percent of remaining portfolio-???????

[engr]

I've been using the Constant Spending Power tab for my calculations thinking this would be the Worst Case condition. I later tried the "% of remaining portfolio" tab just for fun. I got a result that was 35K less than from the CSP result. I was a bit surprised thinking I already had the worst Case condition covered. I started looking into it and I kept getting the same results regardless of whether I put in 95% or 35% for my minimum spending conditions. I saw a few threads suggesting certain settings, etc. but nothing seems to make sense.

I was considering the case for a portfolio value of $1 Meg, annual withdrawal of $84K/year for 30 years. I was using 95% for the minimum year-to-year minimum variation.

Does the % remaining portfolio model use all the variables that the CSP model uses or are there limitations between what each model processes.

This seems like such a simple thing to do yet I don't feel confident with the results I'm seeing.

Can anyone else suggest what I may be overlooking? Thanks.

 

[ERD50]

I'm not following you. $84K on $1M is an extremely aggressive 8.4% WR.

$35K less? Is the $84K the 'lesser amount'?

Maybe you have pensions or SS in there? Hard to comment w/o the full picture.

Use the "link to this set of data" to share the exact set up.

-ERD50

 

[Animorph]

$84k spending with $1M portfolio and all the rest defaults gives 15% success rate.

Changing to % of remaining portfolio with 95% floor gives me:
"Here is how your portfolio would have fared in each of the 113 cycles. The lowest and highest portfolio balance throughout your retirement was $-1,420,813 to $1,000,000, with an average of $111,840. (Note: values are in terms of the dollars as of the beginning of the retirement period for each cycle.) "

% of remaining portfolio uses your spending $ amount, but calculates the percentage of portfolio and uses that for the following years.

Not sure where $35k less comes in. % of remaining portfolio will give you a variable income.

 

[engr]

Let me start over.
Desired withdrawal rate: 110K/yr
Portfolio value: 1 Meg
soc security: 36300. at 2023 (70 yo)
pension1: 60K @2014 noncola
pension2 (actually DW Soc sec disability -presently receiving): 18K @ 2014 cola
Spending1: 21K/yr cola at 2014 (for taxes)
Retire date: 2014, no additional contributions

3.5% inflation, CSP model, fees=0.18%,
Total mkt, 1871, 30 yr treasury, 40% stock, no portfolio changes (0,0,0)
100% confidence level, Spending Level

Result = a spending level of 84.03K (less than my initial w/d rate of 110K but OK)

When I go back and choose the '% of remaining portfolio' model at a min level of 95% (95% rule) and keep all the other variables constant, I get a spending level of 61.1K.

So there is a difference of approx. $22.9K in the spending levels which I don't understand. With such a difference in results between the two models being so large which one do I believe? As mentioned above I thought the CSP model would be the worst case model since it's a constant output (considering inflation effects).

I can't get a handle on why the latter model gives such a lower value (61.1K) and hoping someone could explain it better. Thanks.

 

[Animorph]

OK. I can at least confirm that using the investigate tab and looking for say a 95% success rate by adjusting the spending level, and then switching between the constant spending power and percentage of remaining portfolio spending models results in a lower reported spending number for percent of remaining portfolio.

I think you'll have to look at the fine print for the answer. One that comes immediately to mind is that FIRECalc may be giving you the lowest spending level encountered as the portfolio fluctuates.

Now that there is a simple way to generate the "problem" we might be able to round up some better answers.

 

[Animorph]

There is this note when % of portfolio is used without requesting a success % goal:

"In other models in FIRECalc, "failure" means the portfolio drops to zero. Since you are limiting spending to a percentage of your remaining portfolio, the total balance should never reach zero — but it could become pretty small in some situations. Pay attention to the spending graph, below. Since we can't use portfolio failure as a metric, FIRECalc is following the lead of the 95% Rule from Work Less, Live More, in which one of the goals is for the portfolio to be as big (after adjustment for inflation) at the end of the 30 years as it was when you started. FIRECalc found that 78.8% of the time, the portfolio you would have left behind exceeded the portfolio you started with."

So it is likely using that same "failure" definition when you are searching for a success rate goal. A failure with % of portfolio is ending with less than the original portfolio value while a failure with constant spending power is ending with less than $0. That would definitely favor the constant spending power model.

You can use the "leave some money for my estate" box at the bottom of the investigate tab to boost the allowed portfolio minimum, though it applies throughout retirement, not just at the end.

 

[engr]

Thanks for your perspective Animorph. It gives me a different perspective.

If it's true that a failure occurs when the end value of the portfolio is less than the initial portfolio value then that's not what I want. I expect to eat into the principal and enjoy life a little. What is the purpose of leaving the same amount of money at your death as when you started? To me that seems like you are denying yourself some fun when it's not necessary. We did enough of that in the accumulation phase. Maybe there's a good reason but it's not obvious to me. I plan on leaving something for the kids but I want some enjoyment too.

Thanks again for your insight.