on emotions for guidance and which rules out the use of human reason in execution of the investing project.

Perma-bears are thinking that stocks are the enemy. The historical data does not support that claim. It is the Passive Investing approach to stock investing that is the enemy. The entire struggle lies in getting over the way of thinking that is developed once one accepts that it might be possible to invest successfully for the long term without on occasion making adjustments to one's stock allocation. That idea is the madness at the root of all of the economic turmoil that we suffer from today.

HERE IS MY RESPONSE

Thank you for a most insightful letter.

Here is the referenced article along with its edited version.

Why People Ignore Valuations
Edited: Why People Ignore Valuations

Follow-on to your and Rob's comments to my post...

I received this letter from Michael.

Thank you both for responding. A little more clarification and response to Rob: Please don't take my questions (before or now) as second guessing the VII methodology - you have won a convert! My questions go more toward understanding the boundaries of what you are arguing, and making sure I understand the links you are making and the arguments you are not.

So, following on Rob's "Angel" analogy - the "Angel" that precipitated my original post in fact said something more explicit than "P/E 10 will be lower" - the Angel (John) gave me 10 possible Year 20 levels. And the fact that none were above and most were considerably lower is what I was probing. So based on John's 10 runs, I calculate a range of 20 year real returns of 5.6% to 2.7% with an average of 4.05%. I assume 6% real pa from you, then haircut based upon the final 20 year p/e 10 level - for example, if p/e 10 stayed the same for 20 years at 14, I assume 6% pa, but the highest of John's year 20 p/e 10 is 13, so I solve for the annual real % gain by saying 1.06^20 * 13/14 (the final year 20 over today's) = 2.978 ^ 1/20 - 1 = 5.6%. And so on for the rest.

My surprise was not that year 20 p/e 10 was about the same or slightly lower, but A LOT lower, imho. And this assuming that we in fact bottom out in 5-15 years and by year 20 should be well on our way to higher p/e 10 levels.

So, either 1) I should not surprised at these levels, OR 2) John's 10 runs were too small a sample size and if we had more, we would get a larger distribution, OR 3) some other interpretation that I would like to know.

Also, how do I reconcile my imputed annual % gains based on John's year 20 p/e 10 (again, 2.6% to 5.6% with an average of 4.05%) with the Stock Returns Year 20 "Best Possible" 10.3%, most likely 6.3%, Unlucky of 4.3%, and Worst Case of 2.3% - these Stock Returns seem to be more in line with what Rob is implying (to paraphrase - "around 6%, maybe a bit more, maybe a bit less") and more generous than those levels implied by John's 10 x Year 20 runs.

Thoughts?

HERE IS MY RESPONSE

This is outstanding. Thank you for your analysis.

First: you are right regarding the Year 20 returns in a Bear Market. The Stock Returns Predictor (Button on the left) projects Normal Market returns. I have a separate calculator for predicting Bear Market returns. It uses Ed Easterling’s definition of Bear Markets. Reference: his book Unexpected Returns and his www.CrestmontResearch.com web site.

My Bear Market predictor is built into several of my Retirement Trainers, including the Type 2A Bull Bear Retirement Trainer. They are available for downloading (free) from my Yahoo Briefcase. My Yahoo username is jwr19452000. Look in the Retirement Trainers folder. You will find a variety of retirement trainers. They were the prototypes that led to the Scenario Surfer.

You can read all of my equations (with effort) from this spreadsheet.

Yahoo Briefcase

You should not be surprised at the calculated P/E10 levels. The calculations ASSUME that the Bear Market continues to persist. That is, my calculations do not yet incorporate turning points.

I assume an earnings growth of 6.8% per year (real).

This particular sample turned out to be somewhat pessimistic, but not overly so.

Now some news.

Rob Bennett and I are in the early stages of building yet another calculator. Our thinking is very vague at this time.

Rob likes the notion that high P/E10 levels lead to price drops and that low P/E10 levels lead to rising markets. He would like to teach this lesson. The difficulty is that he wants something realistic and the market does not always work that way. We saw a record bubble that peaked in 2000. Until recently, it was prudent to prepare for an even bigger super bubble. The odds were not high, but they were high enough. After all, people need to prepare for the unusual as well as what is typical.

My idea was to generate a large number of sequences and then sort. The P/E10 Sequences article was my first effort along these lines. I had already chosen twenty year segments to include Mean Reversion properly.

I expect to revisit some of these ideas as I go along. For the moment, I am seeing where they lead.

Once you work with twenty year segments, you can do all sorts of things that the existing calculators do not do well. You can link segments together realistically. If one segment ends at a low P/E10, for example, it is close to a turning point. It is near the beginning of a Bull Market. The next segment will be favorable for retirees.

The Safe Withdrawal Rate calculation over 60 years does not have to be the worst case for two 30 year segments, which is what we now calculate. Rather, we can calculate it using three realistically combined 20 year segments.

Similarly, realistic twenty year segments are likely to give us insights into turning points. We can look at the highs and lows just before the end of each segment.

Returning to Rob Bennett’s training purposes, you might select sequences that go up in the first few years and then sort according to what happens later and use that for training. You might allow the user to do his own sorting. This would translate vague notions about the market into something more tangible and realistic.

Now for some further news: Rob Bennett and I are changing the name of the Reality Checker to The Investment Strategy Tester. The new name is more descriptive of what it does.

Michael Is Asking Cutting-Edge Questions

I received this letter from Rob Bennett.

Michael:

I understand that you are a "convert." I was not thinking of you as the "newcomer" who might become a permabear. I had some others in mind who in the past have gone along with the Passive Investing concept but who now are thinking of getting out of stocks entirely. This worries me because, if too much of this sort of thing happens, we will not learn from our mistakes --we will just keep going around in circles of being overinvested in stocks at times when the value proposition is poor and being underinvested in stocks at times when the value proposition is good. And of course the big-picture reality is that that causes all sorts of economic turmoil. I've been thinking about this problem a lot lately and, when these sorts of things are on my mind, discussion of them tends to come out in my posting on all sorts of topics. In short, I was using my response to your comments to engage in some general venting not intended to be aimed at you in particular.

You are asking some cutting-edge questions. You are indeed pointing to a finding that is counter-intuitive. That suggests that our understanding is not complete and that we need to dig deeper to come to a better understanding.

I agree with the suggestion in John's post that calculators may help us advance in our understanding of the realities. The Strategy Tester will be out soon. That will let us do the sorts of investigations that John often performs in his analyses in much less time and thereby will permit us to do more of them. We haven't made a decision t go forward with a fifth calculator. But, if it looks like the follow-up calculator would help in addressing the sort of question you are raising, that's a strong argument for going ahead.

I cannot at this time offer anything more than vague impressions as to what is going on. I think we (I mean the entire community of investors, not just us three) need to spend some time thinking about the points you have made and try to come to a reasoned understanding of the realities as they present themselves to us in the data. So anything I say now is tentative in nature.

My vague impression is that what you are seeing relates to the nature of the return sequences that come up. John gets at this in his comments. I don't fully have my head around it. I need to continue thinking about it.

I think what makes things tricky is that there are two forces always affecting future stock prices and the two forces pull things in opposite directions. There is always a momentum factor that causes stock prices to continue moving in the same direction in which they are currently moving; when prices have been going up, they tend to continue going up, and when prices have been going down, they tend to continue going down. This factor is primarily emotional in nature. Then there is Reversion to the Mean factor that causes stock prices to move in the OPPOSITE direction from the direction in which they have long been moving once they have been moving in that direction long enough to bring about significant overvaluations or undervaluation problems. This factor is primarily economic in nature. We have little understanding of what causes turning points. The turning point (the point at which the Reversion to the Mean factor becomes more powerful than the Momentum factor) can for the most part only be identified on a looking-backward basis.

I think that you may be in the early stages of developing a means of saying something about when turning points are likely to take place. I don't think that we are ever going to have precise guidance on this question. But it might be that we will be able to identify clues. It does seem highly surprising to me that the odds seem to be high that in 20 years the P/E10 value will be lower. That's not what I would have expected to see. I think it is worth devoting more consideration to this to try to ascertain why such a thing would be so.

Here's a very tentative thought. Say that we have two directions to go today. It could be that efforts to restore investor confidence are quickly successful and that the P/E10 level will rise over the next few years. Or it could be that those efforts will fail and that we will see another 50 percent drop in valuations. Which is better? It could be (again, this is tentative speculation!) that there is some sort of built-in psychological need for us to go lower. If that were so, it could be said that we are better to go lower soon and get it over with. It could be that going higher in the short term would lock us into a scenario in which we would suffer another price crash down the road a bit. By going higher now we would be stretching out the going-lower process. I don't say that this is so. I am just playing with possibilities.

If something like that is so, it might be that the ultimate reality reveals itself only after the completion of 20 years or so. It might be that we choose different paths to recovery each time but that they all end up being in some sense similar at the completion of a long-enough time-period. I doubt very much that things are entirely deterministic or even close to being so. So I think we need to be cautious in what we say about this sort of phenomenon. We might be able to develop some informed guesses re the probabilities of various long-term scenarios. If we place too much confidence in our informed guesses, we may end up doing ourselves more harm than good. I believe that we should explore these questions but in a spirit of humility re our ability to develop truly valuable insights by doing so.

Valuations are emotions. It is critical to understand this. When the developers of the Passive Investing model left out the consideration of valuations, what they were leaving out was the consideration of human emotions. As we move deeper into a Rational Investing analysis (it is rational to accept that emotions matter), we will be confronting more and more questions that are more psychological than statistical in nature (but in many cases looking at the data will help us come to a better informed understanding of psychological realities). We need to become more comfortable mixing the two types of analysis (statistical and psychological). The rules for analysis with the two types of questions are generally very different. That at times makes it hard to know how to proceed and how to state things properly.

The bottom line is that we have no choice. The fundamental realities are in part economic in nature and in part psychological in nature. That tension is always present, whether we like the idea or not.

I believe that you have identified an anomaly of considerable importance. I am grateful to you for doing so and I am going to continue to think about it in the months to come. My experience with this sort of thing is that, when we dig to make sense of this sort of thing, we learn all sorts of things that we were not expecting to learn when we made the decision to do so.

HERE IS MY RESPONSE

Rob, thank you, as always, for a valued letter.

For Michael and others: we really are engaged in cutting edge research at this site. Do not be concerned with the vagueness that typifies this stage. We will reach clarity before long.

Letters to the Editor in 2009

Letters to the Editor in 2009

Letters to the Editor in 2008

Letters to the Editor in 2008

Letters to the Editor in 2007

Letters to the Editor in 2007

Letters to the Editor in 2006

Letters to the Editor in 2006

Letters to the Editor in 2005

Letters to the Editor in 2005