Reasonable Probability?

[intercst]

I was also taken with cut-throats comment about a 100% SWR being safer than your health. As I understand it, risk factors which determine the mathematical probability of any given outcome interact via multiplication. In other words, for a given age the probability of financial failure (defined as being alive with a failed portfolio) would be the product of the actuarial probability of surviving to that age X the probability of portfolio failure.

Thus the "combined risk" of financial failure is actually some fraction of the portfolio risk. (uncertain though it may be) Comforting in a way  :

Any statistical types in the house?

Almost. You multiply the probability of portfolio failure by the life mortality for each year, then sum the products.

Here's a spreadsheet you can use to make the calculation in conjuction with the FIRECalc detailed results.

http://www.retireearlyhomepage.com/SWRLE2004.xls

Maybe dory36 can hardwire this into the next update of the program.

intercst

 

[charlie]

It seems to me that that the probabilities need to
be uncorrelated for that math to work. What about
the possibility of people who affect their life spans
through un-wise health choices that require a
high burn rate to maintain? I don't think you
can apply actuarial averages to individuals in this
case. But ....... I could be wrong.

Cheers,

Charlie

 

[ESRBob]

A survivability of 80% with Monte Carlo is roughly equivalent to the 100% safe historical SWR of 4%.

intercst

Intercst;
I'm really glad you raised this point, since I have been frustrated with Monte Carlos all along. ( I find they don't build any regression to the mean into them, and thus have some pretty extreme values coming out of them). Is this a personal rule of thumb (100 % SWR surviavability rate at 4% roughly equivalent to 80% Monte Carlo survivability)? Do you know of any studies that might back up this relationship more broadly?

Thx

 

[sgeeeee]

. . . Is this a personal rule of thumb (100 % SWR surviavability rate at 4% roughly equivalent to 80% Monte Carlo survivability)?  Do you know of any studies that might back up this relationship more broadly?

Monte Carlo programs that simulate portfolio performance treat each year as an independent event. In other words, they assume that the current year is not causaly related to the prior years. They also assume that stock, bond and inflation rates are independent of each other. In reality, of course, the present is affected by the past and stock, bond and inflation rates are correlated.

By neglecting these complex correlations and causal relationships, monte carlo techniques allow extreme conditions that are not observed in a historical simulator. So they will tend to be more pessimistic than the historical simulations. But I don't know of a general way to quantify the degree of pessimism you should expect. The 80% monte carlo to 100% historical relationship is an empirical observation.

 

[intercst]

Intercst;
I'm really glad you raised this point, since I have been frustrated with Monte Carlos all along.  ( I find they don't build any regression to the mean into them, and thus have some pretty extreme values coming out of them).  Is this a personal rule of thumb (100 % SWR surviavability rate at 4% roughly equivalent to 80% Monte Carlo survivability)?  Do you know of any studies that might back up this relationship more broadly?

Thx

Here's a comparison of Monte Carlo and historical SWR analysis that will appear in the January 2005 update to the REHP web site.

http://www.retireearlyhomepage.com/twoperc.html

Most of the Monte Carlo calculators I've looked at give an 80% to 90% survivability for a 4% withdrawal from a 60% S&P500/40% fixed income portfolio.

intercst

 

[hocus]

Intercst:

Have you looked at the Monte Carlo methodology used by raddr in the research he has posted to the FIRE board at NoFeeBoards.com and more recently to the raddr-pages.com site? My understanding is that his approach DOES provide for reversion to the mean. Yet his Monte Carlo runs ALSO generate a SWR number far lower than the number set forth in the study you published at RetireEarlyHomePage.com.

Also, I don't think you are right in your suggestion that the concerns re the future being worse than the past raised by William Bernstein relate only to Monte Carlo studies. You say at the bottom of the article you link to that: "Bernstein lists the myriad of non-financial events that could sink your retirement (e.g., political failure causing civil unrest or hyperinflation, military action all the way up to a nuclear bomb, etc.) He puts your chances of surviving this list of calamities at about 80%. It's a waste of time and resources to attempt to plan your financial affairs for a higher probability of success." All of the factors that Bernstein lists apply to BOTH conventional studies AND to Monte Carlo studies. Your 4 percent number may well not work if there is a nuclear war.

Bernstein is saying that there is NO withdrawal rate that provides greater than 80 percent safety. This means that your "100 percent safe" number (4 percent) is really only 80 percent safe. Raddr's work (and JWR1945's work too, and also Bernstein's work described in his "Four Pillars" book) says that even this 80 percent safety claim is an overstatement. Raddr and the others are saying that 4 percent is not even 80 percent safe, that to get to 80 percent safe at today's valuation levels, you need to go with a lower take-out number or a portfolio allocation of something other than 74 percent S&P stocks.

I don't hear Bernstein as saying that it is a "waste of time" to calculate the numbers accuraely. He is saying that even if you calculate them with great precision you cannot get to a number providing better than 80 percent safety.

 

[arrete]

Bernstein is saying that there is NO withdrawal rate that provides greater than 80 percent safety. This means that your "100 percent safe" number (4 percent) is really only 80 percent safe.

Sigh. Once again you have the past and the future mixed up. Or maybe you should look up the definition of the words "If" and "Then" as in
"IF the future is no worse than the past THEN 4% is safe approaching 100%.

Only a ninny would actually think the future would be the same as the past. We can only use Monte Carlo or past data .

I will admit that intercst is remiss in not producing future data.

arrete

 

[hocus]

I will admit that intercst is remiss in not producing future data.

William Bernstein did not use any "future data" in the SWR analysis he reports on in his book titled "The Four Pillars of Investing." He reports on Page 234 of that book that the SWR for a high-stock portfolio at the top of the bubble was 2 percent. JWR1945 checked the historical data and found that it backs up Bernstein. JWR1945's research shows that the SWR for an 80-percent S&P portfolio at today's valuation levels is about 2.5 percent.

 

[intercst]

Intercst:

Have you looked at the Monte Carlo methodology used by raddr in the research he has posted to the FIRE board at NoFeeBoards.com and more recently to the raddr-pages.com site? My understanding is that his approach DOES provide for reversion to the mean. Yet his Monte Carlo runs ALSO generate a SWR number far lower than the number set forth in the study you published at RetireEarlyHomePage.com.

Also, I don't think you are right in your suggestion that the concerns re the future being worse than the past raised by William Bernstein relate only to Monte Carlo studies. You say at the bottom of the article you link to that: "Bernstein lists the myriad of non-financial events that could sink your retirement (e.g., political failure causing civil unrest or hyperinflation, military action all the way up to a nuclear bomb, etc.) He puts your chances of surviving this list of calamities at about 80%. It's a waste of time and resources to attempt to plan your financial affairs for a higher probability of success." All of the factors that Bernstein lists apply to BOTH conventional studies AND to Monte Carlo studies. Your 4 percent number may well not work if there is a nuclear war.

Bernstein is saying that there is NO withdrawal rate that provides greater than 80 percent safety. This means that your "100 percent safe" number (4 percent) is really only 80 percent safe. Raddr's work (and JWR1945's work too, and also Bernstein's work described in his "Four Pillars" book) says that even this 80 percent safety claim is an overstatement. Raddr and the others are saying that 4 percent is not even 80 percent safe, that to get to 80 percent safe at today's valuation levels, you need to go with a lower take-out number or a portfolio allocation of something other than 74 percent S&P stocks.

I don't hear Bernstein as saying that it is a "waste of time" to calculate the numbers accuraely. He is saying that even if you calculate them with great precision you cannot get to a number providing better than 80 percent safety.

I'm happy to let folks read the Bernstein article directly and see if the ***** interpretation of what he says holds any merit.

http://www.efficientfrontier.com/ef/901/hell3.htm

intercst

 

[hocus]

I'm happy to let folks read the Bernstein article directly and see if the ***** interpretation of what he says holds any merit.

I agree that people cannot go wrong by reading the source materials for themselves, intercst.

It feels good to be able to put up a post expressing my agreement with you. You may recall that there was a time when you and I agreed on a whole bunch of stuff.

The encouraging reality is that we still do.

 

[sgeeeee]

Here's a comparison of Monte Carlo and historical SWR analysis that will appear in the January 2005 update to the REHP web site.

http://www.retireearlyhomepage.com/twoperc.html

Most of the Monte Carlo calculators I've looked at give an 80% to 90% survivability for a 4% withdrawal from a 60% S&P500/40% fixed income portfolio.

intercst

Hi intercst,

Thanks for the excellent comparison and discussion of simulations techniques. The more one understands the various techniques, the more you realize that they all tell the same story to within the accuracy they offer.


from the Bernstein article:

The historically naïve investor (or academic) might consider reducing his monthly withdrawals to a very low level to maximize his chances of success. But history teaches us that depriving ourselves to boost our 40-year success probability much beyond 80% is a fool’s errand, since all you are doing is increasing the probability of failure for political, economic, and military reasons relative to the failure of banal financial planning.

 

[intercst]

from the Bernstein article: Quote:The historically naïve investor (or academic) might consider reducing his monthly withdrawals to a very low level to maximize his chances of success. But history teaches us that depriving ourselves to boost our 40-year success probability much beyond 80% is a fool’s errand, since all you are doing is increasing the probability of failure for political, economic, and military reasons relative to the failure of banal financial planning.

That must be the part of the Bernstein article that ***** didn't understand, that is, if he understood any of it at all. <LOL>

intercst

 

[hocus]

My hopes rise, my hopes fall. It's like a ride on a never-ending roller-coaster.

 

[brewer12345]

Idiocy boiled down to one sentence

Intercst:

Your 4 percent number may well not work if there is a nuclear war.

I can't think of a better way to concisely illustrate hoco-mania. Oh Lawd, save us from the dribblings of idiots, fools, and those desperately in need of thorazine!

 

[mark]

I keep a copy of "Nuclear War Survival Skills" by Kearny on hand just in case I need a drastic adjustment to my expected return.

 

[mark]

I plan for a 0% after tax real return during retirement.

so (spending + med ins) x joint life expectancy = portfolio needed

portfolio divided by initial life expectancy for rate
age 62 1/62 = 3.3% annual draw

nominal returns just offset inflation

cheesy, but fairly consistent with past world averages and current U.S valuations.

And easy to explain to my wife.

 

[mark]

oops, mistakes in typing

starting age 62 1/29 years = 3.4% rate

 

[ESRBob]

Mark,
I'd like your math more if I weren't in my 40s!
For me, it would give me about 1.6% SWR which would have my wife sending me back to work!

 

[mark]

Hey, I'm 43. Depending on whose demograpic chart you believe, I'm about the last of the Baby Boomers, or the first post-Boomer.

 

[mark]

Run Firecalc with a 1% expense ratio, which was about a cheap as you could invest in the past. I think Firecalc gives an optimistic figure if you input current low index fund expenses into a past history that always had higher investing costs. Surely past investors had some idea of their investment costs and priced securities accordingly.

Note also there are international markets with generally lower returns than the U.S. But we lack the level of information that is available for the U.S market. If any weight at all is given to these excluded results, the 0% after tax real return while withdrawing may not be far off the mark.

Todays higher valuations may be offset by todays lower index investing costs, making past average (but not likely above average) returns possible during accumulation. But the net amounts received by investors was generally lower than most sales brochures indicate.