"I would like to see some more CONCRETE rules to follow based on all this research. "
A question along these lines came up on the "Yahoo Finance Quiz" at the Early Retirement Forum. Here is a link to page 14 of that thread, on which JWR1945 provided his response.
http://early-retirement.org/cgi-bin/yab ... ;start=195
Here's the text of his response.
Logged
JWR1945
Recycles dryer sheets
Posts: 199
Re: Yahoo "Finance Quiz"
« Reply #202 on: Jun 6th, 2004, 6:07pm » Quote Modify
------------------------------------------------------------------------
For Cut-Throat who wants a very definite and very simple answer.
I have no numbers for the DOW. This is what I have for the S&P500 index.
I quote myself from TIPS Switching Surveys from Tue Mar 09, 2004 at 1:28 pm CDT.
http://nofeeboards.com/boards/viewtopic.php?t=2208
These are the optimized results for switching allocations with stocks and TIPS with a 2% (real) interest rate.
..
Optimization turned out to be much more difficult than I had anticipated and my progress was awkward. I am satisfied that I have come reasonably close to the true optimal values. I am satisfied that the highest Historical Database Rate of 5.2% is a consistent ceiling using simple P/E10 thresholds. I am satisfied that the best reason for using additional thresholds and allocations is to reduce the sensitivity of results to minor differences from the optimal values. That is, I expect that the optimal values in the future will be similar to those extracted from historical data, but not identical. My goal is to reduce the effects of errors because we are depending upon historical data.
In the summary, I wrote:
The simplest approach uses two thresholds and one adjustable stock allocation (for a total of three). The best choices are P/E10 thresholds of 11 and 24 and stock allocations of 100%-30%-0%, respectively.
When there is more than one, the most important thresholds are the lowest thresholds.
The best choice with three thresholds and two adjustable allocations (for a total of four) is to use P/E10 thresholds of 9-13-24 ..Using P/E10 thresholds of 9-12-24 is another a good choice.
The best choice with four thresholds and three adjustable stock allocations (for a total of five) are 9-12-21-24 or 9-13-21-24 (with no preference) with 100%-50%-30%-20%-0%, respectively. There is relatively little sensitivity to changing the lowest adjustable allocation or the lower middle P/E10 threshold.
Here are the latest numbers from Yale Professor Robert J. Shiller's web site.
http://www.econ.yale.edu/~shiller/data.htm
Date: November 2003
S&P500 price was 1054.87
P/E10 was 25.898702
E10 is the ten-year trailing average of earnings. P is the current index value or price.
Assuming that E10 has not changed significantly from last November, for P/E10 to equal 24.0 would require that the S&P500 index value fall to 977.53.
The current value of the S&P500 index is 1122.50 and the current value of P/E10 (assuming that E10 has not changed significantly since last November) is close to 27.559.
Therefore, when the S&P500 falls close to 977.53, you may reasonably allocate 20% to 30% of your portfolio to stocks (or, more precisely, to an S&P500 index fund with 0.20% expenses).
This is my best answer for today within the constraints of our existing calculators.
Have fun.
John R.
I am reluctant to hand out investment advice to people. I don't think that I am qualified to do so. I don't possess any degress in the field or anything like that.
What I am trying to do is to provide people a tool so that they can figure out what to do for themselves. SWR analysis, properly done, does not just provide you with a take-out number that worked in the past. The really wonderful thing that it does is to provide you a means of COMPARING different asset classes. That is the primary purpose to which I put the tool.
JWR1945 put up a post here yesterday ("SWR as a Tool") that did an outstanding job of explaining how I used the tool back in 1996, when I first developed it. He put that together by himself, just going by what he knows about what I did from posts of mine he has read and from our e-mail conversations. He sent it to me to check it, and I made two changes, but 90 percent of that he came up with just from reading posts that are already out there.
People have not focused on that post. I wish they would. It's a short post, but one with an awful lot of meat. If you get what is being said in that post, you get the idea of the tool. My sense is that most people today do not get the idea of the tool. That's unfortunate.
It is not because we have not explained what the tool does. The problem is that people are starting out with preconceptions that they just have not been able to get out of their heads. The key purpose of the tool is to COMPARE asset classes in an informed way.
If you believe in the "Stocks for the Long Run" paradigm, you don't need to compare asset classes. You buy stocks. Allocation decisions are easy as pie. They are easy, but they are also poorly informed. The historical data says that, if you make your asset allocations without taking valuation levels into account, there is a good chance you are going to get yourself killed. Those blindly following the "Stocks for the Long Run" paradigm are playing Roussian Roulette with a loaded pistol. They just don't know it yet.
All aspiring early retirees need a tool by which to compare the mertits of various investment classes at various times and for various purposes. To make valid comparisons, you need to form some expectation as to how the various investment classes are likely to perform. The best guide to how they will perform in the future is data showing how they performed in the past. Our tool looks to the historical data to inform you as to how your investments choices are likely to work out.
The trouble we are having has nothing to do with the new tool. The trouble is a consequence of the confidence that people have come to place in the old tool. If people had never heard of these conventional studies, I think that just about everybody would get this. The things that hurt you most are not the things you don't know, but the things that you know for absolutely certain that happen not to be true. Lots of people know with absolute certainty that the SWR for stocks is always 4 percent, and that number is 2 full percentage points off the mark of what the historical data tells us.
You need to get back to zero. You need to ask yourself, why is it that safe withdrawal rate studies were developed in the first place? The core purpose is to tell you what withdrawal rate is safe. The 4 percent number is a high-risk number. There is no informed researcher who denies this. Does it make sense to refer to a high-risk number as a "100 percent safe" number? It makes no sense.
The statistical analysis that JWR1945 presents in the material I quoted above suggests that you should have a zero percent stock allocation today. I personally am reluctant to advise that. It is possible that stocks will go up 30 percent in the next 12 months, and that that will cause you to regret getting out of stocks. You might get back in just at the worst possible moment to do so.
I think it makes sense to always have some skin in the stock game. My circumstances do not permit it at this time. But I tend to think that most others should probably always have 20 percent or 30 percent in stocks.
But not 74 percent! That is a suicidal allocation for the average aspiring early retiree. We regularly inform newcomers to these boards that a 74 percent stock allocation combined with a 4 percent take-out number is "100 percent safe." That advice is so dangerous that I can't think of a word to describe it.
If you are going to tell people what the historical data says is safe, you have to look at what the historical data says re what is safe. I don't see any way around it. If you look at what the historical data says, you won't come up with a number anywhere remotely in the ballpart of the numbers you get from those conventional methodology studies.
Here's an interesting fact to know and tell. There is NOT ONE study of SWRs that has come out since we reached the valuation levels of the late 1990s that shows an SWR for stocks of anything close to 4 percent. There is not one. Every single conventional methodology study was done BEFORE we reached those valuation levels.
Will there ever again be a study showing that the SWR is 4 percent? I doubt it. Why would a researcher want to put out a study giving people the wrong number? What would be the point? I don't think that we will ever again see a new study saying that the SWR is always 4 percent.
So why not just go ahead and do what the whole world is going to do sometime within the next few years and acknowledge that we know for a "mathematical certainty" (Bernstein's phrase) that changes in valuation levels affect long-term returns? If we take that common sense step, it becomes possible for all of us to talk in an honest and informed way about what the historical data says re SWRs. I think that taking that step would be huge.
I understand that there are people who do not like long posts. The truth of the matter is that I have always posted long, even back in the days when I was one of the most loved posters on the Motley Fool site. What people who don't like long posts should do is what you did here, Ben. Put up short posts! That will show old hocus, that will put him in his place. Bury this board in short posts and no one will even notice his lengthy ones anymore. Please feel free to take that as a challenge.
It's true that the board has been too theoretical and complex in the past. The problem has been that JWR1945 has been the dominant poster. His stuff is wonderful beyond words, but, like all of us, he has things he is great at and things he just doesn't do. JWR1945 doesn't do hocus, and hocus doesn't do JWR1945, and hocus doesn't do FoolMeOnce, FoolMeOnce doesn't do Wanderer, and on and on.
We need to get back to the idea of working as a community. When you want long enthusiastic stuff, you dial up hocus. When you want a numbers guy, JWR1945 is the man. When you've got real estate on your mind, it's FoolMeOnce to the rescue. Petey reads lots of books, and is happy to tell you about what he has learned. Wanderer possesses insights on this ex-pat living stuff that few others do.
We all have the same goal, or we all should. Somewhere along the line the idea caught on that we must destroy this new idea that hocus has turned up at all costs, and that has caused a lots of riffs. You can't kill the historical data. If I didn't tell you what it said, it would still say the same thing. You must come to understand that studies that pop up numbers far off from what the historical data says cannot be right. When we get past that point, all the other stuff falls into place again.
Anyway, I hope that the material I quoted from JWR1945 gives you a little bit of an answer to your question. Perhaps some others will expand on that answer or challenge it or whatever.
Here's the key thing that I am trying to communicate. This is not a hocus board. I hate it when people say that. I started the board and I moderate it. But the board is run for the benefit fo the community. For the board to serve the community, the community has to get involved and see that the board does what they want it to do. Change the ratio of short posts to long posts by putting up short posts. Change the theoretical and complex tone by employing a practical and simple tone yourself. And so on.
Board boycotts are stupid. Why the heck would anyone want to put a board out of business? Boards are a free information resource. The idea of putting any of your human energies into trying to kill them off is just plain stupid. Community, stop that stuff! Stop being so stupid!
Make the board what you want it to be. If you find it to be too much about hocus, there is an easy solution. Put up non-hocus stuff, lots of it. Everyone keeps thinking that they can make people stop talking about hocus by demanding that people not talk about hocus. It doesn't work that way. Try this. Don't think about flowers! You're thinking about flowers, are you not? Come on.
We need to de-hocusfy the SWR discussions. That's part of the Normalization process, dehocusification.
That means we need to stop the silliness. When someone uses the word "troll," they are not Normalizing, are they? They are flashing a big red light saying "this issue is different than all the others, this one board members may not talk about, they will get smeared if they talk about it."
Why smear people? If the SWR issue is important enough to smear people over, then it must be important enough to talk through as well. If protecting the conventional methodology from scrutiny justifies smear campaigns, then surely the subject is important enough that we should want to get the number right. Why is it then that few will speak out against the smear campaigns, yet when I suggest that we should get the number right, people say "oh, does it really matter so much, it's only people's retirements at stake afterall?"
I'll stop. I think that we have a tiger by the tale here. People get very emotional about SWRs. The reason is that in the back of their minds they do not believe that the assumptions of the conventional methodology studies could possibly be true. They know those numbers are wrong, but they don't want to know it.
All the angst goes away when we just reach down to the bottom and come up with the truth. The old number is wrong. We all know it. Since it's wrong, we should acknowledge it. Acknowleding the truth about SWRs is the turning point in the discussions. That part is hard, for several different reasons. After we do that one hard thing, the rest off the ride is downhill.
Anyway, thanks for the short post, Ben. Keep 'em coming. You are going to have to come forward with a whole bunch of short ones to counter the effect of my long ones, you know.
I want to see you knock me out of the box. Ker-Pow!