analysis help, please

[aggie]

we had this posted under firecalc support, but perhaps it is better suited to this topic.

We are looking at two scenarios, but would like more pairs of eyes to offer input.

first scenario is
need 74k for expenses (including taxes)
Have 860k in all investments (70/30 split)
we are currently 55, retire in 2008
retirement period is 26 years, one of pre-retire and 25 of retirement (til 80)
pension starts in 2008 at 20k, non-cola
SS at 62 of 17880 and 16452 today's dollars
result is 91.8% success chance (no bernicke, no 95% esrbob rule)

Here is link to scenario 1 using firecalc: http://firecalc.com/calc-research.p...ix5=+40+++&mix6=+10+++&mix7=+15+++&mix8=+5+++

Second scenario is a variation of the "buckets of money" theme which basically says you'll pull out some money up front, put in cash, CDs, money market, etc. and use that for first x years.
need 74k for expenses (including taxes)
have 160k for first bucket at 5.5% will last 3.2 years (at 3% inflation 4500/month)
have 100k for second bucket at 12% (BIL deal) will last 2.8 years (at 3% inflation 4500/month)
(not included in firecalc, calculated elsewhere, these two cash buckets will cover period of 2008 to 2013)
have 600k in investments (70/30)
retirement starts in 2014 in firecalc (instead of 2008 because buckets cover first 6 yrs)
pension starts in 2008 at 20k, non-cola
SS at 62 of 17880 and 16452 today's dollars
results is 99.1% success chance

Here is link to second scenario: http://firecalc.com/calc-research.p...ix5=+40+++&mix6=+10+++&mix7=+15+++&mix8=+5+++

I would have guessed that the chance of success would have been the opposite, greater chance of success in the all investment scenario and less chance of success in the "bucket" scenario, just because the money has a much greater chance of growing and thus being available for longer periods of withdrawal. Maybe it is because of the reduced number of years to be pulling from investments, but there is also a much smaller investment portfolio for the reduced years.

Any analysis, logic, ponderings about why this turns out this way would be useful to us. We're trying to insure we do this correctly and that we understand what we are seeing.

aggie and txdakini

 

[FinallyRetired]

In my opinion, your results make sense.
The first scenario has six more years in retirement, with more opportunity for variable
returns, either positive or negative. The success probability is decreased because there
were some unlucky years in those 6. The tradeoff is, there were also more successful
years, so the range of results is higher than in scenario 2. In scenario 2, you have 6 years
of cash with fixed return, so there is less variability. If you look at the range of possible
returns, you will see what I mean:

Scenario 1 91.8%
And here is how your portfolio would have ended up in each of the 110 cycles.
The range was $-629,273 to $4,206,020, with an average of $1,232,083.
(Note: values are in terms of the dollars as of the beginning of the retirement period for each cycle.)

Scenario 2 99.1%
And here is how your portfolio would have ended up in each of the 110 cycles.
The range was $-13,720 to $3,920,882, with an average of $1,362,208.
(Note: values are in terms of the dollars as of the beginning of the retirement period for each cycle.)

As another comment, why are you planning for such a short retirement period? Maybe
at age 80 you don't care, and I can sympathize, but if you do live longer you'd want at
least a little money to pay for rocking chairs or nursing care.

 

[Finance Dave]

The all investment has a higher standard deviation of returns. As such, it's more volatile. Using a Monte Carlo simulation, this has the potential for higher losses in some years, and therefore may lower your chances of success. With the 2nd scenario, it's "safer", so you have a higher likelihood of success even though you have a lower chance of "breaking out" with huge sums should the investments do well. Makes sense to me. Find the moneychimp calculator and try it...they have a good tutorial on Monte Carlo that explains this.

Dave

 

[jdw_fire]

The all investment has a higher standard deviation of returns. As such, it's more volatile. Using a Monte Carlo simulation, this has the potential for higher losses in some years, and therefore may lower your chances of success. With the 2nd scenario, it's "safer", so you have a higher likelihood of success even though you have a lower chance of "breaking out" with huge sums should the investments do well. Makes sense to me. Find the moneychimp calculator and try it...they have a good tutorial on Monte Carlo that explains this.

Dave

Another thing that helps the second scenario is that for the first 6 years you have fixed the inflation rate
at 3%, which FIRECalc doesn't do unless you tell it to, thus guaranteeing a real rate of return of 2.5% on
bucket #1 and a real return of 9% on bucket #2.

BTW, FYI FIRECalc doesn't use "a Monte Carlo simulation".

 

[txdakini]

Thanks to all of you who answered. I was sure I had made some kind of mistake in setting it up. I really appreciate the opinions.

This is just one of several scenarios we're running. In some we use 85 as end date and in some we use 80 as we're still "playing" with what we will use. but in truth, we come from amazingly short-lived stock...one of the many reasons we want to retire now. All four of our parents died in their 60s. Two at 62 and two at 69. Even with better health care and nutrition, we don't think we'll be one of the ones who lives into their nineties. Probably 85 is a safer number. We're still working on our parameters.

Again, thanks.

 

[FinallyRetired]

jdw, I stand corrected about Firecalc not using Monte Carlo,
it's been a long time since I ran it and had forgotten.
The same effect might apply, though, if the six years omitted
from one of the scenarios decreases the variability of the results.
I posted on another thread about using a number of retirement
calculators, including Firecalc, and finding different answers.
Some were MC, others some different historical starting years,
some used averages. Eventually I understood the differences,
but the real value of the exercise was just that, the understanding.

It's a great relief when you can validate your results across a
number of different tools. Firecalc was correct, in fact they were
all correct, just used different methods and assumptions.

Boy the formatting on this thread is really off.

 

[2B]

Thanks to all of you who answered. I was sure I had made some kind of mistake in setting it up. I really appreciate the opinions.

This is just one of several scenarios we're running. In some we use 85 as end date and in some we use 80 as we're still "playing" with what we will use. but in truth, we come from amazingly short-lived stock...one of the many reasons we want to retire now. All four of our parents died in their 60s. Two at 62 and two at 69. Even with better health care and nutrition, we don't think we'll be one of the ones who lives into their nineties. Probably 85 is a safer number. We're still working on our parameters.

Again, thanks.

Rather than "living forever" consider having a residual at the end of your plan. That would cover nursing care when added to SS. In a nursing home you won't have any other expenses. It would also guarantee a bias towards having a decent portfolio value at the end of the plan. There is a historical glitch in FIRECalc if you put in a very long retirement length. If you don't include the late 60's and early 70's you miss out on a period that was worse than the Great Depression as far as retirement is concerned. Doing a 40+ year plan reduces the impact of the poor market performance with high inflation on your plan.