The Shortcomings of SWR After Retirement

[Midpack]

IMO, SWR calculators/methodology is as good a method as any to help establish "how much do I need?" during the accumulation phase (still working), it figured heavily into my pre-retirement planning.

But once retired and grappling with the various sources of income (taxable accts, tax deferred, pensions, annuities, Soc Sec, inheritance, etc.) and which to withdraw when to minimize tax impact, SWR is no longer of much if any use.

I'm a picture is worth a thousand words type, so below are the theoretical optimal withdrawals by source in our future in the first (bar) graph. [I am currently revisiting (starting) Roth IRA conversions, which will add another income source to optimize.]

The second (line) graph is a repeat of a drum I have beaten here before. While the % of initial portfolio (with and without Soc Sec, the former is far less meaningful IMO) numbers seem very "reasonable." But when you look at what the corresponding "% of remaining portolio" withdrawals that go with following an SWR or constant dollar approach actually are, the percentages increase pretty dramatically in later years. Note the % of remaining portfolio withdrawals in age 80's and 90's, I'd guess most won't have the stomach to stay the SWR course come what may. Imagine withdrawing 30% of your remaining funds at age 97! Though, no one is supposed to blindly follow SWR methodology anyway, and accredited authors have said as much every time they're published. The path through retirement distribution isn't anything like these graphs, non linear to be sure.

For your amusement.

I only wish there were much better retirement income calculators out there to help grapple with all the moving parts...

My purpose is not to debate my many assumptions or actual plans (they're more conservative), this post is only meant to illustrate the subject.

 

[Ed_The_Gypsy]

If I don't do backdoor Roth conversions, MRDs will hit me bigger later. Late as I am, this is a small window.

I use the models to tell me 'How much can I take?", not "How much do I need?"

Nice graphs, by the way.

 

[truenorth418]

Midpack, I always find your observations about withdrawal strategies to be interesting and thought provoking. I may borrow that Retirement Income by Source chart for my own purposes as I have been working on my own lifetime cash flow projections. I am now 49 and I use 95 as my life expectancy, so I have a long time to plan for.

One thing I have discovered is how my Social Security, IRA withdrawals, pension, and iBond maturities will all come on line within a few years of each other, resulting in a potentially large tax impact. I am looking into ways of smoothing and potentially increasing my withdrawals until then. I know Roth conversions might help, but my brackets have been too high even since I retired two years ago. Charting it out as you do might help me to understand the alternatives.

I have also looked into the 1/n WR approach you have posted about. I think it's a reasonable approach for one to take early in retirement and when income sources are limited to withdrawals from taxable accounts, but I agree it leads to potentially huge withdrawals late in life when one might have other sources like SS and might not be much interested in spending, say, 33% of a huge portfolio.

 

[Bestwifeever]

Withdrawing any percentage of a portfolio at any time is not the same as being forced to actually spend it, is it?

 

[clifp]

Dude you making the coolest charts!.

Ok I am curious why the big spike up in withdrawal rates as you get into your 70s. Your withdrawals never exceed 4.25% even before any other sources kick in. Most years, although not all, a normal portfolio will exceed 5% real growth. Once SS kicks in you are only withdrawing 3% of your initial portfolio and it would take a really bad string of years for you to actually lose money. I look at your chart and think most year you should die a wealthy couple.

Am I missing something?

 

[ERD50]

...

Am I missing something?

I guess I'm not following the spike either.

At any rate, I do plan on taking a different look as I age. Assuming I have the mental capacity, I'd do some version of that 'auto-pilot' approach that was linked to a while back. It was a 50-50 combo of a 'SWR' and a conservative RMD table (for someone with a few years more LE than your own estimate).

-ERD50

 

[W2R]

I don't understand the spike either. Maybe there are assumptions or some other information that we are not considering.

I guess I'm not following the spike either.

At any rate, I do plan on taking a different look as I age. Assuming I have the mental capacity, I'd do some version of that 'auto-pilot' approach that was linked to a while back. It was a 50-50 combo of a 'SWR' and a conservative RMD table (for someone with a few years more LE than your own estimate).

-ERD50

I think this is the article with the "auto-pilot" approach. Put retirement savings withdrawals on autopilot - MarketWatch

Very interesting. I like the idea of choosing the RMD value for someone with a few more years' life expectancy than my own, if one is to do it at all.

 

[Animorph]

I'm assuming the spike is due to a bad SWR scenario that just barely passes.

The percent of starting portfolio with inflation adjustments is a great academic scheme since it provides a constant lifestyle, at least in terms of income level. But any other scheme will require you to either spend more or less than that constant value level. Spending more early on is something we might all like, but it will always be riskier than spending less. And who knows if spending less or more later in retirement would be all that welcome.

So I'm planning on the constant value withdrawals. I won't feel too bad if I don't exceed the RMD WR's (with my modifications, including a 10 year minimum and a 90/10 expectancy instead of 50/50). I may try to spend more if things work out great. I have kids, so I'm OK with leaving a large estate if I can't spend it all.

 

[Lsbcal]

...(snip)...
I'm a picture is worth a thousand words type, so below are the theoretical optimal withdrawals by source in our future in the first (bar) graph. [I am currently revisiting (starting) Roth IRA conversions, which will add another income source to optimize.]
...

Yes, very nice charts you have posted Midpack.

FWIW, this is the first year we will be doing aggressive Roth withdrawals to keep down the marginal tax rate given we are also taking SS now. Before this there were some years we did aggressive Roth conversions, so some of the taxes were paid early but hopefully at a reasonable marginal tax rate. For our tax situation SS pushes the marginal tax rates up a lot, little wiggle room unless we had a Roth income stream.

I think maybe the ORP calculator shows some of this sort of stuff. May not get the individual situation exactly correct but gives the flavor.

 

[Ed_The_Gypsy]

Ok I am curious why the big spike up in withdrawal rates as you get into your 70s.

That is the heart transplant.

 

[Independent]

Two questions:
Why does SS disappear in the first graph?
Is the spike the result of picking one, very poor, investment scenario?

If I use FireCalc's history and a 4% SWR, I'll find that in most scenarios my real portfolio grows. I end up with more money than I started with and a declining percent-of-current-portfolio withdrawal.

Real people aren't "comfortable" withdrawing high percents in the later years. Those who planned reasonably but still do it are forced to by surprisingly high expenses and/or surprisingly poor investment results. I'd guess the high expense is the more common reason (LTC expenses).

Most people on this board are conservative planners who position themselves with cushions so they can cut spending if investment returns are poor. The real question is when and how much you cut spending.

 

[ERD50]

I'm assuming the spike is due to a bad SWR scenario that just barely passes. ...

Yes, I took a closer look, would like Midpack to validate this, but it looks to me like the 'DH IRA' runs out ~ age 98.

So it should be no surprise that a scenario that has the source of a large % of WD dropping to zero would also see the '% of portfolio WR' spike up.

The conservative approach would be to aim for an ending portfolio of some significant size even in the worst case for portfolio and long life. For a simple reference point, FIRECalc shows that a 3% WR ($30K/$1M) on a 43 year plan would have a worst-case ending portfolio of $381K, so your WR would go no higher than ~ 7.9% of portfolio at the end. A much milder rise than the spike shown in the OP graph.

Why does SS disappear in the first graph?

Good catch. Is this modelling a DH demise at 98 on an overall plan for age 100 joint? It looks like DW SS increases at that same point?

-ERD50

 

[Midpack]

From post #1: My purpose is not to debate my many assumptions or actual plans (they're more conservative), this post is only meant to illustrate the subject.

That said...

1) The short answer re: "the spike" - it's a result of higher taxes beginning near age 70 in the plan (Soc Sec, RMD, etc.), not higher spending for the most part. The plan seeks to minimize taxes over the whole period and actual spending after taxes is smoothed. So the intended takeaway was that taxes may alter withdrawals irrespective of spending.

The answer to at least one resulting followup question, also from post #1

[I am currently revisiting (starting) Roth IRA conversions, which will add another income source to optimize.]

2) And Soc Sec doesn't disappear, DW's Soc Sec continues. Note the X-axis is DW's age, I'm 2 years older than her, so I go poof when she's 98 in the plan.

3) Thanks for the kind remarks on providing charts. Charts make things clearer to me, and there are many members here who generate and use them often. I appreciate them all.

Hope that helps...

 

[haha]

I doubt anyone is trying to debate anything, just understand its meaning a little better.

Ha

 

[ducky911]

I think making these charts and graphs is keeping your mind sharp.

 

[ERD50]

I doubt anyone is trying to debate anything, just understand its meaning a little better.

Ha

Right. It's hard to understand what is being illustrated w/o understanding the conditions that drive the shape of the graph.

From this, and other threads, I'me probably underestimating taxes. RMDs and delay of SS will probably push me higher than I would have thought. But with such a moving target, it's tough to plan. But that's one reason I try to be conservative in my planning.

I stopped doing ROTH conversions, because it just got too complicated with the kids in school. It's not just tax bracket, it is the loss of education deductions/credits. And then there will be the HI subsidies, and at this point I don't even know if my employee HI will continue.

-ERD50

 

[Chuckanut]

RMDs and delay of SS will probably push me higher than I would have thought. But with such a moving target, it's tough to plan. But that's one reason I try to be conservative in my planning.

+1

My current 'best' plan is to move some tIRA money into Roth IRA's simply to give me a few more options to help control my income level in the future. Will I need to do that for certain? I don't know. Anymore than I know I will need to delay SS to age 70. But, the odds seem to favor it, and the long term downside seems minimal.

 

[Midpack]

I doubt anyone is trying to debate anything, just understand its meaning a little better.

Ha

Debate may have been a poor choice of words, unintentionally. I was hoping the thread would elicit discussion about the best ways to optimize various retirement income sources - IMO it's an "underserved" topic here, that many of us could benefit from, self included.

Though withdrawing taxable first, then deferred when RMD requires, delaying Soc Sec to 70 and the other fundamental concepts may indeed optimize retirement income for some, it's not always optimal - thanks to taxes among other variables. It's a tough exercise to get right, at least in my view - lots of variables and unknowns. That's what the first/bar chart was meant to get at...

 

[Midpack]

I think maybe the ORP calculator shows some of this sort of stuff. May not get the individual situation exactly correct but gives the flavor.

I haven't found anything that I am satisfied with, but I-ORP and ESPlanner (even the free online basic version) together have been helpful. The underlying data from the charts in post #1 are directly from ESPlanner FWIW.

You've provided many great charts over the years, thanks!

 

[H2ODude]

Modeling future withdrawals gives me a headache. Primarily because I believe it is so dependent on what tax rates will be in the future. We have about 60% in tIRA's and at 62 would like to convert as much as possible, but given pension there's not a lot of room in the 15% bracket to move much before hitting the 25. And, we're in a 7% state IT state. Accountant's advice is to never pay any tax before you have to (other than that golden 15%) but I'm seriously considering moving amounts into Roths that will generate some large bills now rather than risk greater ones later. Anyone else think like that?

 

[clifp]

Debate may have been a poor choice of words, unintentionally. I was hoping the thread would elicit discussion about the best ways to optimize various retirement income sources - IMO it's an "underserved" topic here, that many of us could benefit from, self included.

Though withdrawing taxable first, then deferred when RMD requires, delaying Soc Sec to 70 and the other fundamental concepts may indeed optimize retirement income for some, it's not always optimal - thanks to taxes among other variables. It's a tough exercise to get right, at least in my view - lots of variables and unknowns. That's what the first/bar chart was meant to get at...

I posed a similar question fairly recently.
It is really really hard to model all of the moving parts. Interest rates, market returns, inflation. Tax rates: on interest, on social security, on dividends and capital gains. Then include RMD on your IRA. Finally throw in ACA subsidies. You have nightmare for planning purposes.

I have some ability to control income by deciding when to recognize capital gains an loses. I finally decide that while I'l restart making some modest Roth conversion. For the most part I am just going to concentrate on minimizing my existing taxes. Decent chance I could be dead by the time my tax avoidance schemes kick in my 70s.

I know it ain't optimal but it is easier.

 

[teejayevans]

I would try to balance between taxable, roth, and IRA if possible, that way you have choices, each year you can decide how much from each based on that year's tax laws, etc.

 

[haha]

Though withdrawing taxable first, then deferred when RMD requires, delaying Soc Sec to 70 and the other fundamental concepts may indeed optimize retirement income for some, it's not always optimal - thanks to taxes among other variables. It's a tough exercise to get right, at least in my view - lots of variables and unknowns. That's what the first/bar chart was meant to get at...

I think I have never thought too much about "withdrawing", because I have never withdrawn anything from any tax advantaged accounts until I was required to at age 70.

Although at times over the years my draw from taxable accounts exceeded the same period realized income, this was always short term and now my cash income tends to be considerably higher than what I need for a moderately frugal life that I enjoy. I have done some Roth transfers, but I haven't had the long period of low taxable income that many early retirees enjoy, so my Roth conversions cost me plenty in taxes. A couple years ago a poster whose name I have forgotten explained how it is a mistake to voluntarily accept ordinary income when you have LTCGs and dividends, because due to you being pushed into the 25% bracket, in effect favored categories which had been taxed at 0% will now be taxed at 15%, plus the ordinary income will be taxed at best at 15%, and some at least at 25%. So I quit doing conversions. That may change I suppose, though at the moment I don't see why it would.

I know many of us would feel adrift without our SWR spreadsheets. I feel that is derivative, and I pay attention to the safety of my portfolio as to capital value and income generation. If I am getting adequate income without much risk taking I think that is all I need to know. If I couldn't satisfy income needs, I would cut back or try to scare-up some earned income. And I refer to "much risk taking", not no risk taking. I am always amazed at the very conservative investors here who seem to not be particularly moved by risks to life and limb, like mountain biking, bungee jumping, parachuting, walking around in foreign countries where one is a stranger and may not read things very well, etc. Give me a little financial risk!

IMO risk aversion is a rational preference for low risk given the level of return possible. Some of this other stuff does not seem to be rational at all- it is just fear.

Ha

 

[Lsbcal]

Modeling future withdrawals gives me a headache. Primarily because I believe it is so dependent on what tax rates will be in the future. We have about 60% in tIRA's and at 62 would like to convert as much as possible, but given pension there's not a lot of room in the 15% bracket to move much before hitting the 25. And, we're in a 7% state IT state. Accountant's advice is to never pay any tax before you have to (other than that golden 15%) but I'm seriously considering moving amounts into Roths that will generate some large bills now rather than risk greater ones later. Anyone else think like that?

Just a few thoughts. Tax brackets don't go up stepwise I seem to remember so there might be a bit of wiggle room between 15% and 25%. Also it depends on how many dollars get taxed at the higher marginal rate. If I made a bit of a mistake and $1k got taxed at the higher rate, that would be not so bad. But 10k would to me.

I do think that having a Roth income stream to draw from is a very nice thing. If one can manage it over some years maybe before SS kicks in.

My assumption has always been that marginal rates are going to go up in the future. Cannot prove it though.

 

[teejayevans]

My assumption has always been that marginal rates are going to go up in the future. Cannot prove it though.

I don't see the marginal rates going up, I see changes occurring around the edges, for example: for high earners, 85% of their SS is taxed, I expect this to be a 100% or they may change the way they adjust the brackets with inflation. BTW the SS limits used to decide whether SS is taxed at 50% or 85% are not adjusted for inflation, so this is a current example of how your taxes are going up. The bush tax cuts have been made permanent and there is no way they can raise them now.