A question about the Investigate tab in FireCalc

[kyounge1956]

I've just been running some last-minute checks in FireCalc. I put in my portfolio balance, asset allocation, pension income and so on, and got 100% success on any length of retirement up to 40 years, with the ending portfoliio always larger than the starting value.

I then went to the Investigate tab, and without changing any of the inputs, asked what size portfolio I needed for 100% success. The result was about $32K more than my original input. I asked for a spending level and the result was about $5K/year less than my original input. The plots for both Investigate portfolio (below) and Investigate spending level show 100% success even with portfolios quite a lot smaller, or spending quite a bit higher, than my original input.

I'm confused. Why the disparity in the results?
I just tried another little experiment--if I cut the length of retirement in half, from 40 to 20 years, there is no change to the portfolio size needed for 100% success. How can that be right? What am I missing?

 

[ERD50]

I then went to the Investigate tab, and without changing any of the inputs, asked what size portfolio I needed for 100% success. The result was about $32K more than my original input. I asked for a spending level and the result was about $5K/year less than my original input.

This is probably due to the quantization effect of looking for X% success rate. IOW, since there are 101 cycles evaluated, there will be a jump in values between 100 successes, and 101 successes. So it kind of depends on the algorithms - are they coming from below and working up, or the other way round, or incrementing values by some % and re-testing? Not unexpected.

The plots for both Investigate portfolio (below) and Investigate spending level show 100% success even with portfolios quite a lot smaller, or spending quite a bit higher, than my original input.

I'm confused. Why the disparity in the results?
I just tried another little experiment--if I cut the length of retirement in half, from 40 to 20 years, there is no change to the portfolio size needed for 100% success. How can that be right? What am I missing?

We would need to see your actual inputs, but I'm guessing that your pensions are COLA'd? So I'd further guess that the 'danger zone' for you is somewhere in the first 20 years. Then pensions/SS kick in and your actual portfolio withdraws put you in the 'forever' category. Sounds good to me! Congrats!

Also, don't forget that a 20 year run has 20 more data cycles in it than a 40 year run. A 40 year run might eliminate some bad scenarios.

Caveat: I'm on my first cup of coffee, apply more than the usual grain of salt to anything above.

-ERD50

 

[Rustward]

<snip>
We would need to see your actual inputs, . . .
<snip>
-ERD50

. . . which is pretty easy to do . . . http://www.early-retirement.org/for...i-doing-something-wrong-49466.html#post920419

otherwise the best anyone can do is suppose and guess.

 

[rbmrtn]

My guess is that in the first case you already have a 100% success rate and more likely it is above 100%. It is just showing that what you have and plan to spend, nothing fails.

In the second set, it is answering a different question but showing again you have more than enough. You can have a smaller portfolio or spend more and still 100% successful.

 

[kyounge1956]

thanks to everyone, I think the explanation must be somewhere in the comments, especially rbmrtn's, that my original inputs were more than needed for 100% success. I ran several other scenarios later in the evening and got a 100% success rate even with a portfolio of only $50K, although with such a small amount the ending value was not always higher than the starting value.

To answer ERD50's question, my pension is partly COLA'd, but the sum of the COLA'd portion plus SS is only a few thousand dollars per year less than my estimated spending. I did try to link to the page with my inputs, but when I tested the link it didn't go to my results page, so I deleted it.