4% Rule

I think SalaryGuru is right about the overlapping periods - they are correlated.

If you know the return for, say, 1971 - 2001, you can calculate the total return by (1 +r 1971)*(1 + r1972) * ... * (1 + r2000) to get the 30 year return for the years 1971-2000. To compute the return for the next sliding period 1972 - 2001, you only need to divide by (1 + r1971) and multiply by (1 + r2001). The intervening 29 years are common so the two results are related.

[newellcr]

Hello Folks,

I've been following the thread but haven't been able to read every post until today. Would you feel better if it were called the "historically" SWR? Or the "historical condition" SWR? The SWR is just a tool for planning. Planning and safety have little to do with each other.

It's kind of interesting to me that many of you think that the application of the SWR is optimistic. I think it is a very pessimistic tool. It would be more useful to me to have a tool that allowed me to see what history did to a range of withdrawals. For instance, I absolutely need x dollars, but when I can, I'd like to take y dollars. In the down market, the calculator would register the smaller withdrawal and in the up market, the calculator would give you the larger withdrawal.

Interesting thread.

Cheers,

Chris

newellcr,

Quote:

The SWR is just a tool for planning. *Planning and safety have little to do with each other.

I am with you on this one. I see very complex formulas being applied in this search for SWR, but I don't think we're getting a better answer.

I think the FIRE Calc program that applies your numbers against the Stock Market History is about the best planning that you can do.

This is a lot like forecasting the weather. We can say that it will be colder in January than it will be in July on Average. But you have no way of knowing whether your house will get hit with an F5 Tornado!

I agree that the ability to vary withdrawal rates based on the performance of the market, would be a real eye opener. It would show you how pessimistic the FIRE Calc tool really is. In reality when the market is doing poorly, we pull *in our spending horns and when the market is doing well, we spend more money.

I would suggest having 3 inputs. Amount to withdraw after the market had X% gain. And another amount to withdraw when the market has less than this gain.

This would be a good starting point. What do you think Dory?

[bongo2]

"I'm not aware of any mathematical reason that the number of "non-overlapping" periods is important to validity of the analysis."

If you are trying to use standard statistical measures you need independent trials. If you consider the last 130 years of overlapping 30 year time periods to be 100 samples, you are overstating the significance of your results by a wide margin. Try it yourself if you like. Flip a coin 130 times like you suggested, and then write down the 100 overlapping 30 "year" periods. Then test the goodness of fit to the true binomial(0.5,30) distribution. You should find that it isn't even close.

I would also disagree with salaryguru's point. If you think that the 4% rule means that it is completely impossible to exhaust your nest egg using a 4% withdrawal then I guess the implication is that you can pick your highest point. However, I think the 4% rule should be interpreted as a reasonably conservative guideline supported by a reasonable amount of evidence. However, we are not looking at probabilities for the last three years -- we know they were terrible. You have to understand that there is always a chance that your portfolio will be depleted, and once it has fallen in value by 30-50%, then that chance is going to be much greater than it was when you started. In the 130 years of history the stock market has always rallied strongly off of 30%+ drops -- are you going to count on a rally like that this time?

Bernstein makes an important point, that even the best and most powerful civilizations have a less than 90% survival rate over 30 year timeframes. No portfolio is going to withstand hyperinflation, or government collapse, so any survival rate over 90% is meaningless.

[sgeeeee]

Quote:

If you are trying to use standard statistical measures you need independent trials. *If you consider the last 130 years of overlapping 30 year time periods to be 100 samples, you are overstating the significance of your results by a wide margin. *Try it yourself if you like. *Flip a coin 130 times like you suggested, and then write down the 100 overlapping 30 "year" periods. *Then test the goodness of fit to the true binomial(0.5,30) distribution. *You should find that it isn't even close.

That would be true if you were trying to fit the data to some predictive curve as you have indicated. But that's not the point of the historical analysis. In this situation, you are looking for a high probability worst case of successive years. This is very different analysis than predictive curve fitting. It is true that nest egg cash flow analysis of years 1 - 30 are going to share much with nest egg cash flow from years 2 - 31 . . . or years 10 - 40. But the worst case for nest egg survival may fall in years 5 - 36 and if you only use non-overlapping intervals and don't consider all the possible 30 contiguous year intervals, you will dramatically overestimate the safe withdrawal rate.If you don't believe this, or understand it, do the exercise yourself. Try developing a safe withdrawal rate considering only non-overlapping intervals and see what answer you get. It does not make any sense mathematically to require non-overlapping intervals to do this worst case analysis.

Quote:

I would also disagree with salaryguru's point. *If you think that the 4% rule means that it is completely impossible to exhaust your nest egg using a 4% withdrawal then I guess the implication is that you can pick your highest point. *However, I think the 4% rule should be interpreted as a reasonably conservative guideline supported by a reasonable amount of evidence. *However, we are not looking at probabilities for the last three years -- we know they were terrible. *You have to understand that there is always a chance that your portfolio will be depleted, and once it has fallen in value by 30-50%, then that chance is going to be much greater than it was when you started. *In the 130 years of history the stock market has always rallied strongly off of 30%+ drops -- are you going to count on a rally like that this time? Bernstein makes an important point, that even the best and most powerful civilizations have a less than 90% survival rate over 30 year timeframes. *No portfolio is going to withstand hyperinflation, or government collapse, so any survival rate over 90% is meaningless.

This happens every time a discussion of the historical analysis that results in the 4% guidelines begins among ERs. Please read my posts thoroughly before you decide to condemn me and say you disagree with my conclusions. Past performance is no guarantee of future results. Not ever! That statement alone is enough to condemn any historical based analysis. There are plenty of reasons to question the validity of the 4% rule conclusions based on that statement alone. The government could colapse tomorrow and all your assets siezed. Tomorrow may bear no resemblance to yesterday. If that happens, then the historical analysis will not be valid. . . Do these statements ring a bell with anyone?

So . . . if there are valid reasons to believe a historical simulator that considers all key financial performance since 1871 has merit, then you should understand how the 4% figure was arrived at. It was arrived at by searching for the worst case among all the possible 30 year intervals in those years.

1) If you assume that the worst case since 1871 is a cake walk compared to times to come, then you shouldn't believe the 4% rule. Keep working and saving till you feel comfortable.

2) If you assume that the future for retireees who retired in 2000 is not likely to be any worse than the worst cases experienced by retirees in 1929 or 1966 or any other year since 1871, then you should be comfortable retiring based on 4% withdrawal rate of your 2000 nest egg. You can believe that if it were safe to retire in 2000 (and you haven't spent your nest egg) it is still safe to retire today -- regardless of the relative value of your nest egg at those two times. Losses incurred since 2000 are accounted for in the analysis under these assumptions. And just to be clear . . . all of the other assumptions about equity/fixed balances and annual rebalancing have to hold true.

I want to be clear about this too: I'm not giving anyone any advice. If you don't feel comfortable with a 4% initial withdrawal rate based on your 2000 nest egg, then don't do it.

[earlyout]

After reading all these posts I justed wanted to thank everyone for their thoughts and input. Sometimes the best understanding comes from hearing differing points of views.

It is useful to question our own plans and assumptions and understand them or the rationale behind them. My goal is to understand the best and weakest points about any withdrawal system/plan.

To this point, there is a difference between strategy and outcome. It is true that a superior outcome may come from an inferior plan (e.g. russian roulete) and that an inferior outcome may come from a superior strategy.
The way to play though is with an understandable, reasonable strategy. Gurantees don't exist, but good strategies are the way to play it.


Firecalc is one component in assimulating a strategy.

best wishes to all

thanks

earlyout

[Telly]

Whew! I was away for a bit, quite a few posts on this thread!

I would like to thank mikey for his clarification, way up the page!

[bongo2]

salaryguru,

I'm not suggesting that you should only look at non-overlapping intervals. In fact, especially since we're using an inter-year high, I would suggest that you look at the 1200 overlapping 360 month intervals as well to get a feel for how much the inter-year volatility has made a difference. I'm only suggesting that among all the criticisms of using the 4% rule today, one completely valid criticism is that the last 130 years provides very little data about 30-40 year time horizons.

Perhaps I misunderstood your post, but while the 4% rule may be interpreted as merely a historical statement, I don't see why anyone would talk about it unless they think it is some sort of prediction about the safety or intelligence of decisions today.

"You can believe that if it were safe to retire in 2000 (and you haven't spent your nest egg) it is still safe to retire today -- regardless of the relative value of your nest egg at those two times." This is where I disagree. You have more information today than you had in 2000, and what was a prudent decision then may be no longer. In January of 2000 the probability of the 2000 to 2030 period being worse than any 30 year calendar year interval in the last 130 years was quite low. Now that we have had three years of losses that have only been seen twice before in those 130 years it is far more likely that you a facing that "worst case."

What I meant by my comment about government collapse was not that you should be more conservative, but rather that you should be less. Why worry about the last 5% probability that your money will last 30 years, when it's not that unlikely that you'll have much bigger problems by the end of that time?

[hocus]

<b>Mikey</b>: To me, a useful way to think about this problem is to focus on the sustainable cash generated by the portfolio....Would the amount of cash you could take out have anything to do with what valuation was put on the business?

<b>Newellcr:</b>*It would be more useful to me to have a tool that allowed me to see what history did to a range of withdrawals. *

<b>Cut Throat</b>:I agree that the ability to vary withdrawal rates based on the performance of the market, would be a real eye opener.

I formed a board a few months ago at the NoFeeBoards.com site that you guys might want to check out. It's called "SWR Research Group." Here's a link to the page listing the various threads.

http://nofeeboards.com/boards/viewforum.php?f=10

I am the poster who started the Great SWR Debate at the Motley Fool's REHP board back on May 13, 2002. That generated a lot of powerful insights, as well as some ugly posting. The idea for the new board is to continue to explore the insights that have been developed while shutting down the word game posts and ridicule posts that have marred SWR discussions in the past.

Right now, it's just me and a poster named <b>JWR1945</b> who post regularly at the new board. So we would love to have some new posters. Any members of the community here who have thoughts to share on SWRs or who have questions re the research that has been posted at the new board should feel welcome to join in on our discussions.

My recommendation would be to start by reading the "About This Board" post. That explains in general terms what the new board is about. We have big ambitions--I say in that post that we hope to change the world by changing the way that SWRs are determined. It's a bold claim, but I think you will agree after taking a look at some of the threads that have gone up in just a few months that we have built a solid foundation for doing just that. We have already come up with some powerful stuff, and the implications of some of the work we have done stretch forward in many exciting directions.

<b>JWR1945</b> explores the question raised in the comments above--what is the SWR at times of low valuation?--in several threads. For information on that question, I would start with the second thread in the list provided at the link, the one that I put the sticky on. There's a lot of additional research in the 48 threads below that. You probably can get a reasonable sense of which ones are most helpful by taking a look at the thread titles.

SWRs matter. They matter a lot. I have been studying the concept of the data-based SWR (I use this term to contrast the concept being examined at the new board from the concept discussed at the Motley Fool board, which purports to be a data-based concept but which in reality takes into consideration only a portion of the data relating to the question of what withdrawal rate is safe) since 1995, and every year since I have discovered new ways to put this concept to use in informing my investment strategies. The data-based SWR is the most valuable investing tool that I have ever come across, and I am becoming increasingly encouraged to see larger segments of the FIRE community develop an interest in exploring the realities of SWRs.

Another thread that you might want to check out is the one in which Ed Easterling of Crestmont appeared at the new board. Easterling has done some eye-opening research re patterns of stock market returns over time. There is a link to his research at the beginning of that thread, and some great back-and-forth discussion follows it. I hope to be bringing additional experts to the SWR board next year, people like William Bernstein, Scott Burns, and Peter Bernstein. But I need to get more posters participating on a regular basis before I would feel comfortable sending out additional "special event" invitations. So I would be thrilled if one or two posters from this community were to join us (while continuing to participate here, of course).

Hope to see some of you at the new board!

[sgeeeee]

Quote:

I'm not suggesting that you should only look at non-overlapping intervals. *In fact, especially since we're using an inter-year high, I would suggest that you look at the 1200 overlapping 360 month intervals as well to get a feel for how much the inter-year volatility has made a difference. *I'm only suggesting that among all the criticisms of using the 4% rule today, one completely valid criticism is that the last 130 years provides very little data about 30-40 year time horizons.

I'm not sure why some of you have focused on correlation and decided that it is bad. Correlation is, in fact, one of the strengths of a historical calculator -- not a weakness. The financial data (investment returns, inflation etc.) for a given year is highly correlated, not only to itself, but to the financial data and events of previous years. This correlation is important, but so complex as to be nearly impossible to understand and predict. The complexity of this correlation is one of the key reasons why Monte Carlo simulations so badly underestimate safe withdrawal rates. By using historical data in correct chronological order, we are guaranteed to include the complex data correlations without understanding the underlying causes for it or having to describe those correlations mathematically.

As far as the need for more data, that is a request of highly questionable value. The real weakness of a historical simulator is what was referred to as "stationarity" in a previous post. Why do we think that tomorrow will be similar to yesterday? I think there are good reasons to justify throwing out all the data prior to the 1930's. Many laws, policies and procedures have been put in place to insure that the 1929 stock market crash will never happen again. These rules lead one to question whether the data prior to their establishment is really relevant to planning for tomorrow. The further you go back to include historical data, the more important the question of relevance becomes. Do we really believe the financial world prior to 1871 is of critical importance to the finanical world of 2000 and beyond?

Quote:

Perhaps I misunderstood your post, but while the 4% rule may be interpreted as merely a historical statement, I don't see why anyone would talk about it unless they think it is some sort of prediction about the safety or intelligence of decisions today.

Yes. That's exactly the point I've been making. If you believe it has relevance as a predictive tool, then certain conclusions follow. I've stated what some of those are. If you don't believe it has relevance, then it doesn't make sense to do perturbation analysis on it by trying to second guess how to modify it for high P/Es or other worst cases.

Quote:

"You can believe that if it were safe to retire in 2000 (and you haven't spent your nest egg) it is still safe to retire today -- regardless of the relative value of your nest egg at those two times." *This is where I disagree. *You have more information today than you had in 2000, and what was a prudent decision then may be no longer. *In January of 2000 the probability of the 2000 to 2030 period being worse than any 30 year calendar year interval in the last 130 years was quite low. *Now that we have had three years of losses that have only been seen twice before in those 130 years it is far more likely that you a facing that "worst case."

First, please note that the quote you use here is out of context. It was qualified by the preceding sentence in that post: "If you assume that the future for retireees who retired in 2000 is not likely to be any worse than the worst cases experienced by retirees in 1929 or 1966 or any other year since 1871, then you should be comfortable retiring based on 4% withdrawal rate of your 2000 nest egg." But as a matter of fact, I do believe this statement -- because I don't believe that the retirees of 2000 are really going to face retirement financial issues as bad as those who retired in 1929. That's just my opinion and I wouldn't begin to claim I could prove it or to argue with you if you believe otherwise. But please notice that if you believe that a retiree today will not be safe by calculating an initial withdrawal rate based on the high value of his/her portfolio in March 2000, then you must think that he/she is going to face tougher financial challenges than the 1929 retiree. Otherwise your conclusions are inconsistent. And that is the point I have been making stated in yet another way.

[sgeeeee]

Quote:

I'm not suggesting that you should only look at non-overlapping intervals. *In fact, especially since we're using an inter-year high, I would suggest that you look at the 1200 overlapping 360 month intervals as well to get a feel for how much the inter-year volatility has made a difference. *I'm only suggesting that among all the criticisms of using the 4% rule today, one completely valid criticism is that the last 130 years provides very little data about 30-40 year time horizons.

I'm not sure why some of you have focused on correlation and decided that it is bad. Correlation is, in fact, one of the strengths of a historical calculator -- not a weakness. The financial data (investment returns, inflation etc.) for a given year is highly correlated, not only to itself, but to the financial data and events of previous years. This correlation is important, but so complex as to be nearly impossible to understand and predict. The complexity of this correlation is one of the key reasons why Monte Carlo simulations so badly underestimate safe withdrawal rates. By using historical data in correct chronological order, we are guaranteed to include the complex data correlations without understanding the underlying causes for it or having to describe those correlations mathematically.

As far as the need for more data, that is a request of highly questionable value. The real weakness of a historical simulator is what was referred to as "stationarity" in a previous post. Why do we think that tomorrow will be similar to yesterday? I think there are good reasons to justify throwing out all the data prior to the 1930's. Many laws, policies and procedures have been put in place to insure that the 1929 stock market crash will never happen again. These rules lead one to question whether the data prior to their establishment is really relevant to planning for tomorrow. The further you go back to include historical data, the more important the question of relevance becomes. Do we really believe the financial world prior to 1871 is of critical importance to the finanical world of 2000 and beyond?

Quote:

Perhaps I misunderstood your post, but while the 4% rule may be interpreted as merely a historical statement, I don't see why anyone would talk about it unless they think it is some sort of prediction about the safety or intelligence of decisions today.

Yes. That's exactly the point I've been making. If you believe it has relevance as a predictive tool, then certain conclusions follow. I've stated what some of those are. If you don't believe it has relevance, then it doesn't make sense to do perturbation analysis on it by trying to second guess how to modify it for high P/Es or other worst cases.

Quote:

"You can believe that if it were safe to retire in 2000 (and you haven't spent your nest egg) it is still safe to retire today -- regardless of the relative value of your nest egg at those two times." *This is where I disagree. *You have more information today than you had in 2000, and what was a prudent decision then may be no longer. *In January of 2000 the probability of the 2000 to 2030 period being worse than any 30 year calendar year interval in the last 130 years was quite low. *Now that we have had three years of losses that have only been seen twice before in those 130 years it is far more likely that you a facing that "worst case."

First, please note that the quote you use here is out of context. It was qualified by the preceding sentence in that post: "If you assume that the future for retireees who retired in 2000 is not likely to be any worse than the worst cases experienced by retirees in 1929 or 1966 or any other year since 1871, then you should be comfortable retiring based on 4% withdrawal rate of your 2000 nest egg." But as a matter of fact, I do believe this statement -- because I don't believe that the retirees of 2000 are really going to face retirement financial issues as bad as those who retired in 1929. That's just my opinion and I wouldn't begin to claim I could prove it or to argue with you if you believe otherwise. But please notice that if you believe that a retiree today will not be safe by calculating an initial withdrawal rate based on the high value of his/her portfolio in March 2000, then you must think that he/she is going to face tougher financial challenges than the 1929 retiree. Otherwise your conclusions are inconsistent. And that is the point I have been making stated in yet another way.

[sgeeeee]

Quote:

What I meant by my comment about government collapse was not that you should be more conservative, but rather that you should be less. Why worry about the last 5% probability that your money will last 30 years, when it's not that unlikely that you'll have much bigger problems by the end of that time?

I agree with this statement. Rather than look at the 4% rule as optimistic, I see it the other way around. I do believe that it is unlikely that we will see financial environments as bad as those faced by previous worst case retirees in history. I also think it unlikely that we will see investment boom times for at least 10 or 20 years, but that limits upside potential - not downside.

To be sure, there are events that concern me regarding my retirement over the next 40 or so years. Recent political actions have transferred exhorbitant amounts of debt from the wealthy of today to the working class of tomorrow. I can't claim to know the impact that will have on us, but I don't see anything good happening to middle class retirees because of it. Still, I have trouble seeing a future as bleak as that faced by the retiree of the Great Depression era.

[haha]

Quote:

I'm not familiar with the term "stationarity", but I am familiar with statistcal analysis, ....

Stationarity of data series in the analysis of time series relationships has been a major interest in the statistical community. For your interest, here is a definitive explication, which describes the work of Nobel Laureates Clive Granger and Robert Engle.

http://www.nobel.se/economics/laurea...03/public.html

[sgeeeee]

Quote:

Stationarity of data series in the analysis of time series relationships has been a major interest in the statistical community. For your interest, here is a definitive explication, which describes the work of Nobel Laureates Clive Granger and Robert Engle. http://www.nobel.se/economics/laurea...03/public.html

Thanks for the link. Interesting article.

I had assumed that nonstationarity referred to a condition where some economic variable suddenly shifted to a new norm and stayed there. But this article describes something quite different.

I'm not sure how this concept relates to historical simulation at all. The article describes nonstationarity to be the characteristic of a variable that follows a long-run trend, where temporary disturbances affect its long-term level. This is a problem if you are trying to fit this variable to a predictive curve. The combined contributions of Granger and Engel overcome this problem and allow them to develop predictive time series expressions to variables that were previously difficult to model.

But a historical simulator avoids the problem of nonstationarity by using actual data. This data clearly includes any nonstationarity properties within it. No time series is ever developed, so a time varying error term is not relevant to the model.

So why do you see nonstationarity as a problem for this analysis?

[bongo2]

"if you believe that a retiree today will not be safe by calculating an initial withdrawal rate based on the high value of his/her portfolio in March 2000, then you must think that he/she is going to face tougher financial challenges than the 1929 retiree."

While that may have seemed highly unlikely March of 2000, I don't think it's all that unlikely today. What does FireCalc say? If I look at the default options for FireCalc and raise my initial portfolio to $750k (so I have a 4% withdrawal), then what do I look like at the end of three years. If I look at 2002 (not in FireCalc, but we have the information that we need) we have spent $32k and have a year-end portfolio of $436k. We are withdrawing at a rate of about 7.5%. Now look at 1929. After three years we have a portfolio of $367k, worse right? Ahh, but because of deflation we have only spent $17k in 1931, so our withdrawal rate is 4.6%. It takes 11 years for us to reach a 7.5% withdrawal rate after 1929. 1966 is similar with a 4.7% withdrawal rate after three years. So, with the additional data of the last three years, it doesn't seem that unlikely at all that march/2000-2029 will be worse than any period in the last 130 years.

[sgeeeee]

Quote:

. . . So, with the additional data of the last three years, it doesn't seem that unlikely at all that march/2000-2029 will be worse than any period in the last 130 years.

And as I said before . . . I wouldn't begin to argue with you about that. If you think retiring in 2000 is likely to be worse than retiring in 1929, then you certainly shouldn't use your portfolio value in March 2000 to compute a safe withdrawal rate today. Of course, I'm not sure what you should use if you think tomorrow will be worse than any time in the past. I guess you have to answer the question, "How bad is it going to be?"

Good luck with figuring out the answer to that.

[dory36]

I just want to throw out a comment about the statistical issues and the SWR concepts.

First, remember that the "S" stands for "safe" - meaning we're planning for worst case, not trying to predict the actual outcome. Second, we're stuck with the assumption that the span of our retirement won't occur during years that are worse than the worst ever seen in recorded market history. (Any better ideas?)

In a past life, I ran a statistics & computer department at a research-oriented medical school. Every funded study by the medical researchers had to have our phd-level statisticians [which did not include me!] develop and execute the statistical portion of the study. The goal was to insure that the statistics behind any conclusion were rock-solid, so that the conclusions could be relied upon.

The conclusions were usually along the lines of, if you do abc, then xyz will happen, with a high degree of probability.

The SWR concepts and the tools that implement them (such as Firecalc) don't do anything of the sort. Instead, they look at all of the data to verify that something has never happened. (That something being, of course, a certain level of depletion of a nest egg, given a defined withdrawal amount.)

Let's look at weather for an analogy.

If I had to predict the temperature for Cape Canaveral Florida for next July 4th at noon, I would have to use lots of statistics to even get a wide range of possible temperatures. Even a wide range prediction would be suspect, unless everyone agreed on the statistical approach I had taken - and even then I could be pretty far off.

On the other hand, if the question is "can I count on it not being below 60 degrees at noon on next July 4th, assuming the weather won't be any stranger than the strangest weather we've ever seen", then it is no longer a statistical prediction, but simply a reference to historical records.

I can look at all of recorded weather history and see that it either was or was not ever below 60 on 7/4 in any year in the past.

Y'all keep on with the discussions - they are fascinating. But just note that the SWR concept is historical and not statistical.

Dory36

[moguls]

Thank you Dory36!!

I agree with you. This is all interesting discussion, but many seem to be searching for that crystal ball that will predict the future and give them 100% certainty. The bottom line is that no one knows what turn of events might occur. All we can do is look at the data we have; develop plans that will not keep us up all night; jump in and be prepared to adjust if/when something unplanned for occurs.

I know REing is a scary thought for many, including me. But we made it this far. And those of us considering ER are in much better shape than most. So do your analysis, but have some confidence in your abilities that got you this far and go for it. Or just keep working until you feel comfortable to take the leap. Unfortunately for some that will be too late.

Moguls
(204 Days and counting)

[hocus]

[i}In a past life, I ran a statistics & computer department at a research-oriented medical school.[/i]

You were one of my favorite posters in the days when we were posting together at the old REHP board, Dory36. So I have always considered you a bit of an "expert" on the subject of early retirement. I was not aware of your statistics background, however. Learning about this gives me an idea.

As noted in my earlier post in this thread, one of the initiatives that I am pursuing at the new SWR board is putting together a series of Special Event discussions with experts that can add to the FIRE community's understanding of the realities of SWR analysis. The thread at which we had a Special Event discussion with Ed Easterling can be found at the link set forth in my earlier post. I am hoping to be able to put together Special Event discussions with <b>Gummy</b> (he has a web site that deals with SWRs and other matters); the financial columnist Scott Burns; and the authors Peter Bernstein and William Bernstein. If those efforts are successful, I have thoughts of holding additional Special Event discussions. If we can get enough participation at the SWR board to make it realistic, I would like to hear what people like Robert Shiller and Jeremy Siegel and the authors of the Trinity Study have to say. We are at the early stages of the Great SWR Debate, but I believe that we are going to take this thing to some amazing places as we work to develop the insights that have been put forward in the first 17 months.

Given your significant contribution to the FIRE movement (providing a place for aspiring early retirees to share ideas with each other is a life-enhancing act of great significance) and your deep knowledge of statistical issues, you would be a great Special Event guest. Would you be willing to participate as the "expert" at our next Special Event discussion at the SWR board?

If you announced the event here, my expectation is that we could get some posters here to join in during the special event. There clearly are posters here capable of putting forward some powerful insights re SWRs. So you woould be helping me get the new board get off the ground if you participated.

Anyway, thanks for the work you do with the board here. I have not been able to participate in the past, and I don't expect to have time to do so much more anytime real soon. But I have hopes of seeing some additional time for posting open up next year, and perhaps I will be able to step in here from time to time then. Even when I am not posting here, I try to stop by from time to time to see what you guys are up to. I am a big believer that the world needs as many FIRE boards as can possibly be created. We all need to do what we can to make this movement grow bigger and bolder, faster and better, and you are clearly in the group that has made an wonderful contribution in getting things off the ground in the early days of the movement.

As one friend of the Wave to another, I hope that you will be able to join us for a Special Event discussion, or, if not, to at least stop by and put up a post from time to time. In any event, it's been nice to have a chance to share some thoughts with you once again. It's been too long!

[dory36]

Yikes - I created the wrong impression about my statistical expertise! That will teach me to respond to messages during a sleepless early AM.

I didn't mean to claim statistical expertise at all, but rather to relate situations in which statistical expertise might be needed, compared to those where such expertise might not make such a contribution.

The department I ran was a combined computer science/statistics/computer service department. The statistics faculty worked for me (in theory - anyone with academic background knows that faculty work for themselves only! ) so I was frequently involved at a high level, but I am certainly not an expert.

As far as participating, thanks, but I'll have to beg off for now. I am in the final hours of moving, and am even reading email and writing this on a PC in a marina office, having packed up all my stuff. I'll certainly "drop by" once I get back to Texas and get reestablished.

Dory36