January 16, 2008 Letters to the Editor

Updated: February 9, 2008.

Stupid Assertions

I received this letter from Rob Bennett.

I enjoyed reading the "Stupid Assertions" article again when you highlighted it in the Notes section. I too find charm in the frankness of the title. The points you made in the article are of great significance.

I understand why many investors fall into the various traps. Most people are busy. Most people don't find it all that exciting to study investing. Most people don't ask hard questions and dig enough to resolve contradictions in what they read. This is all unfortunate but not so terribly shocking once you think it over a bit.

I struggle to understand why the "experts" fall into these traps. That's the killer for me. The experts have the time to dig deep. They should be aware of the traps. They should avoid them. They should point them out to others.

My sense is that the experts are to some extent telling people what they want to hear (I find this shameful behavior) and to some extent really are taken in by the traps themselves. My sense is that they allow themselves to be fooled on a few core points and then they genuinely lose the capacity to reason things through carefully on more subtle points.

I don't have a great deal of confidence in my judgments in this area. I do the best I can to make sense of what I see appear before me.

I don't view most of the people today referred to as "experts" as being genuine experts. They know many things. But they do not know what they need to know to give effective advice. If the advice you are giving does not stand a realistic chance of working in the real world, you are not a true expert, in my assessment.

HERE IS MY RESPONSE

Very well said.

Thank you.

Stupid Assertions

Is the Stock Market a Closed System?

I received this letter from Rob Bennett.

You recently highlighted an earlier article that contained these words:

"Dr. Hussman proved another important theorem. Money never enters or leaves the stock market. All dollars invested to buy shares of stock come out of the market as dollars removed by the sale of those shares. "

I'd like to explore in a bit more depth the question of the extent to which the stock market is a closed system. If you are able to provide a reference to where Hussman said what he said, I would be grateful.

I have become a big believer in recent years in the idea that what investors need to focus on is the fundamentals. Investing is a field in which all sorts of misleading claims are accepted because investors are emotionally inclined to go along with any claim suggesting that stocks offer a better value than they really do at times of high prices and a worse value than they really do at times of low prices. It's only by coming to possess a sure grasp of the fundamentals that we can develop the fortitude needed to see through deceptive claims.

One of the basic questions that trips many of us up is the question of where returns come from. The returns that we see reported on our portfolio statements seem real. If we were to cash in our shares today, we really would obtain a dollar figure close to the dollar figure stated as the value of our shares. That impresses many people as convincing evidence that those portfolio statement numbers are legitimate numbers.

Those saying that valuations matter are saying that those numbers are not legitimate numbers. We don't usually make our point that way. But, when you boil away the verbiage, that really is what it comes to. We are saying that there is more integrity to the portfolio numbers when stocks are at fair value than there is when stocks are wildly overvalued. I really think it is fair to say that that is the bottom-line point we are making. We are saying that at times of overvaluation investors need to mentally adjust down those numbers to have them reflect reality or else they are fooling themselves.

Investors trying to understand the valuations question are stumped by the fact that the portfolio numbers are treated as real -- you really can cash in your shares and obtain the statement amount. Are those numbers real or are they illegitimate? I think we need to try to answer that question as clearly as we possibly can.

I wrote a blog entry ("Poof!") a week or two ago aimed at beginning an effort to answer this question. The blog entry argued that the portfolio numbers are not legitimate, that any one investor can by offering $1 more for a share of stock than the newspaper price increase by $1 million the combined value of all shares as reflected in the new newspaper price. If there are 1 million shares of ABC Company outstanding, a decision by one shareholder to pay $1 more than the current stated price translates into $1 million of added wealth for owners of the shares of that company. The example is artificial in some respects; I simplified the trading process to highlight the point being made -- newspaper prices reflect the emotions of investors as much as they do economic realities. Prices can change without anything of economic substance having taken place, just by a change in investor emotions. Stock prices are not necessarily connected (at least for a time) to any great extent to economic realities.

I believe that the extent to which the stock market is a closed system or not has a bearing on the question. I believe that it is easier to analyze the question if the market is a closed system. If the market is a closed system, all that is going on is that investors are for a time kidding themselves into believing that their shares are worth more than they are and then at a later time kidding themselves that their shares are worth less than they are. To the extent that the system is not closed, it becomes more difficult to explain clearly what is going on. Then there are all sorts of exceptions and caveats that need to be noted that confuse the explanation of the most significant forces at play.

All of this is a long-about way of saying that I'd be grateful for any further input you can provide on the question of whether the market can fairly be described as a closed system or not.

HERE IS MY RESPONSE

Here is the reference: There's No Such Thing as Idle Cash on the Sidelines by John P. Hussman, Ph.D dated July 10, 2006.

There's No Such Thing as Idle Cash on the Sidelines

” In any case, stock prices don't change because money goes “into” or “out of” the market. Prices change because buyers are more eager than sellers, or vice versa. If a dentist from Poughkeepsie is eager to buy a single share of General Electric (which has about 10 billion shares outstanding), and pays $33.30 instead of $33.20 for that single share, that one trade will increase the stock market's capitalization by a billion dollars. But at the end of the day, all securities that were originally in existence are still in existence, and there is just as much “cash on the sidelines” as there was before.”

Here is a link to John P. Hussman’s Archived Comments.

John P. Hussman’s Archived Comments

Here is a link to Rob’s “Poof” articles. Look at January 25, 2008 12:33 - Newspaper Prices and January 31, 2008 12:46 - A Strange Babyland. They are excellent.

Rob Bennett’s “Poof” Articles (January 2008 blog)

The Trouble with "Mechanically Defined Algorithms"

I received this letter from Rob Bennett.

The thoughts below were prompted by a sentence that appeared in your most recent article ("6% is NOT Normal"). The sentence is: "Training improves results over mechanically defined algorithms."

I certainly agree with the thought. I am a big believer in the idea of using historical data to inform one's investing strategies. I am highly skeptical of the idea of becoming a slave to the message of the historical data.

Still, there's something in the sentence that troubles me. It seems to me that the lessons that one picks up through "training" (using tools like the Scenario Surfer) can be used to inform one's "mechanically defined algorithms." Doesn't it seem that over time we should be able to improve the mechanically defined algorithms until they do the job as well as training/experience? If not, why not? Where does the logic chain break apart?

I have two tentative thoughts.

A big problem is that numbers that show how effective a strategy is in terms of the returns it generates do not show the risk that had to be taken on to attain those returns. When I am using the Scenario Surfer, I find myself at times of moderate valuations being tempted to go with stock allocations higher than what my common sense tells me are right. The numbers that follow from going with such allocations are extremely appealing in the vast majority of cases. If one thinks of investing as a "game," (most algorithms are designed pursuant to a "investing as game" mindset), you would yield to this temptation.

But of course in reality investing is not a game. In a game, you play the odds. In life, you only get one chance to play; there are not always enough tries offered to get the odds on your side. Say that a 100 percent allocation works in 9 out of 10 plays. Do you bet that way? Game theory says "yes." I'm not so sure about life theory. It might be that the small chance of a loss argues for a measure of moderation.

The trouble we face is that we humans are always trying to rationalize our decisions. I want to make the Valuation-Informed Indexing approach sound as appealing as possible. If I tell people to go with a 100 percent stock allocation at a P/E10 of 17, that will make the numbers look better in the case I put forward for how this strategy pays off long term. It's hard for me to resist the pull to at least tell people what the numbers say (and I sometimes do indeed do just this). The other side of the story is that the few cases in which there are losses might do enough harm as to make it not a good idea to go with a 100 percent allocation no matter what the numbers say. I don't know that that argument can be reduced to numbers, though. It's a tricky business.

There are Behavioral Finance studies that show that investors are "irrationally" averse to losses; they are willing to give up more in the way of gains than they should be to avoid losses. I question whether this human instinct is truly irrational. To some extent, it no doubt is irrational; there is no question but that people sometimes apply the rule in irrational ways. My guess is that there is something "rational" driving it, however. It's an irrational element within a generally rational framework. This could be compared to how the Supreme Court is an anti-democratic institution that operates in a nation generally governed by democratic principles. There is something in us that doesn't like the idea of being entirely rational (or entirely democratic). My strong hunch is that that "something" within us is rooted in something real.

My foggy thought is that what might be going on is that the numbers by themselves don't show the risk that was taken to obtain them. When a certain strategy puts up good numbers over and over again, the numbers make too strong a case. It's not that the numbers are wrong. It is that the full reality is that there needs to be a counter to them. There was some risk attached to getting those numbers, and the gains added by going from an 80 percent allocation to a 100 percent allocation might not be great enough to justify taking on even that small amount of unnecessary risk. But to show that, we need to go beyond numbers and mixing numbers-based analyses with non-numbers-based analyses makes us uneasy. So we adopt a "veto" rule, where we tell ourselves "I will never go above 80 percent stocks no matter what the numbers say."

Is that rational or irrational?

I think there is some irrationality built into it, but I also think that it is ultimately more rational than an approach that in nominal terms is purely rational. I did not decide who to marry based on a purely rational analysis. Yet I am confident that I made the right decision. I suspect that had I gone by pure reason I might have messed up. I believe that some of the logical fallacies criticized by Behavioral Finance advocates are rooted in intuition and that there are circumstances in which intuition trumps reason.

Another factor is that where you are in the 30-year cycle (the Surfer examines 30-year time-periods) needs to affect your allocation calls if the analysis is to be true to life. Say that I am trying to figure out whether a stock allocation of 100 percent makes sense at a P/E10 of 17. It might be that the best answer is that it does make sense if one's portfolio has not yet grown too large (the risk here is less) but that it does not make sense if one's portfolio has grown large. I don't think it is possible for any one algorithm to make sense in both circumstances. The strategic considerations are different, so the answer to the question posed must be different.

If a 100 percent stock allocation really makes sense even at times where the portfolio is large (and the amount at risk is therefore large), then it logically follows that a stock allocation of greater than 100 percent must make sense when the portfolio amount is smaller. People don't like to say that a stock allocation of greater than 100 percent (borrowing to buy stocks) can ever be right, for quite proper and understandable reasons. The reality remains that the logic follows. If borrowing to buy stocks is wrong in some sense at times when the portfolio allocation is small, then going with 100 percent stocks when the portfolio amount is large must be wrong as well.

HERE IS MY RESPONSE

Thank you for some excellent comments and observations.

My biggest problem with mechanically defined algorithms is building calculators that implement them. It is not easy. In addition, I want to avoid algorithms that are too complicated. Such algorithms might show what was best in the past, looking back historically, instead of what will be best looking forward.

Using the Scenario Surfer can help avoid the second problem. Still, the standard for comparison will always be the historical record combined with commonsense.

I was surprised when I built my first retirement trainer at how naturally I reacted better than with previously examined algorithms. I had thought of Latch and Hold as being sophisticated. Yet, it ignores some obvious factors such as the current balance and the time remaining. For example, if you do exceedingly well early, why not lock in success by buying TIPS?

Behavioral Finance researchers have succeeded in getting different responses to the same question from groups of people depending upon the exact wording. We see similar behavior in political polls. Change the words slightly and the percentages change.

I believe that the Scenario Surfer’s greatest benefit is as a trainer. After picking up a reasonable level of skill, it is a good idea to make a few runs SLOWLY, focusing on all of the details each year. By going through slowly, you reduce the game-like atmosphere. You can concentrate on training. You learn about yourself and how you react to numbers. That is the key. You learn about yourself.

Making enough runs tempers our actions. If you choose a 100% stock allocation when P/E10=12, how do you feel when P/E10 drops to 6? With enough runs, it happens. P/E10=7 or 8 happens even more often. Similarly, when starting with P/E10=26 and allocating nothing to stocks, how do you react when P/E10 climbs to 40 or 50? I now set a minimum allocation around 20% until P/E10 rises above 30. At some point, enough is enough and I want to be out of stocks. But no longer at P/E10=26.

Letters to the Editor in 2008

Letters to the Editor in 2007

Letters to the Editor in 2006

Letters to the Editor in 2005