Another Look at SWR, Near Worst Case

[Midpack]

This is again the classic example of a 65 yo retiree with a $1MM portfolio (and no other income sources) planning to follow the 4% SWR methodology.

When I run FIRECALC and get a 94.6% probability of success, that sounds really good - and it is [There's no such thing as 100% probability of success WRT retirement plans to begin with, just statistics]. And I look at the lines on the chart and see that most of them never drop below zero, that looks really good. But for me at least, taking that worst case that does not fail (95% success threshold) and looking at the actual withdrawals year by year is easier for me to relate to - the last column. Presumably none of us would blindly continue with the 4% SWR methodology over the course of this real scenario (I know I couldn't), but it's interesting to me that had you done so, you never would have run short of funds until the well into the last year (of course you'd have to die on schedule)!

And it's interesting to think about where on would lose his/her nerve, undoubtedly it would vary considerably.

Not making a point or looking for a debate. I do this for my own curiousity, assuming others may enjoy the mental exercise...numbers is hard.

Age	Portfolio $ Worst Non-Failure (1906)	Annual Income (w 3% Inflation)	SWR Worst Non-Failure (1906)
65	$1,000,000	$40,000	4.0%
66	$966,598	$41,200	4.3%
67	$765,346	$42,436	5.5%
68	$938,009	$43,709	4.7%
69	$1,002,902	$45,020	4.5%
70	$940,813	$46,371	4.9%
71	$928,227	$47,762	5.1%
72	$936,350	$49,195	5.3%
73	$865,240	$50,671	5.9%
74	$794,262	$52,191	6.6%
75	$921,087	$53,757	5.8%
76	$920,366	$55,369	6.0%
77	$742,218	$57,030	7.7%
78	$761,788	$58,741	7.7%
79	$775,932	$60,504	7.8%
80	$624,182	$62,319	10.0%
81	$590,219	$64,188	10.9%
82	$623,613	$66,114	10.6%
83	$566,990	$68,097	12.0%
84	$583,230	$70,140	12.0%
85	$595,216	$72,244	12.1%
86	$557,547	$74,412	13.3%
87	$607,283	$76,644	12.6%
88	$712,250	$78,943	11.1%
89	$592,492	$81,312	13.7%
90	$435,821	$83,751	19.2%
91	$254,883	$86,264	33.8%
92	$188,114	$88,852	47.2%
93	$178,973	$91,517	51.1%
94	$108,585	$94,263	86.8%
95	$60,993	$97,090	159.2%

 

[stevelb]

I appreciate your postings; very helpful

 

[Gone4Good]

And it's interesting to think about where on would lose his/her nerve, undoubtedly it would vary considerably.

I went through the same exercise a while back and came to the conclusion that many of the runs FIRECalc considers successful wouldn't feel all that successful to a real person living with that financial situation in real time. In periods of stress, it is impossible to know whether you're on a crash and burn path or one that recovers.

The common response is that folks will cut their spending and all will be well. The first part is almost certainly true. But the second part also might depend on their ability and willingness to keep rebalancing and pushing money into a market that seemingly chews up their retirement savings like a wood chipper. A perusal through the ER.Org archives circa March 2009 gives an indication of how truly uncommon that kind of discipline (or faith?) really is.

FIRECalc is useful, but it relys on many assumptions. One commonly overlooked assumption is that people will act like automatons instead of like emotional people who often make bad investment decisions during periods of stress and euphoria. An additional margin of safety is probably needed to not only offset such "user error" but to also avoid many of the "successful" scenarios that feel anything but.

 

[Alan]

Thanks for posting that table, very interesting, and scary.

 

[audreyh1]

Yes definitely! Looking at the real numbers in a bad scenario brings a lot of perspective.

To me, it tells me that my annual fixed % approach each year does keep one from getting into territory where the withdrawal gets way out of whack with what the portfolio can support over a long period - like finding oneself withdrawing 10% plus over long periods of time.

Another advantage of the fixed % approach is that in sequences of GOOD market years you are withdrawing more, as opposed to allowing it to accumulate to the point where you have a large remainder at the end. If you are smart you bank some of those higher withdrawals in the good years to help out with expenses in the bad years.

Even though taking a fixed percentage is a continuous "reset" in the sense that eventually you will run into a scenario where you have a series of bad years, that doesn't matter. Because you just keep on "resetting" as your portfolio shrinks. This last sentence probably doesn't make sense to some but I don't know how to say it better.

Audrey

 

[W2R]

I can't imagine just blithely withdrawing the inflation adjusted amount year after year like that! Whew. That would take so much courage.

Looking at the figures, it would be hard not to cut back severely for a year at age 67, and then again for an indeterminate period of time at age 77.

 

[haha]

The common response is that folks will cut their spending and all will be well. The first part is almost certainly true.

Even this is much more open to question than commonly assumed. It is a lot easier to cut spending if you are living on $80,000/yr than on $25,000. The other aspect is that there is a lot of non-discretionary spending. Also there is spending that can be put off, but only by accepting loss. For example, for homeowners, with the sort of trajectory shown the first thing that might go is home upkeep. As time went on, it would dawn on them that they might well die in rundown place with buckets stationed here and there to catch the rain leaking through the worn out roof. Not a sweet future.

Ha

 

[Gone4Good]

Another advantage of the fixed % approach is that in sequences of GOOD market years you are withdrawing more, as opposed to allowing it to accumulate to the point where you have a large remainder at the end. If you are smart you bank some of those higher withdrawals in the good years to help out with expenses in the bad years.

That's an interesting, straight forward, approach which seems to approximate something I arrived at through a bit more convoluted manner.

My basic thought is that Asset Allocation should (or at least can) change as the value of your portfolio changes relative to your withdrawal requirements. As equity prices run up, my portfolio grows, my withdrawal rate declines, and I need less equity in my portfolio to meet my new lower withdrawal rate. (explained a bit more here and here)

Your strategy seems to mechanically achieve the same thing (if I understand you correctly). After periods of good returns, you're reallocating some of the excess to something safe instead of plowing them back in to your existing AA. Your equity allocation falls as your portfolio grows relative to how much you spend.

This makes sense to me. In an extreme case (think Bill Gates) the need for equities drops to zero, although you may still choose to own some for other reasons.

 

[Gone4Good]

Even this is much more open to question than commonly assumed. It is a lot easier to cut spending if you are living on $80,000/yr than on $25,000.

Very true. Some of the bare bones ER plans don't have much flexibility.

 

[Aiming_4_55]

Very true. Some of the bare bones ER plans don't have much flexibility.

Very true - each person/family may define "bare bones" very differently. I'm not retired, but targeting a bare bones budget of $60k for a family of 4 (paid off home and cars). I know some people will disagree with me, but I include $5k for 529 college funding, $5k for family vacations, $5k for cell, cable, internet, entertainment in my annual bare bones budget. In tough years, some of that could be trimmed.

 

[Chuckanut]

Midpack, Thanks for an informative look. It reminds us that one of the consequences of our "goal" is to have zero dollars ($0) left at the end of the period. In reality, is that what any of us would want? Well... maybe it is time to look into long life insurance.

 

[ERD50]

Yes, very interesting, thanks for posting.

First, let me get a bit of nit-picking out of the way, the nit-pick probably doesn't change the basic point, but I don't want to completely ignore it.

You are using a fixed 3% inflation number for the 'spend' - some of those years were actually deflationary, right? Some might be much higher than 3%? Again, probably doesn't change things materially, but I wonder if the actual 'spend' numbers are available from the spreadsheet?

I went through the same exercise a while back and came to the conclusion that many of the runs FIRECalc considers successful wouldn't feel all that successful to a real person living with that financial situation in real time. In periods of stress, it is impossible to know whether you're on a crash and burn path or one that recovers.

Exactly. I started a thread a few years back about 'Scary Dips in Net Worth' - there were lots of scenarios where the NW dipped to 50% of the buying power you started with (and sometimes didn't take long), and yet, those were 'successes'. Different AA didn't really have a dramatic effect.

The common response is that folks will cut their spending and all will be well. The first part is almost certainly true. But the second part...

I wonder about something more basic. Will cutting spending make it 'all well'? Or does it just help a little? I don't know if that spreadsheet can be manipulated to show what a drop in spending would accomplish. I've tried to take some very rough cuts at simulating this, and I suspect that many people here over-estimate the benefit of cutting spending if you are on a failure or near-failure path anyhow.

My gut read of that sheet, while pretending I can't see the future years - if I saw I was exceeding my planned WR for 6-8 years consistently, I think I'd be concerned. Yet, the portfolio shown is still pretty close to the original value 8 years later (I guess that ignores inflation?), and fully recovered once along the way. That might lean one to think 'this too shall pass'. But I would love to see the effect of say, cutting spending to 75% of originally planned starting in year 9, where you see another dip that would likely move you to take action (but then, a decent recovery in years 11-12!).

-ERD50

 

[Midpack]

Midpack, Thanks for an informative look. It reminds us that one of the consequences of our "goal" is to have zero dollars ($0) left at the end of the period. In reality, is that what any of us would want? Well... maybe it is time to look into long life insurance.

I didn't mean to imply that. While some would indeed like to 'die broke', I gather most would like to plan on a residual to bequest for family and/or charity. Either can be baked into a plan.

 

[Gone4Good]

I wonder about something more basic. Will cutting spending make it 'all well'? Or does it just help a little?

I built a spreadsheet that approximates FIRECalc, but gives me more flexibility to play around with scenarios and see what is going on "under the hood."

I just dusted it off and ran a failure scenario: 1966 retirement, 4% withdrawal, 60% equity. Using the normal WR + Inflation, the portfolio crashes in 1998. (I guess that is success by normal FIRECalc standards). If I permanently cut spending 10% in the first year my WR exceeds 6% (1974) the portfolio lasts until 2007.

 

[clifp]

Dory (FIRECalc) author has been running the case of the poor Y2K retiree for many years now. This is certainly a scary table. Doubly so for me being both a very early retiree and a Y2K retiree. With only $367K remaining in the 1Mil portfolio odds are very good the retiree will run out of money by 2020 or so.

 

[FIREd]

This is again the classic example of a 65 yo retiree with a $1MM portfolio (and no other income sources) planning to follow the 4% SWR methodology.

When I run FIRECALC and get a 94.6% probability of success, that sounds really good - and it is [There's no such thing as 100% probability of success WRT retirement plans to begin with, just statistics]. And I look at the lines on the chart and see that most of them never drop below zero, that looks really good. But for me at least, taking that worst case that does not fail (95% success threshold) and looking at the actual withdrawals year by year is easier for me to relate to - the last column. Presumably none of us would blindly continue with the 4% SWR methodology over the course of this real scenario (I know I couldn't), but it's interesting to me that had you done so, you never would have run short of funds until the well into the last year (of course you'd have to die on schedule)!

And it's interesting to think about where on would lose his/her nerve, undoubtedly it would vary considerably.

Not making a point or looking for a debate. I do this for my own curiousity, assuming others may enjoy the mental exercise...numbers is hard.

Age	Portfolio $ Worst Non-Failure (1906)	Annual Income (w 3% Inflation)	SWR Worst Non-Failure (1906)
65	$1,000,000	$40,000	4.0%
66	$966,598	$41,200	4.3%
67	$765,346	$42,436	5.5%
68	$938,009	$43,709	4.7%
69	$1,002,902	$45,020	4.5%
70	$940,813	$46,371	4.9%
71	$928,227	$47,762	5.1%
72	$936,350	$49,195	5.3%
73	$865,240	$50,671	5.9%
74	$794,262	$52,191	6.6%
75	$921,087	$53,757	5.8%
76	$920,366	$55,369	6.0%
77	$742,218	$57,030	7.7%
78	$761,788	$58,741	7.7%
79	$775,932	$60,504	7.8%
80	$624,182	$62,319	10.0%
81	$590,219	$64,188	10.9%
82	$623,613	$66,114	10.6%
83	$566,990	$68,097	12.0%
84	$583,230	$70,140	12.0%
85	$595,216	$72,244	12.1%
86	$557,547	$74,412	13.3%
87	$607,283	$76,644	12.6%
88	$712,250	$78,943	11.1%
89	$592,492	$81,312	13.7%
90	$435,821	$83,751	19.2%
91	$254,883	$86,264	33.8%
92	$188,114	$88,852	47.2%
93	$178,973	$91,517	51.1%
94	$108,585	$94,263	86.8%
95	$60,993	$97,090	159.2%

I was wondering: it is my understanding that all portfolio balance values returned by FIREcalc are in constant dollars (1906 dollars in this case). Therefore shouldn't the annual income remain also constant at $40,000 a year?

 

[Mulligan]

Interesting chart to look at. I have a pension so I do not have to worry so much about this, but it is very scary to someone like me. I am at the age where if I had paid into a 401k, instead of a pension, I would have been thinking throughout my career, save a million and live comfortably off the 6% CDs for the rest of my life. Times have indeed changed!

 

[Midpack]

I was wondering: it is my understanding that all portfolio balance values returned by FIREcalc are in constant dollars (1906 dollars in this case). Therefore shouldn't the annual income remain also constant at $40,000 a year?

I may be mistaken, but the results in the written summary and the Excel dump are very different. It appears the written summary results are constant/current dollars and the Excel data are real dollars (much larger)! I think I was able to confirm same, though I'll spare you the details. And note the FIRECALC wording below.

It would be easy enough to produce the chart in constant or real dollars, I hope I haven't mixed them up.

From FIRECALC

Here is how your portfolio would have fared in each of the 111 cycles. The lowest and highest portfolio balance throughout your retirement was $-400,986 to $5,679,475, with an average of $1,764,891. (Note: values are in terms of the dollars as of the beginning of the retirement period for each cycle.)

Open an (unformatted) Excel spreadsheet showing the inflation-adjusted end-of-year portfolio balances for every year in each of the cycles tested by FIRECalc. Open a spreadsheet showing the year by year inputs, data, and formulas for the cycle beginning in 1960 -- Note: these spreadsheet links will expire in about 15 minutes.
For our purposes, failure means the portfolio was depleted before the end of the 30 years. FIRECalc found that 6 cycles failed, for a success rate of 94.6%.

 

[audreyh1]

Your strategy seems to mechanically achieve the same thing (if I understand you correctly). After periods of good returns, you're reallocating some of the excess to something safe instead of plowing them back in to your existing AA. Your equity allocation falls as your portfolio grows relative to how much you spend.

That's correct in the sense that I see no reason to put "excess" - i.e. money withdrawn but not spent in a given year - back into a long term portfolio where it is subject to market risks. Those are monies that might very well be spent in the short term, so they shouldn't be invested back in equities IMO. Once withdrawn, keep it out of the portfolio, save it for a "rainy day".

People looking to optimize what their heirs inherit may feel otherwise. But I'm not looking to have a lot left over if possible. Even gifting I would prefer to do before I die.

As for the portfolio (less whatever has been withdrawn) - the asset allocation stays constant via rebalancing, or you might deliberately, gradually change your AA to match the years you expect to have remaining to live. This is a separate issue from what to do with excess withdrawals.

And any "excess" short term funds that you accumulate outside of your portfolio could very well be used up suddenly in a year or two that your portfolio was hit hard by a bad market event.

Audrey

 

[ERD50]

It would be easy enough to produce the chart in constant or real dollars, I hope I haven't mixed them up.

I think you may have. Look at the last two years. You start with $108,585, and if you were to spend $40,000, you'd have $68,585. Pretty close, and I'll assume the difference is investment changes. But it doesn't seem realistic to have $108,585, spend $94,263 ( or $97,090?) and still have $60,993.

94 $108,585 $94,263 86.8%
95 $60,993 $97,090 159.2%


-ERD50

 

[donheff]

That's correct in the sense that I see no reason to put "excess" - i.e. money withdrawn but not spent in a given year - back into a long term portfolio where it is subject to market risks. Those are monies that might very well be spent in the short term, so they shouldn't be invested back in equities IMO. Once withdrawn, keep it out of the portfolio, save it for a "rainy day".

People looking to optimize what their heirs inherit may feel otherwise. But I'm not looking to have a lot left over if possible. Even gifting I would prefer to do before I die.

I have been plowing my excess back into the portfolio (or not withdrawing it in the first place). I simply earmark it as "excess" on my spreadsheet and don't consider it in calculating subsequent withdrawals. I also adjust the total excess pool up or down annually based on the growth rate of the overall portfolio. I keep a cash cushion within the portfolio that I can tap in a bad year but I figure why not let my "excess" savings sink or swim with the portfolio.

 

[Midpack]

I think you may have. Look at the last two years. You start with $108,585, and if you were to spend $40,000, you'd have $68,585. Pretty close, and I'll assume the difference is investment changes. But it doesn't seem realistic to have $108,585, spend $94,263 ( or $97,090?) and still have $60,993.

94 $108,585 $94,263 86.8%
95 $60,993 $97,090 159.2%

May be explained by your earlier observation and conclusion. Look for yourself, FIRECALC does not kick out actual income withdrawals or corresponding inflation so I used their default 3% inflation constant (knowing it's not right year by year) for lack of anything else to illustrate.

You are using a fixed 3% inflation number for the 'spend' - some of those years were actually deflationary, right? Some might be much higher than 3%? Again, probably doesn't change things materially, but I wonder if the actual 'spend' numbers are available from the spreadsheet?

 

[donheff]

I am curious about what the withdrawals in these same years would be if the retiree took a fixed 4% of the total portfolio out each year. It would be interesting to see side by side. Ready to modify your spreadsheet?

 

[Midpack]

I am curious about what the withdrawals in these same years would be if the retiree took a fixed 4% of the total portfolio out each year.

We know what you really meant...

 

[jdw_fire]

thank you midpack, for posting that table.

from earlier threads i am thinking that many here would think that if someone retired with a $500k porfolio and a $20k COLAed pension (often considered equivalent to a $1M portfolio for retirement spending purposes) there would be less worrying about the exhausting of their portfolio and reductions in WD $ over that same timeframe.

it seems to me that because of these types of scenerios (portfolio exhausting successes), buying an SPIA to cover "barebones" retirement spending would be a good plan A (especially for the people that agree with my last paragraph). i guess i dont really understand all the objections to this approach, the whole point of picking a SWR is to make sure the portfolio doesnt disappear before the retiree dies and buying a SPIA (if it is the correct size) should increase the odds of income stability and continued portfolio existance.