Wednesday, May 20, 2009

Critique: Does Long Term Market Timing Really Work?

Yesterday, Rob of A Rich Life wrote a guest post on Money and Such, arguing the case for long term stock market timing. In effect he argues that when stocks are over valued wise investors should reduce their exposure to that asset class. 

Boiled down to their essence, Rob's arguments are based on the assumption of regression to the mean. Regression to the mean is the assumption that given a large enough sample, things tend to average out. Yes, there maybe someone out there who is 7'3", but it is unreasonable to expect the next guy in your sample to be a giant as well. Similarly, the argument goes, if in a given period of time stocks dramatically over perform or under perform, you can expect that over the long term results will shift in the opposite direction such that the average matches to long term trends. A reasonable enough conclusion, and very much in line with common sense.

So where does this argument break down? Well, first of all, I am not sure that it does. Having said this, there seems to be plenty of empirical evidence to suggest that this common sense argument does not translate well into the real world. Here is a good one: take for example so-called balanced funds that have the freedom to follow Rob's proposed strategy by re-allocating investments between different asset classes. I have seen no evidence that these outperform the market over the long term. These fund managers are well paid and the whole idea behind their full time jobs is that they can identify inflection points in the market and shift asset allocation to take advantage of such changes. They are free to follow long term market timing, so why don't they? Or maybe they think that they are doing exactly that but are simply not successful? I don't know the answer, but this should raise some skepticism. 

The trick, as always, seems to be understanding where the inflection points in the market are, or as Rob would put it, understanding when the market is overpriced. Rob's entire strategy seems to rely on his ability to understand when the market is overpriced, and to do so before others detect the same problem and the market self-corrects. If he can suggest a convincing strategy for achieving this goal, there may be something to his arguments. However, I tend to be skeptical of theories that are built on the assumption that everyone else is subject to mass delusion, hypnosis or hysteria, yet we seem to be the only ones retaining our sanity and keeping our wits. It just doesn't feel right.

I will make one more point that is completely irrelevant to the current discussion, but that is important in the context of evaluating evidence. One of the statements Rob made in his recent guest post did not sit right with me: 
"Given that the studies are silent, I believe that we should default to our common-sense take that timing MUST work". 
That line of reasoning does not pass muster. It boils down to: "the other side doesn't have proof, so our position wins by default". Hmmm. That doesn't work. You don't get to claim victory because the other side does not have proof. You have an equal burden of proof. In fact, if you are trying to go against accepted theory, the burden of proof is mostly yours to bear. You know the saying "extraordinary claims require extraordinary evidence". I am sure that Rob did not mean to use the statement in this fashion, but I wanted to make sure that the ground rules are clear.


While we are talking about investing, here are a few more posts on the subject from other PF bloggers:

My First Million recommends selling your energy stocks... now that sounds like short term market timing to me. He is certainly an aggressive trader. Rob and I are both Index guys. Right Rob?

The Sun's Financial Diary (actually it's a guest writer) is preaching real estate investing. Yeah, I think REITs are good for 5% to 10% of your portfolio, but going gangbusters on the stuff seems like a really bad idea.

Finally, Five Cent Nickel (or is it Fivecentnickel) reviews a new peer-to-peer lending site. You know, I have been meaning to look into p2p lending, but never seem to get around to it. It just seems like a lot of work, for a considerable level of risk and a bit more return. I think the idea has merit in principle, but I think may not quite ready for prime time.


Enjoyed this post? Please consider subscribing to Money and Such by free RSS Feed or by email. You can also follow me on Twitter.

14 comments:

Rob Bennett said...

Thanks for posting the critique, Shadox. As I said re the earlier one, I find it intelligent and fair-minded. I think you did a great job of arguing "the other side" while agreeing with those points that make enough sense to you to win your acceptance.

Rob's arguments are based on the assumption of regression to the mean. That's right. The book that got me on the path that I am on today is Bogle's Common Sense on Mutual Funds. It was from Bogle that I learned about reversion to the mean and it was by exploring the implications of reversion to the mean that I discovered all the rest that I have discovered (for good or for ill).

There seems to be plenty of empirical evidence to suggest that this common sense argument does not translate well into the real world.I don't agree with this statement. I don't see that there is any empirical evidence arguing against the common-sense view of how stock investing must work. There is no question but that many smart people believe that there is empirical evidence that does this. I attribute this to the hundreds of millions of dollars that have been directed to the marketing of the Passive Investing concept. Academic research is often reported to say things that it really does not say. When you check the actual studies, you learn things very much contrary to what we are told by the "experts" employed by The Stock-Selling Industry.

They are free to follow long term market timing, so why don't they? The primary job of a fund manager is to attract dollars to the fund. That is done through marketing, not effective investing. You don't want the fund to do well for investors in the long term. You want investors to see short-term appeal in buying into the fund. Those are very different objectives.

A fund manager who used the historical stock-return data as his guide to what his stock allocation should be during the years of insanely dangerous stock prices would have been fired from his high-pay job. The purchasers of that fund would have been able to retire years sooner. But the fund manager would have been out of work. It is not reasonable to expect fund managers to give up their means of earning a livelihood just to invest in ways that are of greatest benefit to the purchasers of their funds.

We need to educate investors as to what works before we can improve the performance of the fund managers. When investors understand what works in the long term (paying attention to price is critical), the fund managers will be perfectly happy to invest more effectively. Fund managers are followers, not leaders. It is the paying customers who call the tune and most paying customers today do not understand the extent to which valuations affect long-term returns.

Rob Arnott has said that we need to see a "revolution" in our understanding of how stock investing works. That's the right word. We need to go back to the basics and begin again. Passive Investing (sticking to the same stock allocation regardless of price) is exactly what does NOT work. It is hard for any fund manager to go against what his customers want. We need to help the customers learn what works first.Rob

Rob Bennett said...

The trick, as always, seems to be understanding where the inflection points in the market are, or as Rob would put it, understanding when the market is overpriced. I don't see those two as being the same thing, Shadox.

I believe in investing in stocks heavily when the long-term value proposition is strong and not so heavily when the long-term value proposition is poor. I don't think it is possible to identify inflection points and I make no effort to do so. I only look at the long-term value proposition (as revealed by valuation levels). Trying to identify inflection points is short-term timing, which I reject.

Rob's entire strategy seems to rely on his ability to understand when the market is overpriced, and to do so before others detect the same problem and the market self-corrects. I don't try to do anything "before others detect the same problem." The reality is quite to the contrary. I want everyone to know about the problem and to act appropriately. I want to see the market self-correct as quickly as possible.

Why? Because when stocks are overpriced, I go with a low stock allocation and I prefer to be going with a high stock allocation (stocks as a general rule offer the best means to achieving financial freedom early in life).

I view the stock market as a public asset, like the environment. I want to protect that asset and to encourage all others to do so. When the asset is destroyed (great damage has been done during the Passive Investing Era, in my view), we all suffer the consequences. I attribute today's economic crisis to the widespread promotion of Passive Investing in recent decades. If the U.S. economy goes over a cliff, we all lose, Passives and Rationals both.

Some will laugh, but I believe that we have a patriotic duty to warn our fellow investors of the dangers of Passive Investing. I'm being entirely serious when I say that.

I tend to be skeptical of theories that are built on the assumption that everyone else is subject to mass delusion, hypnosis or hysteria, yet we seem to be the only ones retaining our sanity and keeping our wits.Your skepticism is healthy, Shadox. But please understand that what you are describing is precisely what those of us who argue for the importance of valuations are seeing. At the top of the bubble, the historical data told us that the most likely long-term return on stocks was a negative number. At the time, TIPS were paying 4 percent real. To give up 5 percentage points of real return every year for 10 years running for the fun of investing in a high-risk asset class was insane.The people who did this were not insane, of course. Lots of smart and good people did this. It is the investing strategy that encouraged people to do this that is insane.

The problem is that the investing strategy that causes the insanity says that the insanity is an impossibility. So long as you see things through a Passive Investing filter, you can't see the insanity because the core premise of Passive Investing is that the markets are rational. Once you step outside the dominant school of thought, you see it clearly.

To persuade others, you need to get them outside the dominant school of thought. They need to accept that the dominant model is flawed to be able to see the flaws in the dominant model. Believing in Passive Investing is like being caught in The Matrix. It's a closed system of thought.

Rob

Rob Bennett said...

You don't get to claim victory because the other side does not have proof. You have an equal burden of proof. In fact, if you are trying to go against accepted theory, the burden of proof is mostly yours to bear. You know the saying "extraordinary claims require extraordinary evidence". I am sure that Rob did not mean to use the statement in this fashion, but I wanted to make sure that the ground rules are clear.I did mean it the way it sounded, Shadox.

I view the Passive Investing claim (that there is no need to make an allocation change when prices change dramatically) as the extraordinary claim. I believe that the common-sense claim (that allocation changes are required) should be the default claim.

In a practical sense, you are right. Passive Investing is the dominant model. So it is going to take a supreme effort to bring it down from its perch. From the standpoint of logic, however, I think we should always start with a belief that what common sense tells us is probably so. I am willing to go against common sense when strong evidence is presented for a counter-intuitive claim. But when nothing is presented for the counter-intuitive claim, I feel more comfortable going with common sense that going against it.

If he can suggest a convincing strategy for achieving this goal, there may be something to his arguments. If you have any interest in reviewing a Guest Blog Entry on the implementation aspects of the Valuation-Informed Indexing strategy, please let me know and I will shoot some more words your way, Shadox. I don't recall anyone with whom I have engaged in such discussions that I have felt has done a better job of expressing skepticism in a civil and friendly and intelligent way.

I do not want to press on your kindness. I am grateful that you have already done so much to make your readers aware of these ideas. Please don't feel pressed. I just want you to know that if now or at some later time, you have an appetite for further exploration of these ideas, I would be glad to do whatever I could to be responsive to questions from you or from any of your readers.

You have done a fantastic job, in my view. There were moments when you almost had me thinking of going back to a belief in Passive Investing!

Rob

Frank Curmudgeon said...

Not to get too technical, but Rob is actually arguing against regression toward the mean. Regression toward the mean is driven only by random chance. In a nutshell, it is the idea that as you increase your sample size your average observation gets closer to the true average. But this is not because having drawn high observations in the past you are more likely to draw low ones in the future. (E.g. rolling 12 on a pair of dice does not increase the likelihood of rolling 2 later.) It is simply based on the idea that with enough tries the exceptional cases average out into a milder mean. In contrast, Rob believes that previous high returns in the market actually do cause lower returns in the future.

Rob Bennett said...

Rob believes that previous high returns in the market actually do cause lower returns in the future.You're making a fascinating point, Frank.

There is absolutely no question whatsoever but that the words of yours quoted above describe my position. I believe that high returns cause low returns and that low returns cause high returns.

Bogle uses the term "Reversion to the Mean" all the time and I believe (I am questioning this belief a little after reading your words) that he too suggests that there are causative factors. It would to a large extent explain my differences with his thinking if it turned out that that is not the case. If Bogle believes that Reversion to the Mean (I believe that Bogle uses the term "Reversion to the Mean" rather than "Regression to the Mean") is driven by random chance, that would explain a lot. I don't think he does going by my memory of things he has said (and it is possible that it will turn out that he is not entirely clear on this point), but I cannot say that I have researched this particular question.

The is potentially a big deal, in my assessment. The question of whether Reversion to the Mean takes place as the result of random chance or as the result of a causative process could well be the big difference between the Passive investing mindset and the Rational Investing mindset. It's possible that confusion over this point explains a good bit of the confusion that we have seen in the discussion of these ideas.

In any event, Frank is right about what I am saying. I am saying that high returns cause low returns.

Rob

Michael Harr @ Wealth...Uncomplicated said...

@Shadox - A couple of notes on your critique. There is no published study that I've seen to date that supports a superior return from Rob's strategy that I liken to dynamic asset allocation. However, there is a load of evidence that it reduces the risk in a portfolio. I posted a follow on to Rob's guest post that you can find below. Assuming you use 15 p/e10 as the baseline and 8 and 23 as the 'brackets' to identify really low and really high prices, you find that most of the 1990s and 2000s were spent with minimal allocations toward stocks. You would also have sat out the ugly parts of the Great Depression and been underweighted in the 1973-74 bear. One can argue, as Rob does, that this strategy can provide higher returns if properly applied. The example I put forward was a simplified model, but I think it more clearly defines a good chunk of what Rob is trying to get across.

Also, balanced funds were not created to take advantage of market pricing to improve returns, but exist to give shareholders exposure to the market with a fraction of the risk. In addition, by prospectus, they are limited in their abilities to speculate based on market valuations.

At any rate, here's the post:

http://www.wealthuncomplicated.com/wealthuncomplicated/2009/05/dynamic-asset-allocation-a-follow-on-to-rob-bennetts-guest-post.html

BTW, good catch on Digerati...there was a good deal of confusion with MPT and EMH

Shadox said...

Guys = thank you all for participating in a very lively and intelligent discussion.

Rob - we can certainly continue our debate in a future post. Let's take it up in a couple of weeks - I have a pile of other post ideas I would like to explore, so folks don't think Money and Such is only about batting around investment theories.

Frank - your comment is exactly right. I misread the argument Rob was making. Thank you for the correction.

Michael - I am extremely suspicious of P/E as a useful indicator, simply because I know that both the P and the E are subject to a lot of potential manipulation and fluctuations.

I want to make a final point regarding this whole issue. The mark of a useful theory is that (i) it is falsifiable; and (ii) it makes accurate predictions about the future... predicting the past is sort of easy. Finding patterns and models that fit past performance is nice, but does not prove any predictive power.

Maybe we should put this to the test. Let's make some predictions and see if they hold in 5 years. Hopefully, Money and Such will still be here (but with many more readers...)

Rob Bennett said...

we can certainly continue our debate in a future post. Let's take it up in a couple of weeksThat sounds great, Shadox. I look forward to it.

I am extremely suspicious of P/E as a useful indicator, simply because I know that both the P and the E are subject to a lot of potential manipulation and fluctuations. I don't think that we have perfect knowledge re the best valuation indicators. It amazes me how many people still use P/E1 years after it has been shown that P/E10 is superior. But the problem here (in my view) is that Passive Investing has become so dominant that many don't see a need to spend much time studying valuation metrics. As people come to see how important understanding the effect of valuations is to long-term success, I believe that much more energy will be devoted to identifying and perfecting the best possible valuation-assessment tools. That will be a plus for everyone.

The mark of a useful theory is that (i) it is falsifiable; and (ii) it makes accurate predictions about the future... predicting the past is sort of easy. This is a strong point. I wish that people would apply it to Passive Investing. Is Passive Investing falsifiable? I sure cannot see how. It is an entirely closed system. There is no way to disprove it because no proof has ever been offered for the core principle (that there is no need to change your stock allocation in response to big price changes). The core principle is an assumption. There is no way to falsify an assumption.

And please understand that none of the claims made by Passive Investing enthusiasts have ever worked on a going-forward basis. The entire model is the product of backtesting of the historical data (in which some factors were taken into consideration and some were ignored).

The idea that you don't need to be concerned about valuations always becomes extremely popular during wild bull markets and then is forgotten in wild bear markets and then is brought back again (under some new name) in the next out-of-control bull. It always works looking backwards (because those who want to convince themselves that it works look only at those factors that make it look like it works). It never works on a going-forward basis (except for brief stretches of time, like the time-period from 1995 through 2000).

Rob

Rob Bennett said...

Let's make some predictions and see if they hold in 5 years.We can do that, Shadox. But please understand that it is not possible to make precise predictions. And we cannot make statistically significant predictions that only go 5 years out. For statistically significant predictions, you need to go 10 years out. We can say that returns in 10 years will be between x and y and we can assign rough probabilities to different points on the spectrum of possibilities. There's huge value in doing that. But that's all that we can do. We cannot do everything that we would like to be able to do.

There actually has already been a publicly announced test done. Valuation-Informed Indexing is all rooted in the research that Robert Shiller has been doing for years (John Walter Russell has done much more detailed research in recent years, but his research is rooted in Shiller's insights and findings). Shiller predicted in public testimony in 1996 that those going with high stock allocations were going to regret it within 10 years because of the valuations that applied at the time and because of the effect that valuations have on long-term returns.

Passive Investing enthusiasts argue that Shiller has been proven wrong and Rational Investing enthusiasts argue that Shiller has been proven right. How could it be both?

The return that was achieved in 2006 was within the scope of the "prediction" you got from performing a regression analysis on the historical data in 1996. Stocks started performing poorly in 2000. And stocks performed very poorly in 2008. Those facts support the Rational Investors.

The Passive Investors say that the 10-year return in 2006 wasn't all that bad. They are right. The 10-year return was on the lucky side of the range of possibilities identified by performing a regression analysis in 1996.

This highlights the point that it is the premises you start with that determine the conclusions you reach. This prediction was public and the 10-year time-period has expired. But there still is not agreement on whether the prediction "worked" or not. I say that investors would have been smart to follow Shiller's 1996 advice because those invested heavily in stocks were taking huge risk to obtain what were not likely to be very good returns. The Passives don't buy it. They see things through a different filtering mechanism, a Passive Investing filtering mechanism. Everyone sees the same facts but interprets them differently.

Rob

Rob Bennett said...

There is no published study that I've seen to date that supports a superior return from Rob's strategy that I liken to dynamic asset allocation. Your article is a fine piece of work, Michael. It is going to help a lot of people come to a better understanding of this stuff. I am grateful to you for taking the time to put it together.

We are not in agreement that Valuation-Informed Indexing does not necessarily increase returns; I say that it does, at least if implemented in a reasonable way. We are not entirely in agreement re the "study" question, but that's really just a question of what you consider a "study."

John Walter Russell has been doing research on these questions for seven years now. His entire web site (www.Early-Retirement-Planning-Insights.com) is devoted to his original research. John is doing "studies." Some are going to say that they don't count because they are not peer-reviewed. I don't buy it. Our investigations began when I discovered analytical errors in the Old School Safe-Withdrawal-Rate Studies. Numerous big-name experts have since confirmed that John and I were right about the SWR studies and yet not one of those peer-reviewed "studies" has been corrected in the seven years since. So I question what it really signifies for a "study" to pass peer review. My sense at this point is that all that it means is that the "study" follows the dominant academic theory of the day. If the Passive model falls, all of the studies that passed peer-review because they followed the Passive model fall too.

A better test of the worthiness of a study is -- Does it make sense? I have been working these issues daily for seven years and I can say that John's research passes that test. It's not just me. He is the best respected and best liked researcher in the Retire Early and Indexing discussion-board communities. All of his work has been published in real time on the internet. And there are many who would love to be able to find flaws in it. Not one of the thousands who would love to be able to find a flaw in it has yet found anything significant. That's not absolute proof that the research will forever stand up to scrutiny but it is powerful evidence in my eyes. My view is that the fact that no flaws have been found in John's work for seven years counts for more than a "Peer-Review" stamp of approval from people who, however well-intentioned, have a vested interest in not seeing the Passive model fall (any academic who published research of his own under the Passive model is biased re these questions).

We have produced two calculators that permit investors to test whether Valuation-Informed Indexing yields higher returns than Passive Indexing. The Investor's Scenario Surfer lets you go year by year through a 30-year returns sequence that is in accord with the sequences we have seen throughout the historical record and see whether your particular VII strategy beats the various rebalancing strategies or not. VII wins at the end of 30 years about 90 percent of the time, often by large margins. The Investment Strategy Tester lets you run 1,000 tests of any particular VII strategy and see what the best and worst possibilities are. Again, it shows that there is no comparison between VII and Passive Indexing.

You're right that VII reduces risk. What I think you are missing is that in the long term to reduce risk is to increase return. When you reduce risk, you suffer less in the way of losses. When you suffer less in the way of losses, you have more money to invest in stocks at times when they are reasonably priced. Having more money in stocks translates into enjoying a greater compounding of returns over the years. Given the passage of enough time that puts you far ahead.

Rob

Rob Bennett said...

Rational Investing (changing your allocation in response to big price changes) permits you both to reduce risk AND to increase returns. The Passives will tell you that this is impossible. The Passives teach that greater returns come from taking on more risk. You must choose EITHER lower risk OR higher return but you can never have both. The Rationals dispute this. The Rationals say that reduced risk and higher returns come in a package -- BOTH benefits go to those willing to take valuations into consideration when setting their stock allocation.

How can it be?

The resolution of the mystery lies in consideration of a point made by Shadox in the critique. Shadox said that he finds it hard to believe that so many investors could have been as wrong as I am saying they would have had to have been for prices to get where they were during the crazy years. I agree that my claim is hard to swallow. But, if I am right that millions were indeed investing irrationally (that is indeed my claim), then I don't think it is at all hard to accept that it would be possible for those investing rationally to obtain both higher returns and less risk as a result of doing so.

The root question in all of this is -- Is investing primarily a rational endeavor (that's the root premise of the Passive model) or is investing primarily an emotional endeavor (that's the root premise of the Rational model). If investing is primarily an emotional endeavor, then you cannot trust 90 percent of what you are told about investing during a time when prices are insane. Emotional investors demand "findings" that support their emotional biases. I am saying that 90 percent of the "research" done during the past 30 years was done for primarily emotional reasons -- the researchers wanted to find that Passive works and so they (unintentionally) spun the numbers to formulate "studies" that told the message that they wanted to hear.

How do I know when I can trust the research being generated? I look at the P/E10 level. The P/E10 level tells us how emotional investing has become at any given time. We went in recent years to P/E10 levels far higher than any we have seen in history. That tells us that most of the investment research generated in recent years is the least trustworthy investment research ever produced.

Passives don't see this. Because Passives don't think that valuations matter much. You cannot see something that you are not willing to look at.

I see it because I think that valuations matter. The evidence is there for those who want to see it. But once you buy into the Passive model it disappears. To buy into the Passive model is to conclude that valuations don't matter. To buy into the Passive model is to lose the ability to see why the Passive model does not work.

The Passive model is a closed system. The only way out is to accept (at least for a moment) that valuations matter and then see how different the world looks when you see through eyes that appreciate the importance of valuations.

Rob

Market Timing said...

Hey this is a good article thank you for the read. I stumbled across this good market timing site myself that I have been using the past year. It's great I love market timing.

Anonymous said...

Bennett provided the world with a gift of his "logic" in the following tight little package of words:

"I attribute this to the hundreds of millions of dollars that have been directed to the marketing of the Passive Investing concept. Academic research is often reported to say things that it really does not say. When you check the actual studies, you learn things very much contrary to what we are told by the "experts" employed by The Stock-Selling Industry."

So, according to Bennett:
1) Hundreds of millions of dollars were spent by the "Stock Selling Industry" selling the concept specifically of "Passive Investing"
2) We should trust no one's claim about such things until we see the source documents ourselves!

Fine! I will take the rope Rob has played out for us here, and return it to him, fashioned as a cravat, fit for his wearing:

Rob: Provide a reputable primary source, containing data and specifics, to support your oft-made claim that the "Stock Selling Industry" has spent "hundreds of millions of dollars" on marketing "Passive Investing."

I'll be waiting to be educated.

Market Timing Signal said...

Whether or not market timing signals work or are proven is really a matter of application and time frame involved. Most professionals use a form of market timing everytime they enter a trade... yet when timeframes are longer people seem to have issues with it.

Its a great debate that will probably never end.