MPT discussion

Research on Safe Withdrawal Rates

Moderator: hocus2004

User avatar
Alec
Admin Board Member
Posts: 31
Joined: Wed Sep 10, 2003 4:00 am
Location: Crofton, MD

MPT discussion

Post by Alec »

The previous conversation [started by unclemick] got me thinking about Modern Portfolio Theory [MPT]. Personally, I have some concerns that people [not people here, just people in general] may be mis-interpreting it, how to use it, its limitations, etc. I was wondering what you guys thought of it, how you thought it is/was being misused/misunderstood, etc.

To start off, I'll just throw out two of my gripes w/ MPT:

1. It assumes that portfolio volatility is the only source of risk, when, in reality, the risk I'm most concerned about is accumulating [or not] $X by "x" date. So, I don't really care about the yearly volatility of my portfolio, as long as its value converges to a future value [note: stocks don't do this, but stripped TIPS do].

2. The math can seduce people into millions of portfolio maximizations, when the underlying assumptions [future correlation, SD, returns] are totally unknown and constantly changing. Piggy backing on this is for people to conclude that there is one "uber" risky portfolio that everyone should hold [in proportion to their risk aversion], when, in reality, each person holds different risks b/c of different jobs, different living areas, etc.

Thought it'd be an interesting conversation, and give me something to think about on the drive to work. :D

- Alec
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"I have some concerns that people [not people here, just people in general] may be mis-interpreting it, how to use it, its limitations, etc. "

Thank you for saying this, Alec.

"I don't really care about the yearly volatility of my portfolio, as long as its value converges to a future value"

Thank you for saying this, Alec.

"Piggy backing on this is for people to conclude that there is one "uber" risky portfolio that everyone should hold [in proportion to their risk aversion], when, in reality, each person holds different risks b/c of different jobs, different living areas, etc. "

Thank you for saying this, Alec.

That's three line drives in a single post. I hereby proclaim you the Ted Williams of the SWR Research Group!
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

Here are some thoughts that are a little bit out of the ordinary.

The mathematical foundation behind Modern Portfolio Theory is seriously flawed.

From The (Mis)Behavior of Markets by Benoit Mandelbrot and Richard Hudson, chapter 5, The Case Against the Modern Theory of Finance.

Benoit Mandelbrot is the inventor of fractal geometry.

Mandelbrot identifies these Shaky Assumptions on pages 82-107.
1) People are rational and aim only to get rich.
2) All investors are alike.
3) Price change is practically continuous.
4) Price changes follow a Brownian motion.

In his remarks about the fallacies behind assumption 1), Mandelbrot mentions the new field of behavioral economics.

Assumption 2) requires that investors have identical goals, time horizons and measures of success. Mandelbrot points out that as soon as an assumption of perfect homogeneity is relaxed, the mathematics goes out the window. It only takes two types of investors, fundamentalist who look at intrinsic value (to buy low and sell high) and chartists who try to exploit short-term price trends, for bubbles and crashes to show up.

Assumption 3) is demonstrably false. Significant news, such as the approval of a new drug or a buyout offer, causes big jumps in prices. Even without anything special, prices tend to move in increments of a nickel (i.e., prices most often end with either 0 cents or 5 cents) per share.

Assumption 4) requires that price statistics are independent, stationary and Gaussian. This has caused major blunders in estimating risk. One of Mandelbrot's main points is that market risk is much higher and huge price changes occur much more frequently than anything closely related to these assumptions would permit. Market prices are best characterized as having turbulent behavior, not the gentle behavior associated with Brownian statistics. Coin tossing is out. Weather patterns and hurricanes are in.

Have fun.

John R.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

Alec wrote:2. The math can seduce people into millions of portfolio maximizations, when the underlying assumptions [future correlation, SD, returns] are totally unknown and constantly changing.
More from Mandelbrot. From page 233:
The same kind of simulations can be done for stock, bond, or other financial prices. According to the standard model of finance, in which prices vary according to the bell curve, the odds of future ruin are about 10^(-20). Translation: One chance in ten billion billion...But if prices vary wildly, as I showed in the cotton market, the odds of ruin soar: They are on the order of one in ten or one in thirty. Considering the disastrous fortunes of many cotton farmers, which estimate of ruin seems most reasonable?
Regarding the patterns that Alec mentions (..when the underlying assumptions [future correlation, SD, returns] are totally unknown and constantly changing..), it may be that they are actually the result of long-term time dependence in the data. What happened a long time ago might be making the patterns (i.e., the correlations, SD, returns, etc.) that we think that we are seeing today. That is, events from long ago plus an element of chance introduce false patterns and correlations that are incredibly deceptive.

On pages 244-247, he writes that Market are Deceptive. From that section:
As described earlier, the long-range dependence in prices creates a kind of tendency in the data--not towards any particular price level, but towards prices changes of a particular size or direction. They can be persistent, meaning that they reinforce each other; a trend once started tends to keep going. Or they can be anti-persistent, meaning that they contradict each other, a trend once begun is likely to reverse itself...In our research..generated such records by the purest operations of chance. Nevertheless, they all appeared to display a long, slow, up-down cycle of three; upon those long waves, smaller and more numerous cycles seemed to interpolate themselves. When we looked at a small section of the record, we again saw three waves, each a third shorter than the section.
A major benefit of Mandelbrot's research will be improved random number generators for our Monte Carlo models. They will present a much more realistic picture which includes the risky, turbulent behavior typically found in market prices. At the moment, researchers are limited to running stress tests [or running sequences taken from limited sequences of historical data].

Have fun.

John R.
User avatar
Alec
Admin Board Member
Posts: 31
Joined: Wed Sep 10, 2003 4:00 am
Location: Crofton, MD

Mandelbrot

Post by Alec »

Hey John,
1) People are rational and aim only to get rich.
2) All investors are alike.
3) Price change is practically continuous.
4) Price changes follow a Brownian motion.
This sounds more like attacking the CAPM rather than portfolio theory.

Porftolio theory certainly does allow for people to be different and hold different portfolios, simply because people have different risks. Heck, my finance textbook specifically says this. Its the CAPM that says that everybody is the same, has the same holding period, and thus holds the same risky portfolio.

Economic theory says that people seek to maximize Utility, not necessarily aim only to get rich. It's really interesting, to me anyway, that most people aren't utility maximizing machines [see most of the behavioral finance literature]. Heroine addicts don't see that it's not in their best interests to keep using. I used to get more utility out of getting hammered, but now I don't. ;) It was pretty funny that my econ professors would start out teaching Utility theory by saying, "This is how people should make decisions, but no one does".

I think portfolio theory is more like common sense. Common sense would say not to hold to concentrated a portfolio, like owning a whole lot of stock of your employer or of your industry. Or holding all nominal assets if you've got a nominal pension. Too bad people don't have a lot of common sense. It's all about hedging risks, and it certainly doesn't help that there is a whole financial services industry out there giving people bad advice about how to go about hedging these risks.

- Alec
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »


hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

The biggest flaw in the conventional methodologiy studies is the failure to take into account the effect of changes in valuation levels. What the conventional studies report is not the SWR that applies at the time the aspiring early retiree's retirement begins, but a mix of the SWRs that apply in low valuation years, medium valuation years, and high valuation years. I once asked JWR1945 what the PE10 would need to be for an SWR of 4 percent to apply. My recollection is that the answer was that you would have an SWR of 4 percent if you retired at a time when the PE10 was 18. I believe that he said that a PE10 of 18 is a moderate sort of PE10 number. So the conventional studies are giving you a sort of average SWR, not the SWR that actually applies for your particular retirement.

Is there anyone who is familiar enough with the literature on Modern Portfolio Theory to say whether it was a belief in MPT that caused this analytical error on the part of the first person who put out a conventional methodology study? My sense is that supporters of the conventional methodology are generally sympathetic to Modern Portfolio Theory. But is it fair to say that the theory is what caused them to get the SWR wrong? Is there something inherent in Modern Portfolio Theory that requires its proponents to deny that the intrinsic value of a stock purchase varies with changes in valuation levels?.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

hocus2004 wrote:Is there anyone who is familiar enough with the literature on Modern Portfolio Theory to say whether it was a belief in MPT that caused this analytical error on the part of the first person who put out a conventional methodology study?
Let me come from the opposite direction.

One of the biggest advantages of looking at historical sequences [which is basic to the conventional methodology] is that you do not have to make assumptions about statistical distributions. This is in sharp contrast with Monte Carlo models, which have random number generators built into them [and, most of the time, they generate the standard bell curve (also known as the normal distribution or the Gaussian distribution)].

Using historical sequences takes away the most obvious, big problem with Monte Carlo models.

Using historical sequences does not solve all problems. Instead, it replaces the most obvious problem associated with Monte Carlo models with a different set of problems.

Proponents of the historical sequence method perpetrated a fallacy by treating a plausibility argument as fact: the argument that the existing historical sequences provide sufficient information to answer all questions about Safe Withdrawal Rates directly and that there is no need to investigate further. They assumed, but never proved, that nothing of significance was yet to be discovered. Valuations were irrelevant because they were assumed to be irrelevant.

Have fun.

John R.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

hocus2004 wrote:My sense is that supporters of the conventional methodology are generally sympathetic to Modern Portfolio Theory. But is it fair to say that the theory is what caused them to get the SWR wrong? Is there something inherent in Modern Portfolio Theory that requires its proponents to deny that the intrinsic value of a stock purchase varies with changes in valuation levels?.
IMHO, it is the belief in the Efficient Market Hypothesis makes intrinsic value irrelevant.

There is a strange pick and choose process involving which assumptions one chooses to believe going on in the background. We are led to believe that we cannot improve investment returns, no matter how clever we are, but that we can easily identify and reduce volatility by careful asset allocation.

It is assumed that anything that might produce a better return will be arbitraged away. It can never persist. But actions that reduce risk will always persist. They will never be arbitraged away.

In this strange world of assumptions, there can never be any such thing as a variation in intrinsic value. It would automatically disappear. On the other hand, the risk premium for investing in stocks will always remain the same. It does not matter that just about everybody knows that stocks are always best in the long run. You still get your bonus, the risk premium.

Have fun.

John R.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"Proponents of the historical sequence method perpetrated a fallacy by treating a plausibility argument as fact: the argument that the existing historical sequences provide sufficient information to answer all questions about Safe Withdrawal Rates directly and that there is no need to investigate further."

That sounds right. I would add an important wrinkle that relates specifically to the study published at RetireEarlyHomePage.com.

The early conventional methodology studies did a very good thing. William Bernstein described the Trinity study as "breakthrough" research, and I share his assessment of the Trinity study. There are two critical ways in which the Trinity study and other conventional methodology can be distinguished from the REHP study.

One, the earlier studies were produced at times of lower valuations. The number in the earlier studies were off a bit because of the invalid methodology used. But they were generally in the right ballpark. The earlier studies came out at a time when 4 percent was a not entirely unreasonable rule of thumb.

Two, the earlier studies expressed a note of caution re stock investing. Most investors had thought that it was safe to use a take-out number of more than 4 percent, and the Trinity study set them straight. So I see the authors of the Trinity study as heroes to aspiring early retirees. They did a very positive thing with their research.

Intercst has used the REHP study for a purpose exactly the opposite of the purpose to which the earlier studies were used. Intercst has used it to shut down any discussions of the realities of stock investing. There are many people on the various boards who have tried to make arguments that there are circumstances in which real estate can be a good investment or in which TIPS or ibonds can be a good investment, or whatever. The intercst line is that the historical data shows beyond any doubt whatsoever that anyone who even considers any investment other than stocks is suffering from mental illness. These claims are absurd and dangerous and have been used (combined with intimidation tactics) to block discussions on dozens of topics that aspiring early retirees very much need to discuss.

One of my biggest concerns is the damage that intercst and his REHP study has done to the credibility of SWR analysis. I have used the data-based approach to SWR analysis since the mid-90s and have found it to be a powerful took for building safe long-term income streams. Intercst and the other DCMs have turned SWR analysis into a cartoon. When you manipulate the numbers in the manner in which the DCMs do, you can "prove" that any take-out number you want is "100 percent safe." The DCMs employ different meanings for all the important terms involved in SWR analysis. They describe a take-out number that the historical data reveals as high-risk as being "100 percent safe." They claim to have looked at "the worst-case scenario" when in fact they consider only data from the month of January because considering data from other months makes the case for stocks far less strong. Intercst refers to the practice of looking at data from non-January months as "cherry picking" when most researchers would refer to the practice of limiting the scope of your investigation to a single month as more akin to "cherry picking" than looking at all the data available. DCMs refer to a 74 percent S&P allocation as "optimal" although the data says that this is an allocation likely to produce poor intermediate and long-term results in the event that stocks perform in the future as they have in the past. And so on.

My point here is that the real danger is the REHP study, not SWR analysis in general. SWR analysis is wonderful. Some of the early researchers in this field made some mistakes, but that is to be expected. It is not reasonable to expect the first person who uses an analytical technique to get everything right. There is no reason to believe that researchers other than intercst would respond in the manner in which he has to questioning of the conventional methodology. My guess is that the authors of the other conventional methodology studies will be happy to have the flaws of their methodology pointed out to them and will work with us to improve the SWR tool over time. Most of these people are competent and well-intentioned researchers who are trying to help people with their research, not to place large numbers of early retirees at risk of busted retirements because of their pursuit of some strange personal agenda of exploring the outer reaches of extreme ego-puffery.

You are absolutely right that there is a need to investigate further. That is going to be the case for a long time to come. SWR analysis is still in its early days of development. The DCMs have tried very hard indeed to turn this analytical tool into a joke, but it is not a joke. It is the DCM claims that are a joke. It is my expectation that aspiring early retirees will be making productive use of valid SWR analyses for many years to come.
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"IMHO, it is the belief in the Efficient Market Hypothesis makes intrinsic value irrelevant."

OK. It's not clear to me where the Efficient Market Hypothesis and Modern Portfolio Theory intersect, if in fact they do intersect at some points.

"In this strange world of assumptions, there can never be any such thing as a variation in intrinsic value."

OK. Well, I think it is fair to say that, if the Efficient Market Hypothesis says that, then the Efficient Market Hypothesis is a big fat hairy donkey, if you get my drift. I think there are times when it is a good idea to take this stuff out of the ivory tower and try to get across to fellow community members the message that they most need to keep in mind when putting together their Retire Early plans. My suggestion to aspiring early retirees is, when you hear a DCM-dominated thread suggesting that the intrinsic value of your investments is irrelevant to your hopes of enjoying a successful early retirement, RUN AND DO NOT WALK to another, more reasoned and better informed thread.
unclemick
*** Veteran
Posts: 231
Joined: Sat Jun 12, 2004 4:00 am
Location: LA till Katrina, now MO

Post by unclemick »

Math and history? History or math?

My concession to MPT is of course Charles DeGaul:

"God looks after Drunkards, Fools, and The United States of America".

Hold the low cost pie chart appropriate for your age(low cost balanced index). Now if you plan to ER at 45 instead of 65 - then appropriate is a judgement call.

When in ER - learn to play some defense ala Bear Bryant: "agile, mobile, and hostile". Hence my semi Ben Graham 50ish percent stock/fixed supplemented with REITs and dividend stocks to get 3-4% SEC yield.

Even a cursory look at some of the data strings published on this forum showing div yield down history lane in comparason with Vanguard's Balanced Index 2.55% current yield ought to tell you something. So - look to your span time to ER/and or span time IN ER and adjust accordingly. Remember Geraldine Weiss: "Dividends Don't Lie".
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

Alec wrote:I think portfolio theory is more like common sense. Common sense would say not to hold to concentrated a portfolio, like owning a whole lot of stock of your employer or of your industry. Or holding all nominal assets if you've got a nominal pension. Too bad people don't have a lot of common sense. It's all about hedging risks, and it certainly doesn't help that there is a whole financial services industry out there giving people bad advice about how to go about hedging these risks.
We may have a variety of definitions of (modern) portfolio theory: one for each of you and half a dozen for me.

Managing risk does seem to be a big part of modern portfolio theory. And Alec is right to question the usual definition of risk. It is mathematically convenient, but not necessarily what we are interested in.

My biggest concern is an underlying inconsistency in what we read. Sometimes, we are told that there is arbitrage. At other times, we are to act as if there is none. That does not make sense.

In addition, I agree that just about any advantage in the market can and most likely will be subjected to arbitrage. I do not agree that the advantage will automatically disappear entirely. It will be reduced, but not necessarily be eliminated.

I expect arbitrage to reduce the advantages of dividend-based strategies, but not to eliminate them. Similarly, I expect that the advantages of value strategies will continue to persist.

Sometimes, an advantage can turn into a disadvantage. This can happen when just about everybody knows the same thing. As with the Dogs of the DOW, salesmen learn how to sell a financial product and extract their fees while others engage in front-running to extract excess profits.

I do not understand why I am supposed to believe that risk reduction techniques are free from arbitrage. If we can reduce the risks associated with stocks, why should we expect for the risk premium to remain the same? Shouldn't our risk reduction techniques be as vulnerable as everything else?

Have fun.

John R.
JWR1945
***** Legend
Posts: 1697
Joined: Tue Nov 26, 2002 3:59 am
Location: Crestview, Florida

Post by JWR1945 »

unclemick wrote:Remember Geraldine Weiss: "Dividends Don't Lie".
Sometimes they do. And then you can lose your shirt.

Make sure that those dividends are secure. Take a look at a company's credit rating. Otherwise, watch out below!

Have fun.

John R.
unclemick
*** Veteran
Posts: 231
Joined: Sat Jun 12, 2004 4:00 am
Location: LA till Katrina, now MO

Post by unclemick »

True true - G.W. uses Ben Graham type metrics and Dreman likes credit quality/low P/E to identify quality companies. They both overtrade - relative to my 7-10 yr holding period. And they lean toward large value companies instead of small. Me too.
unclemick
*** Veteran
Posts: 231
Joined: Sat Jun 12, 2004 4:00 am
Location: LA till Katrina, now MO

Post by unclemick »

I'm skimming through a library copy of Dreman's 1982 opus - brings back some memories of certain 'less than stellar' investments I made during the period - succumbing to recency, groupthink, etc, etc. Luckily it was mad money - I paid virtually no attention to max 401k deductions 50/50 ala Ben Graham since they were auto deducted and boring (heh,heh - lucky me). At least I can look back and laugh at myself with the distance of time. I can remember being disappointed only getting 14% interest on my Putnam closed end bond fund while Treasuries were shooting up - and still worried about besting inflation. Gold, silver, real estate, collectibles - hmmm - 22 yrs ago - ??due for a repeat??
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

The idea that the intrinsic value of a stock purchase does not vary with changes in valuation levels is the most unusual of ideas. It exerts a great influence on the thinking of most middle-class investors today. It has been incorporated into thousands of studies and calculators and money advice books and articles. It does not stand up to even the tinest bit of informed scrutiny. The historical data flatly contradicts it. The best informed stock analysts dismiss it out of hand as nonsense. And it seems to be almost impossible to identify anyone who is willing to put his or her name to a direct defense of it. That's quite a combination of attributes for a single idea to possess!

Imagine that a similar claim was being made in reference to real estate investment. Say that two houses in the same development go up for sale on the same day. They each have a market value of $100,000. When the first propesptive buyer enters House A, he is panicky about losing it to someone else and he offers $150,000 and the house is his. When the first prospective buyer enters House B, it is the seller who gets panicky and offers to let go of it for $50,000 and this buyer too takes the property. Is there any sane person who would argue that in the long term the buyer of House B does not possess an edge in terms of the long-term return he is likely to obtain from his investment? It is simple common sense that it is better to pay $50,000 for an asset worth $100,000 than it is to pay $150,000 for an asset worth $100,000. Move on to a discussion of stock investing, and this obvious truth becomes highly controversial, controversial enough to result in the burrning of several board communities to the ground.

One of the most amazing things about the nonsense claim that valuations do not matter is that, as influential as this idea has become (it is the core "insight" of the 'Stocks for the Long Run' Paradigm), it is hard to find someone willing to step forward and claim credit for having come up with the idea. Usually, there are all sorts of people rushing forward to claim credit for influential ideas. Not in this case.

This idea is the root assumption of the REHP study, but intercst obviously did not come up with it. Who did? There's a temptation to point to the authors of the earlier conventional methodology studies. But my guess is that, if you asked them how they came up with the idea that valuation doesn't matter, they would say that it is part and parcel of the Efficient Market Hypothesis or some such thing. And who is it who can speak on behalf of the Efficient Market Hypothesis? We could try that fellow who was mentioned in the Wall Street Journal article that was linked from here not too long back. He was supposedly one of the fathers of the Efficient Market Hypothesis. But it was clear enough from the article that he is capable of some slippery footwork when asked direct questions about the meaning of his pronouncements. The article was about a speech he gave that seemed to contradict earlier claims, but the Efficient Market guy denied the contradiction. My sense of things is that each time some aspect of one of these wonderful academic theories is disproven, the "theory" itself is rewritten or reexplained or rethought or whatever so that it can continue to employ academics even though it has not been able to survive real-world tests.

I can't say for sure whether the Efficient Market Hypothesis says that valuations do not matter or not. I don't even know who to ask to find out for sure. What I know for certain is this: Whoever it is who came up with this idea is a fool. And not just any old kind of fool. This individual is a dangerous fool. Because this idea caught fire and was incorporated into hundreds of studies, and it is an idea that is based on nothing but hot air. There is not a sliver of data or common sense backing it up. The whole matter would be comic if it were not for the sad reality that, in the event that stocks perform in the future at all in the way in which they have performed in the past, millions of middle-class investors are going to lose large portions of their life savings as a result of their misplaced confidence in this nonsense idea that became so popular during the 20 years of the hottest bull market in history. If it is hard to find anyone willing to claim credit for the "valuation doesn't matter" insight today, I have a strong sense that it is going to be 10 times harder still to find someone willing to claim credit for it After the Fall.

I'm tempted to take a new look at "A Random Walk Down Wall Street" and see if that book identifies the author of this most unfortunate of investment "insights." Here's my guess as to what we would find if we did so (it's been a long time since I read it so I cannot say for sure). The book probably cites the research indicating that it is extremely difficult to profit from short-term timing. Then it probably shifts almost imperceptively to making suggestions that is is also difficult or impossible to engage in long-term timing (which is something very, very different, according to the best-informed experts and according to the historical data). So we would be right back where we started. We would have one more source encouraging the misperception that long-term timing is impossible. But we would be no closer to determining who it was who fiirst dreamed up this nonsense idea. It's an idea that everyone "knows" is so, but that no one is willing to defend in a reasoned manner. What was that saying about how it's not what you don't know that hurts you most, but what you know for certain that just isn't so?

On top of all that, we have the incredible phenomenon in which we see posters who both know it is so that valuations matter and also know that it is NOT so. I'm talking about people like raddr here. There is a post from his pre-DCM days in which raddr refers to the REHP study as "bogus." That doesn't sound too far off from the claim of William Bernstein that the conventional methodology studies are "highly misleading," does it? Or from the claim of JWR1945 and me that the conventional methodology studies are "analytically invalid." If there is a difference of critical substance between these three ways of describing the REHP study, I fail to see what it is.

There's a big difference in the mind of raddr, however. To raddr, the difference between a "bogus" study and a "highly misleading" or "analytically invalid" study is all the difference in the world. It is a difference so great that raddr has seen fit to ban me and JWR1945 from his site for life before either of us ever put up a single post there. He has not publicly stated that William Bernstein is banned for life as well, but it would seem to logically follow.

Has anyone ever asked raddr what this distinction is that he finds so compelling? No. It is a question that inspires no curiosity in the mind of any poster. Raddr thinks that the study is "bogus" but not "analytically invalid" and that's that. Just as a lot of people think that changes in valuation can never possibly have any effect on long-term stock returns, that it is "100 percent safe" to assume that they will never have any effect, even though every sliver of evidence available to mortal man cuts the other way. None of the people who make this claim are able to offer a sliver of data in support of it. But they know it is so. And you had better "know" it is so too if you know what is good for you.

Not this boy. I know it's a lie. A dangerous lie. A lie that in all likelihood is going to cause losses of hundreds of millions to the middle-class investors taken in by it. I'm on the side of the middle-class investors trying to put together realistic plans for winning financial freedom early in life. That puts me at cross purposes with a puffed-up ego case of a board founder who wrote a study telling people that a take-out number that the data reveals to be highly risky is in fact "100 percent safe." It also puts me at cross purposes with a bunch of people who defend the study, including people who in straight-talking moments put forward straight-talking claims that the study in question is "bogus."

So be it.

What matters is the reality. The reality is that raddr was right the first time. He was once one of our finest posters. He is now a DCM. That's a huge step down, and we have all suffered the effects of his unfortunate choice to take that step.

I am not anti-intercst. I am pro-Retire Early community. There's a difference.
User avatar
Alec
Admin Board Member
Posts: 31
Joined: Wed Sep 10, 2003 4:00 am
Location: Crofton, MD

Post by Alec »

Is there something inherent in Modern Portfolio Theory that requires its proponents to deny that the intrinsic value of a stock purchase varies with changes in valuation levels?
Hocus,

The value of anything is what someone else will pay for it, plain and simple. It doesn't matter what you think the value of something is, it only matters what others think the value is. That is how assets are priced. If you're talking about something like book value as intrinsic value, then the value of a company doesn't change from one second to another if the stock price changes. The expected return changes. The book value of a company is the book value of the company. The thing that changes is how much people will pay for it. As the price drops, the expected return goes up.

Certainly no proponents of MPT or EMH would say that the expected return doesn't change with valuation changes. That is a simple holding period return formula.

Does the expected return going up [valuation levels going down] mean that SWR, or future returns, will go up? Nope. There have been numerous countries where valuation levels going down hasn't led to higher returns. Even though valuation levels in 1990 where higher than they were in 1980, the return in the 1990's was higher for stocks.
It is assumed that anything that might produce a better return will be arbitraged away. It can never persist. But actions that reduce risk will always persist. They will never be arbitraged away.

In this strange world of assumptions, there can never be any such thing as a variation in intrinsic value. It would automatically disappear. On the other hand, the risk premium for investing in stocks will always remain the same. It does not matter that just about everybody knows that stocks are always best in the long run. You still get your bonus, the risk premium.
No, it is assumed that any "free lunch" return will be arbitraged away. For example, a stock trading on two different exchanges at different prices. A trader would just buy at the lower price and sell and the higher price until the profit to doing so was zero, not necessarily until the two prices equaled each other. We do have trading costs after all.

EMH doesn't say anything about investors being rational or irrational, just the level of information that is already priced into securities. Anything producing higher returns certainly can persist for a variety of reasons. For example, the "value premium" story can be because of more risks associated with value stocks, investors irrationality favoring glamor stocks over dogs, the fact that the deep value stocks had the most trading costs associated with them, or a little of all three. Neither of these would violate EMH.

Actions that reduce risk can certainly be arbitraged away, until there is little, zero, or negative return from purchasing those assets that reduce risk the most. Does anyone have an insurance policy with a positive expected return?

Can the "intrinsic value" of a company, or stocks in general, stay the same while the risk premium changes? Yes. People simply bid up the price of a company, or stocks, while the intrinsic value [book value, future earnings, etc.] of the company(ies) doesn't change, b/c people view these assets as less risky investments.

Can the intrinsic value of a company, or stocks, change while the risk premium doesn't? Yes. Book value changes (through something like a terrorist attack destroying assets), or future earnings change while people don't view these assets as any more or less risky.

But again, the intrinsic value of a company doesn't matter, it's what someone else is willing to pay for that company.

EMH & MPT don't say anything about stocks being safe long-term investments, or that the value of assets never changes. I don't know anyone who teaches this stuff that says this. It's the people that don't pay attention is class, that don't understand, or those for whom it is more lucrative to say this, that say this.

- Alec
hocus2004
Moderator
Posts: 752
Joined: Thu Jun 10, 2004 7:33 am

Post by hocus2004 »

"Certainly no proponents of MPT or EMH would say that the expected return doesn't change with valuation changes....

"Does the expected return going up [valuation levels going down] mean that SWR, or future returns, will go up? Nope.


The second statement is in direct contradiction to the first. When the expected return changes, the range of long-term return possibilities changes. The SWR is the take-out number that works in the worst-case returns sequence (the worst that we have seen in the past, but no worse than that). That is the number at the low end of the range of possibilities. When the range of possibilities shifts downward, the number at the low end of the range shifts downward too.
User avatar
Alec
Admin Board Member
Posts: 31
Joined: Wed Sep 10, 2003 4:00 am
Location: Crofton, MD

not really

Post by Alec »

hocus,

Perhaps I wasn't clear [shocker, eh?]. The second statement simply says that just because the future expected return goes up, that doesn't mean that the future realized return will actually go up as well. Prices could always stay depressed for 50+ years. Valuations have crashed in other equity markets and not recovered.

The worst case scenario is that all of your assets become worthless at once, and you're left starving. You cannot get worse than this. This is a real possibility, however very small, no matter what valuation levels are. It is also a real possibility, however small, that valuation levels keep increasing. Just b/c this may have not happened yet doesn't make it any less of a possibility.

If valuation levels decrease, and expected returns increase, then the realized SWR could increase if valuations stabalize. Or the realized SWR could decrease because valuations keep dropping, and drop to zero. OTOH, the realized SWR could increase because valuation levels keep increasing from high valuations. Personally, I don't believe that there is some normalized risk premium, or normalized "intrinsic value", or normalized earnings, dividend, or book ratio that investors will "force prices" back to. There certainly could be, but I'm not willing to bet my $$ on that.

Proponents of MPT would say that when the expected return of risky assets, and their risk premium, go up [when valuations decrease], an investor is getting more reward for each unit of risk. In MPT speak, the line from the risk free asset to the tangency portfolio on the efficient frontier [i.e. Capital Allocation Line] becomes steeper. Thus, given the same level of risk aversion, an investor will allocate more to risky assets when expected return goes up. How much more an investor allocates to risky assets when their expected return, and risk premium, rises depends on how much more utility that person gets out of each increased unit of return per each unit of risk. A really risk averse person would most likely not change their allocation all that much, but a really risk tolerant person probably would.

- Alec
Post Reply