Set forth below is the text of a comment that I recently posted to the discussion thread for another blog entry at this site:
Uh, Oh. The RESEARCH says that VII and CAPE don’t work. Looks like that $500 million windfall just went out the window
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3142575
My view of the paper is that it is making a valid and important point. The paper does not conclude that long-term timing (price discipline) doesn’t work. I think the statistical point needs to be considered. Personally (please understand that I do not possess expertise in the statistics area), I lean toward saying that the statistical point being made is a good one so long as it is not overstated. I don’t think the paper overstates the point in a direct and clear way. But I think that overstatements are suggested. That’s my one big criticism of the paper. I don’t think it properly described the context in which these questions are being considered and thus might cause Buy-and-Holders to be misled into having more confidence in their strategy than the strategy merits.
The authors are saying that there are not enough independent data points to draw grand conclusions about long-term return predictability. I half agree and half disagree. Returns do not fall in the pattern of a random walk. It is my strongly held view that that has been demonstrated beyond any reasonable doubt whatsoever. If the authors are trying to say that returns play out in the form of a random walk both in the short-term and in the long-term, I reject that out of hand. I did not see a statement that clear in my quick read of the paper. If they are indeed not saying that, then I am more or less okay with what they are saying.
Stock returns play out in long bull/bear cycles of something in the neighborhood of 37 years each on average (we have data for 148 years and we are near the end of the fourth cycle in the record). We have 148 years of data supporting that general claim, the entire historical record. But when we try to get more specific, we get into trouble. Shiller predicted that returns would be negative in the ten years from 1986 forward — that prediction failed. I predicted that we would see a second price crash by the end of President Obama’s second term — that prediction failed. Both of those predictions were supported by the historical data. I believe that part of the problem with such predictions is the point being made in the paper — we don’t have a lot of independent data points.
There are only four bull/bear cycles in the historical record. So, when you try to identify when the current cycle is going to end, you don’t have many cases to look at. You can say “no cycle has lasted longer than x number of years.” But does that really show that the current cycle will not be the first exception to the general rule? Say that you studied four alcoholics and warned a friend of yours that in every one of the four cases the alcoholic ended up dead or in prison within 30 years of becoming an alcoholic and that thus he had better shape up because he only had five years to go before hitting the 30-year mark. Is it possible that your friend would remain alive and out of prison for more than another five years? It’s possible. You only studied four cases. Each case is different. If there were zero alcoholics who remained alive and out of prison at the end of five years out of a group of 10,000 alcoholics, you would have a very strong case re your five-year prediction. But, when you only have four cases, you have to be very cautious in your predictions. That’s the important point that I think is legitimate here.
The possible conclusion that I am uneasy about is any suggestion that alcoholism is not a very serious problem or that failing to exercise price discipline when buying stocks is not a very serious problem. I think that our failure as a society to exercise price discipline when buying stocks is killing us. I think that is is in the process of causing millions of failed retirements. I think that it was the primary cause of the 2008 economic crisis. I think that it has destroyed dozens of once top-notch investing boards and blogs. I think that it has played a big role in bringing on the political frictions on both the left and the right that have become characteristic of our political life in recent years. So I don’t at all go along with any suggestion along those lines.
The bull/bear cycle has applied for the entire history of the market. If the market were efficient (that is, if price changes were caused by economic developments. as the Buy-and-Holders posit), prices should play out in the form of a random walk, there should be no bull/bear cycle. But there is! And there has been for the entire history of the market! Stock returns have never played out in the manner in which they must always play out if the core belief of the Buy-and-Holders – market efficiency — is valid.
Short-term returns cannot be predicted at all. We all agree on that. Long-term returns cannot be predicted with precision. We all also agree on that. But long-term returns CAN be predicted; a range of possible returns can be identified and rough probabilities can be assigned to all points on the range. All of the evidence available to us shows that. There are legitimate questions as to the degree to which investors should be using predictions to influence their stock allocation decisions. We don’t know it all.I can go along with that much. But the Buy-and-Holders say something more than that — they say that it is perfectly okay for investors to fail to make any allocation changes whatsoever in response to valuation shifts. I view that one as irresponsible in the extreme.
The authors at one point refer to “the null of no predictability.” I believe that that phrase points to the core dispute that I have with Buy-and-Holders. The Buy-and-Holders seem to ASSUME that there is no need to engage in price discipline when buying stocks and then set up very high hurdles that they insist that those arguing in favor of price discipline (long-term timing) must leap. But why that assumption? In every market other than the stock market, price discipline is essential. It is price discipline that makes markets work. Without price discipline, we should expect markets to collapse. And, sure enough, on every occasion in history in which Buy-and-Hold has become popular, the stock market has collapsed. Could it be because the core assumption of the Buy-and-Holders is false, that price discipline is every bit as essential when buying stocks as it is when buying anything else? That’s what I believe.
In some scientific investigations, people test things with mice. If we were testing with mice and we had significant evidence that long-term timing always works and is always required but a limited number of independent data points, I would argue that we should bring in more mice and run more tests — we should try to get those data points up to where they need to be. But that’s not an option in this case. We just have to watch things play out. We will have more independent data points in time. But, given that one bull/bear cycle can take 37 years or even longer to play out, it is going to be a long time before we have as many independent data points as we would like to have. None of us engaged in these conversations is even going to be around anymore when that magical day arrives.
What to do?
I think we need to listen to people coming at these issues from all sides. I think we need to open up every investing discussion board and blog to honest posting on every possible question.
I think we all need to make more of an effort to become aware of the assumptions (assumptions are something different than evidence) driving our beliefs about how stock investing works. Wade Pfau spent months trying to find a single peer-reviewed study backing up the core Buy-and-Hold belief that long-term timing does not work. He couldn’t find one. That’s the “null” that the authors of this paper are referring to. The Buy-and-Holders ASSUMED that long-term timing doesn’t work and the authors of this paper are saying that there are not enough independent data points to entirely disprove that belief statistically. The other side of the story is that there is zero support in the historical record for the Buy-and-Hold belief. It is an assumption. There never has been any evidence for it.
Should Buy-and-Hold be the “null”? I sure don’t think so. Long-term timing is price discipline. Make Buy-and-Hold the null and you are ruling out the use of price discipline in the stock market. Huh? What the f? Where is that going to take you? I think it would be fair to predict that it is going to take you to a very bad place. And of course the historical record shows that indeed it always has. The widespread popularity of Buy-and-Hold ALWAYS causes economic crises. There is not one exception in the (limited, to be sure) historical record.
My null is that price discipline (long-term timing) always works and is always required. The entire historical record supports that null. I think that I can go along with a claim that the historical record does not alone PROVE that null. I think that that is the claim that this paper is making and my tentative sense is that that is probably a good point to make. But I worry that the claim will be misinterpreted by many Buy-and-Holders. I believe that there are Buy-and-Holders who will conclude that it is okay not to practice price discipline when buying stocks. It is my belief that nothing could be further from the truth. I believe that IN PART because of what we have learned in recent decades from investigations of the historical data, which indeed are not 100 percent conclusive by themselves. But I also believe it because the historical data just supports what we would all know from common sense if we were thinking clearly. Of course price discipline is required! It is required in every market that ever existed! Markets cannot function without price discipline! Why would anyone think it would be any different in the stock market from how it is in every other market that ever existed?
Again, I am grateful to you for bringing the paper to my attention. I will plan to write a column on it in the not-too-distant future.
And, as always, my best and warmest wishes to you and yours, my dear Goon friend.
Go Phils! It’s Nola day. The historical record shows that this one will with certainly be another win for the good guys!
Rob
Anonymous says
“The paper does not conclude that long-term timing (price discipline) doesn’t work.”
Actually, it does.
Rob says
Okay, Anonymous.
My advice to those reading these words either now or at a later time is to read the article for themselves and to develop their own take as to what is being said.
My best and warmest wishes to you.
Rob
Anonymous says
“The authors at one point refer to “the null of no predictability.”…My null is that price discipline (long-term timing) always works and is always required. The entire historical record supports that null.”
Simply breathtaking ignorance. “Null” does not mean “premise” or “position”. In the paper’s context it is short for “null hypothesis”, meaning “no relationship”. As in CAPE has no statistically proven relationship to stock market returns.
More generally, null means no value at all. As in “Rob Bennett’s reading comprehension skills are null.”
Rob says
You go right to the heart of things with this comment, Anonymous.
You say: “In the paper’s context, it is short for ‘null hypothesis,” meaning no relationship.” We are in complete agreement re this point.
The null hypothesis behind Buy-and-Hold was discredited when Shiller published his “revolutionary” (Shiller’s word), Nobel-prize-winning research showing that valuations affect long-term returns. If valuations affect long-term returns, there IS a relationship. Price discipline (long-term timing) always works. And price discipline is always REQUIRED for investors seeking to keep their risk profiles roughly constant over time.
Wade Pfau spent months searching the literature seeking to identify a single peer-reviewed study showing that long-term timing might not work or might not be required. He came up empty-handed. There is nothing. Not one study. Nada. Zilch.
Buy-and-Hold is a marketing gimmick. Nothing more, nothing less. There is zero support for it in the peer-reviewed research or in the historical return data on which the research is based. Buy-and-Hold is a MISTAKE. Before 1981, people didn’t know. So I think it would be fair to say that Buy-and-Hold was an honest mistake in earlier days. But since 1981, anyone who follows the peer-reviewed research knows the story. 100 percent of the research supports Valuation-Informed Indexing, 0 percent supports Buy-and-Hold.
Are most Buy-and-Holders suffering from cognitive dissonance? It sure seems so to me. Is that an important reality here? It sure seems so to me. Do we still need to find a way to get accurate and informed investing advice out to millions of middle-class investors? It sure seems so to me.
I hope that helps a small bit.
Reading-Comprehension-Challenged Rob
Anonymous says
Tediously predictable, you double down on the stupid with the same insane gibberish. Face it Rob. It’s not working anymore. Your act has played out.
Rob says
Um — That makes a lot of sense, Anonymous.
Truly outstanding!!!
Take good care, man.
Tediously Predictable Rob