Help Calculate Magic Number for Young Family

[Apex1]

Assume a young family with expenses of 100k for the first 25 years, and then expenses of 50k for the next 25 years after that (empty nest period).

How would you calculate a retirement number for this whole 50year period?

Of course 3.3+mill, should be sufficient with a 3%swr if one wanted to be very conservative as the family would have 50k annual surplus during the empty nest period.

I know the magic number should be less...... but how much less? How low a number can the family "shoot for" and still be 99% confident?

Anyone care to take a stab. Thanks in advance.

 

[REWahoo]

Have you tried running those numbers through FIRECalc? Input $100K annual spending and on the Other Income/Spending tab, input a pension of $50k starting 25 years after you retire, leaving the inflation adj box checked. This should have the effect of reducing your spending from your portfolio to $50K at that point.

Oh, and on the Investigate tab, select the Starting Portfolio Value button under the "Given a success rate, determine spending level for a set portfolio, or portfolio for a set spending level" option.

 

[Apex1]

REwahoo: thank you so much for helping me! My head was beginning to hurt trying to figure it out on my own. Took me awhile to figure how to input the data into Firecalc and also never realized how powerful Firecalc is. Besides, i would have never figured out that adding 50k pension would have the same effect as 50k reduction in spending.

So, to answer my own question..... Firecalc spit out 2.6m (100k spending to start, 50k pension, 99%confidence)....... pretty neat stuff

 

[REWahoo]

FIRECalc isn't perfect by a long shot, but it can do some really neat stuff if you play around with it for a while.

BTW, I haven't attempted to run your numbers through the calculator but your $2.6M number sound reasonable to me.

 

[Animorph]

Careful running 50 year periods with FIRECalc. That won't give you many tested periods.

 

[REWahoo]

Apex, to better understand the caution note from Animorph take a look at FIRECalc: How it Works, especially the section titled "How about describing FIRECalc step by step?". Bottom line, the longer the retirement period the fewer the retirement cycle 'test runs' by FIRECalc.

 

[Apex1]

Careful running 50 year periods with FIRECalc. That won't give you many tested periods.

Thanks Animorph,

Right and i did re-read the explanation of how FIRECalc works, but what other alternative is there? Isn't Firecalc already "conservative enough" if I input and look for a 100% success despite a 50 year period??

I guess you could mitigate the risk over such a long period by adjusting the spending along the way or possibly side jobs/re-entering the workforce if there is a shortfall. Or padding the initial requirement numbers even more leading to the OMY syndrome.....

 

[Animorph]

Thanks Animorph,

Right and i did re-read the explanation of how FIRECalc works, but what other alternative is there? Isn't Firecalc already "conservative enough" if I input and look for a 100% success despite a 50 year period??

I guess you could mitigate the risk over such a long period by adjusting the spending along the way or possibly side jobs/re-entering the workforce if there is a shortfall. Or padding the initial requirement numbers even more leading to the OMY syndrome.....

You won't get any scenario that includes starting retirement in 1999, for example, if your retirement period is more than 12 years or so (depending on whenever the last year of data is). For a 50 year period, your last starting year is roughly 1961. Maybe not one of the better times to start, but maybe not one of the worst. One thing you can do is try a shorter period but require a specific ending value. In your case two 25 year periods might be useful.

The is nothing "conservative" about FIRECalc, only about how you use it. It uses actual historical returns and inflation. If you think future returns will be less than historical averages, FIRECalc may be optimistic.

 

[Live And Learn]

One thing you can do is try a shorter period but require a specific ending value. In your case two 25 year periods might be useful. .

What a GREAT idea ! I'm going to do that right now for myself !

 

[youbet]

Apex, to better understand the caution note from Animorph take a look at FIRECalc: How it Works, especially the section titled "How about describing FIRECalc step by step?". Bottom line, the longer the retirement period the fewer the retirement cycle 'test runs' by FIRECalc.

Yes, and longer test periods can actually look better than shorter periods in some cases. Intuitively you wouldn't think so, but in some of my runs with 100% success rates, I have a higher minimum ending balance using 30 yrs than with 25 yrs.

 

[Apex1]

One thing you can do is try a shorter period but require a specific ending value. In your case two 25 year periods might be useful.

The is nothing "conservative" about FIRECalc, only about how you use it. It uses actual historical returns and inflation. If you think future returns will be less than historical averages, FIRECalc may be optimistic.

Requiring a specific ending value seems arbitrary to me. What would you suggest for the assumptions given in the OP? I initially considered the idea of breaking the calculation into two 25 periods with 100% success rate in each period but that gave me a number of 3.75m that seems too high (2.5m for the first 25 year period of 100k expense and 1.25m for the second 25 year period of 50k expense) ...... even higher than 3.3m (3% swr of 100k throughout the entire 50 years)

 

[Apex1]

Yes, and longer test periods can actually look better than shorter periods in some cases. Intuitively you wouldn't think so, but in some of my runs with 100% success rates, I have a higher minimum ending balance using 30 yrs than with 25 yrs.

Yes I found the same results with many of the longer test periods.

My intuition tells me that FireCalc is less reliable in these cases (because of less sample size/test runs) but on the other hand,my intuition also tells me the longer the test period, the greater the likelihood of regression to the mean.

 

[Apex1]

The is nothing "conservative" about FIRECalc, only about how you use it. It uses actual historical returns and inflation. If you think future returns will be less than historical averages, FIRECalc may be optimistic.

I was using "conservative" in reference to using 100% success rate in FIRECalc which usually results in an average ending portfolio that is significantly higher than the starting portfolio.

 

[Animorph]

Requiring a specific ending value seems arbitrary to me. What would you suggest for the assumptions given in the OP? I initially considered the idea of breaking the calculation into two 25 periods with 100% success rate in each period but that gave me a number of 3.75m that seems too high (2.5m for the first 25 year period of 100k expense and 1.25m for the second 25 year period of 50k expense) ...... even higher than 3.3m (3% swr of 100k throughout the entire 50 years)

I don't have a great answer for the OP. My best quick answer would be to run the second period first with the normal $0 ending. Find the 100% success rate starting value. Then run the first period with an ending value equal to the 100% success rate second period starting value. Then you can find the first period starting value that gives you 100% success. Or, you could try 90% or whatever. Given these are not exactly random trials, I think the combined probability of 90% and 90% is open to interpretation.

 

[Apex1]

I don't have a great answer for the OP. My best quick answer would be to run the second period first with the normal $0 ending. Find the 100% success rate starting value. Then run the first period with an ending value equal to the 100% success rate second period starting value. Then you can find the first period starting value that gives you 100% success. Or, you could try 90% or whatever. Given these are not exactly random trials, I think the combined probability of 90% and 90% is open to interpretation.

Thanks or spelling that out or me! Great idea!! I got slightly ober 3mill doing it this way (90% or the irst period).

 

[Amethyst]

The effect of pension income, for us, is very different from the effect of reduced spending. Our pensions are/will be entirely taxable - Federal, state, local. Reduced spending, OTOH, is an exact dollar-for-dollar gain.

Your mileage may vary, as not all pensions are taxed, or taxed the same way.

Amethyst

esides, i would have never figured out that adding 50k pension would have the same effect as 50k reduction in spending.

 

[ERD50]

Requiring a specific ending value seems arbitrary to me. What would you suggest for the assumptions given in the OP? I initially considered the idea of breaking the calculation into two 25 periods with 100% success rate in each period but that gave me a number of 3.75m that seems too high (2.5m for the first 25 year period of 100k expense and 1.25m for the second 25 year period of 50k expense) ...... even higher than 3.3m (3% swr of 100k throughout the entire 50 years)

The 'problem' (better to call it an anomaly?) with two consecutive 25 year runs is you get hit with the the absolute worst 25 year period, followed by the exact same absolute worst 25 year period. A double-whammy!

It's as if we had two consecutive Great Depressions, or two 70s-80-s market/inflation cycles spaced 25 years apart. I don't think any 50 year period has actually been as bad as two of the worst 25 year periods stacked together.

It might still be useful, just keep that in mind when looking at the numbers.

-ERD50

 

[Apex1]

The effect of pension income, for us, is very different from the effect of reduced spending. Our pensions are/will be entirely taxable - Federal, state, local. Reduced spending, OTOH, is an exact dollar-for-dollar gain.

Your mileage may vary, as not all pensions are taxed, or taxed the same way.

Amethyst

Enter the after-tax pension amount?

For me personally, I am trying to predict all my personal expenses on an after tax basis (and not include the taxes as an expense). However, the spending amount ("start here tab")I enter into Firecalc is grossed up to a before-tax amount. At least, thats I how look at it.

 

[Apex1]

The 'problem' (better to call it an anomaly?) with two consecutive 25 year runs is you get hit with the the absolute worst 25 year period, followed by the exact same absolute worst 25 year period. A double-whammy!

It's as if we had two consecutive Great Depressions, or two 70s-80-s market/inflation cycles spaced 25 years apart. I don't think any 50 year period has actually been as bad as two of the worst 25 year periods stacked together.

It might still be useful, just keep that in mind when looking at the numbers.

-ERD50

Thanks for putting this into context! At the same time, if you string two consecutive bull markets, you might have been able to pull the plug ALOT sooner.

 

[FinancialDave]

To be honest I really don't think any of these calculators are going to be any more accurate than a general rule of thumb that if you save 15% of your salary from in your mid twenties to age 66 you will do fine. If you want to retire earlier then just up the saving rate - the goal being to hit 25 times your retirement expenses by retirement time. When you get within 15 years or so of retirement, you can start fine tuning your glide path, but trying to do it over 25 years is a bit much.

fd

 

[Apex1]

To be honest I really don't think any of these calculators are going to be any more accurate than a general rule of thumb that if you save 15% of your salary from in your mid twenties to age 66 you will do fine. If you want to retire earlier then just up the saving rate - the goal being to hit 25 times your retirement expenses by retirement time. When you get within 15 years or so of retirement, you can start fine tuning your glide path, but trying to do it over 25 years is a bit much.

fd

Thanks FD! With a few adjustments to my current lifestyle (losing the nanny/housekeeper, no private schools for the 2 kids and downsizing to 75% of my current home)....... I'm already there. If I maintain my current lifestyle, then I am 5 years away (at which point, I am still looking at a 50year retirement)...... hence, its crunch time for me to be fine-tuning and deciding "more money versus more life"

 

[FinancialDave]

"more money versus more life"

That is always the choice that we make, and to some degree if we make that one wrong, for ourselves, it is tougher to live with, if we went too conservative, than "I had to downsize" a little later in life.

fd