A "Ratcheting" SWR - Kitces

[audreyh1]

It seems that a lot of these schemes go to considerable trouble to avoid the "reduction in income" scenario.

But, IMO, it's not that big a deal if you have plenty of discretionary spending in your budget (a good idea if you retire very early). Facing an occasional drop in income is far from a disaster, particularly if it has recently increased due to the portfolio doing well.

We have mostly taxable investments. I have tracked the after tax income from our portfolio for many years. Taxes create an interesting "income smoothing" effect on taking a % of remaining portfolio. In years where the portfolio has a big jump, taxes tend to be higher because more is paid in cap gains distributions. In years where the portfolio suffers a loss, cap gains distributions drop significantly, even to zero, and there may even be options for tax loss harvesting.

No such smoothing by taxes effect if you are withdrawing income from an IRA or 401K

 

[rayvt]

That is, if I'm willing to live with variable withdrawals only on the excess over 4%, I can be quicker to increase.

The problem I see with these various systems is the complexity of figuring the annual draw. 4% of starting value adjusted for inflation, plus X% of the previous year portfolio gain. UGH!

Getting Guyton-Klinger into a spreadsheet was bad enough, and G-K is conceptually simpler than a lot of these other schemes.
I ran into some difficulty converting G-K's description into a set of rules. Seems that there was a bit of handwaving there. You don't find out if everything is covered until you try to reduce it down to a set of rules for a computer.


Another problem is the rapid jumps -- both up and down -- of the draw. If the market has one very good year, your draw could go from $40K to $43K the next year, back down to $40K the following year.

 

[audreyh1]

For this reason I prefer the simple approach of % of remaining portfolio. So easy - X% of Dec 31 value each year.

I never liked the idea of inflation adjusting from the initial portfolio value. If the portfolio keeps up with inflation, great, if not, annual withdraw will stay behind until the portfolio catches up. If the portfolio zooms ahead of inflation, then I'd like to withdraw the increase sooner rather than later.

I have no problem at all with variability in annual income. IMO this issue is way blown out of proportion and makes the withdrawal calcs much more complex for early retirees. If someone's budget is so tight that they can't handle the variability, then early retirement may not be the best choice. If someone is retiring at a normal age and the budget is tight, then I can see opting for the Trinity Study method to avoid variability.

 

[Independent]

Except when income streams come online at various points, then the whole SWR exercise is useless.

I posted an example of using SWR with a deferred income stream. See (B) here:

http://www.early-retirement.org/for...f-maximize-my-ss-com-77660-6.html#post1606094

 

[Independent]

The problem I see with these various systems is the complexity of figuring the annual draw. 4% of starting value adjusted for .

I thought that my "4% x Max[current portfolio, inflated original portfolio]" was pretty simple.

 

[Andre1969]

The problem I see with these various systems is the complexity of figuring the annual draw. 4% of starting value adjusted for inflation, plus X% of the previous year portfolio gain. UGH!

I don't think it would be too hard. For instance, let's suppose you retired with $1M in 2007, and used a 4%WR, adjusting for inflation.

And now, let's suppose your ending portfolio on 12/31/14 was $1.5M. Well, 4% of $1.5M is $60,000. To find what 4% of $1M is, adjusted for inflation, just plug the number into an inflation calculator. I just did, and $40K in 2007 dollars comes out to about $45,671 in 2014 dollars. Or, $45,877 in 2015 dollars (not sure which figure you should use to calculate your 2015 wd, but in this case they're close).

So, in this hypothetical scenario, if you wanted to take a bonus, you could take $60K if you wanted, instead of ~$46K. At least, that's the way I'm reading into it.

 

[Independent]

Another one that's similar to this is Gummy's "Sensible Withdrawals". You use whatever SWR makes sense, but in good years you get a bonus. A good year is defined this way: if, after the year's withdrawal, your portfolio is greater than it was a year ago (after inflation) then take a % of that growth as a bonus. He suggests 3% SWR, but I think it would work with 4% as well. This one has been on my "top 5" list as a variable method as it tends to produce more bonuses early in retirement than later.

I may not have interpreted the rule correctly, but this is what I think he's saying.

Suppose my portfolio at the end of year 8 is 140% of the inflation-adjusted original portfolio.
And, my portfolio at the end of year 9 is 150% of the inflation-adjusted original portfolio.
Then I take a bonus of X% of the 10% growth. If X=10, I'll get a bonus equal to 1% of my original portfolio. A nice bonus compared to my 4% SWR.

Now suppose that at the end of year 10, my portfolio is again 150% of the inflation-adjusted original portfolio.
I get no bonus, because there is no growth.

But, it seems to me that 150% is enough cushion to justify a bonus. For example, if my portfolio just happens to stay at 150% in each following year, I'll never get another bonus.
Basing the extra payouts on one-year growth creates (IMO) unnecessary volatility in the extra payouts.

But, I'd have to build a model to compare carefully.

 

[utrecht]

I don't think it would be too hard. For instance, let's suppose you retired with $1M in 2007, and used a 4%WR, adjusting for inflation.

And now, let's suppose your ending portfolio on 12/31/14 was $1.5M. Well, 4% of $1.5M is $60,000. To find what 4% of $1M is, adjusted for inflation, just plug the number into an inflation calculator. I just did, and $40K in 2007 dollars comes out to about $45,671 in 2014 dollars. Or, $45,877 in 2015 dollars (not sure which figure you should use to calculate your 2015 wd, but in this case they're close).

So, in this hypothetical scenario, if you wanted to take a bonus, you could take $60K if you wanted, instead of ~$46K. At least, that's the way I'm reading into it.

I agree its not very hard. You dont even need a calculator. Nowadays we make a spreadsheet that figure all of this stuff for us.

 

[David1961]

We have mostly taxable investments. I have tracked the after tax income from our portfolio for many years. Taxes create an interesting "income smoothing" effect on taking a % of remaining portfolio. In years where the portfolio has a big jump, taxes tend to be higher because more is paid in cap gains distributions. In years where the portfolio suffers a loss, cap gains distributions drop significantly, even to zero, and there may even be options for tax loss harvesting.

I have noticed this as well. Plus, after a real good year, if you end up owing taxes the following April, somehow I feel like I should add this to the previous years expenses, not this years. Although if you track your distributions and pay estimated taxes accordingly, it should not be a big surprise.

 

[GTFan]

For this reason I prefer the simple approach of % of remaining portfolio. So easy - X% of Dec 31 value each year.

How do you factor in taking divs and cap gains throughout the year instead of reinvesting? Your WR % calc is then a little more complicated because you don't know how much of that % will be in your distributions. I guess you're simply deducting what you got this year from the % when you withdraw in January of next year?

For me it's going to be a fair bit more complicated anyway because I'm trying to limit my cap gains to stay under the 200% FPL threshold to maximize ACA subsidies. Subject to the Supremes letting me take advantage of them, of course. I don't think I'll ever be using a standard WR % of port as long as that's the case.

 

[audreyh1]

I have noticed this as well. Plus, after a real good year, if you end up owing taxes the following April, somehow I feel like I should add this to the previous years expenses, not this years. Although if you track your distributions and pay estimated taxes accordingly, it should not be a big surprise.

Yeah, most of it gets paid ahead in estimated taxes - at least for 3 quarters. But the remainder, plus set aside for current year's estimated taxes is the first thing that gets pulled out after I take the withdrawal. What remains is what we get to spend.

 

[audreyh1]

How do you factor in taking divs and cap gains throughout the year instead of reinvesting? Your WR % calc is then a little more complicated because you don't know how much of that % will be in your distributions. I guess you're simply deducting what you got this year from the % when you withdraw in January of next year?

For me it's going to be a fair bit more complicated anyway because I'm trying to limit my cap gains to stay under the 200% FPL threshold to maximize ACA subsidies. Subject to the Supremes letting me take advantage of them, of course. I don't think I'll ever be using a standard WR % of port as long as that's the case.

No factor. Any distributions paid out during the year remain part of the portfolio until Jan 1(2) so they are included in the Dec 31 portfolio value. They usually aren't reinvested and sit in cash, but they aren't withdrawn either. No portfolio withdrawals during the year.

 

[Options]

I posted an example of using SWR with a deferred income stream. See (B) here:

http://www.early-retirement.org/for...f-maximize-my-ss-com-77660-6.html#post1606094

Yes, I saw that and I should have added YMMV to my post. In my case SWR is useless as I have enough assets today with another lump sum income increase in around 5 years to delay SS until 70. I am using a suggestion I saw elsewhere to monitor WR until 70 via FIDO. Will also use ESPlanner monte carlo mode and upside modes to monitor PF health as well. Could change/might change depending on circumstances. At this point, not far enough into decumulation to firmly commit to any path. SWR will probably become more relevant after last income stream of SS at 70 comes online.

 

[clifp]

Yes, my comments assume the regular, inflation adjusted SWR.

I should have said "The greater of 4% of your inflation-adjusted initial fund or 4% of your current fund"

That does seem like a pretty reasonable rule. I'm curious how often the excess withdrawal early on would hurt the long term sustainability.
So for instance somebody retiring in 98 would have a two years of increased spending followed by three years significantly decreased spending I wonder if they'd be almost as bad off as the Y2K retiree?

If have been toying with this idea.
Withdrawal rate (WR)= 4% of current portfolio.
IF WR > 110%* last years WR THEN WR= 110% of last year WR
IF WR < 95% * last years WR THEN WR = 95% of last years WR

Now this requires some significant belt tightening early on if you retire into a bear market, but also recognize that not many of us need/want to increase spend say 25% year over year like we would have seen in 2013.

Comments?

 

[audreyh1]

That does seem like a pretty reasonable rule. I'm curious how often the excess withdrawal early on would hurt the long term sustainability.
So for instance somebody retiring in 98 would have a two years of increased spending followed by three years significantly decreased spending I wonder if they'd be almost as bad off as the Y2K retiree?

If have been toying with this idea.
Withdrawal rate (WR)= 4% of current portfolio.
IF WR > 110%* last years WR THEN WR= 110% of last year WR
IF WR < 95% * last years WR THEN WR = 95% of last years WR

Now this requires some significant belt tightening early on if you retire into a bear market, but also recognize that not many of us need/want to increase spend say 25% year over year like we would have seen in 2013.

Comments?

you don't have to spend the "raise" all in one year.

 

[big-papa]

I may not have interpreted the rule correctly, but this is what I think he's saying.

Suppose my portfolio at the end of year 8 is 140% of the inflation-adjusted original portfolio.
And, my portfolio at the end of year 9 is 150% of the inflation-adjusted original portfolio.
Then I take a bonus of X% of the 10% growth. If X=10, I'll get a bonus equal to 1% of my original portfolio. A nice bonus compared to my 4% SWR.

Now suppose that at the end of year 10, my portfolio is again 150% of the inflation-adjusted original portfolio.
I get no bonus, because there is no growth.

But, it seems to me that 150% is enough cushion to justify a bonus. For example, if my portfolio just happens to stay at 150% in each following year, I'll never get another bonus.
Basing the extra payouts on one-year growth creates (IMO) unnecessary volatility in the extra payouts.

But, I'd have to build a model to compare carefully.

Built a model a while back (time to revisit it), but for Gummy's I don't think it was based on the growth relative to the original portfolio - just the last year's growth. Here's a quote from the site:

"I'm suggesting that after we've withdrawn, say 3% for example (to pay the bills), we check to see if our portfolio has increased since a year ago (even after withdrawing that 3%) and, if it has increased by at least the inflation rate, we withdraw some more ... "

Now I think what you described above is a nice buffer, especially if you want to start with a higher withdrawal rate than what Gummy suggested - that is, you don't start even applying the bonus calculation until/if your portfolio has grown for a few years. I think Gummy's method started off with the lower withdrawal rate instead. He treated it as a bonus, much like some of us who are still w*rking have. You (portfolio) do well you get a bonus. If not, wait till next year. The idea is to never count on the bonus as being necessary to cover basic expenses, but to truly treat it as some extra cash to do with as you will (including not withdrawing it).

 

[big-papa]

Reading this interesting paper did remind me a lot of the twins paradox. Especially because if your assets have increased year over year, your life expectancy has also decreased. So if 62 year with 1 million dollars in assets and 4% (40K) withdrawal who find herself with 1.1 million at age 63 logically should be more likely to be able to withdraw 4% (44k) without running out of money in her life than a 62 year old taking 4%.

I suspect that 50% figure before adjusting the withdrawal is almost certainly much too conservative.

The more I think about withdrawals and observe how real world retirees react, the less important I think your initial starting portfolio should be on your current withdrawal.

I am curious how many people who have been retired lets 5+ actually based their current withdrawal on their initial portfolio? Heck I don't even have records of what my portfolio was until 2 years after I retired.

You touched on something I've always wondered about in the retire again and again method. Say you're looking at a 30 year retirement. If you go through the calculation and decide you want to "reset" why would you end up with the same % withdrawal vs. something slightly higher. After all, you now have 29 years because you're one year closer to death. I don't mean that literally because few of us know exactly how many years we have left. But the point is all else being equal, shorter withdrawal periods should result in higher safe withdrawal rates.

 

[big-papa]

For this reason I prefer the simple approach of % of remaining portfolio. So easy - X% of Dec 31 value each year.

I never liked the idea of inflation adjusting from the initial portfolio value. If the portfolio keeps up with inflation, great, if not, annual withdraw will stay behind until the portfolio catches up. If the portfolio zooms ahead of inflation, then I'd like to withdraw the increase sooner rather than later.

I have no problem at all with variability in annual income. IMO this issue is way blown out of proportion and makes the withdrawal calcs much more complex for early retirees. If someone's budget is so tight that they can't handle the variability, then early retirement may not be the best choice. If someone is retiring at a normal age and the budget is tight, then I can see opting for the Trinity Study method to avoid variability.

Yep, I know a lot of people who use % of remaining portfolio. An easy twist to that one was discussed here a while back and which showed up in a Scott Burns article last year. His specific example was 6% of your portfolio or 90% of last year's withdrawal, whichever was largest. The 90% helps reduce how far your withdrawal can drop year to year. You can even modify that slightly by changing the 90% to be 90% after inflation if you like, which will can reduce the "effective" year-to-year drops even further. You can of course play around with both the 6% or 90% number and see what would have happened historically.

 

[clifp]

So I played around with this idea a bit. ******** has some but not all of the capabilities to model, so I had to use a spreadsheet.
I took data from Raddr poor Y2K retiree (same as me) with 75% of his assets in the S&P 500 index in 25% in 3 month tbills. This is is someways a worse case analysis, 75% in Vanguard Total Stock Market and say 20% in Total Bond 5% in T-bills would significantly better although probably not good enough to avoid running out of money in 30 years (much less 40 years for an early retiree)


The withdrawals were calculated to be the 4% of the end of year portfolio value, with a floor of 95% of the previous years+ inflation and ceiling of 10% of the previous years withdrawal. (I probably should add inflation)
Note that your initial start portfolio value is irrelevant to your withdrawal.

This table shows the nominal withdrawals and portfolio values.

1999   0  $1,000,000 
2000 $40,000   $909,360       
2001 $39,292   $799,375 
2002 $38,373   $637,340       
2003 $37,037   $730,118
2004 $35,995   $752,256 
2005 $35,118   $748,871
2006 $34,497   $806,886 
2007 $33,820   $813,265       
2008 $35,207   $564,870       
2009 $33,480   $637,269       
2010 $32,665   $672,320       
2011 $31,373   $647,677
2012 $30,818   $690,204 
2013 $29,804   $825,005       
2014 $32,784   $873,760

As you can see the Y2K retiree starts withdrawing less and less money before getting a pay raise in 2008, and then 2014. The low point 2013 is 26% below 40K in nominal terms..

Including inflation the 2013 withdrawal is only $22,038 which is significant drop from 40K. The portfolio value at the beginning of 2015 is $636K in 2000 $. Which is a big improvement from $390K real for the Retiree blindly following the Trinity study methodology, a withdrawal rate of >10% for the next 16 years seems very likely that blindly following the 4% SWR methodology you'd go broke in the next dozen years.

 

[audreyh1]

Yep, I know a lot of people who use % of remaining portfolio. An easy twist to that one was discussed here a while back and which showed up in a Scott Burns article last year. His specific example was 6% of your portfolio or 90% of last year's withdrawal, whichever was largest. The 90% helps reduce how far your withdrawal can drop year to year. You can even modify that slightly by changing the 90% to be 90% after inflation if you like, which will can reduce the "effective" year-to-year drops even further. You can of course play around with both the 6% or 90% number and see what would have happened historically.

I am comfortable with an occasional big drop in income. We already don't spend all we withdraw, so the excess is set aside for that rainy day or a splurge, etc.

 

[David1961]

I am comfortable with an occasional big drop in income. We already don't spend all we withdraw, so the excess is set aside for that rainy day or a splurge, etc.

Audrey, I apologize if I have asked you this before. Say at the end of the year, you spend $10k less than you withdraw. Is that $10k not counted as part of your portfolio anymore? I have always thought of the withdraw as the money that you spend, so if you spend less in a certain year, your withdraw is lower.

 

[ERD50]

So I played around with this idea a bit. ******** has some but not all of the capabilities to model, so I had to use a spreadsheet.
I took data from Raddr poor Y2K retiree (same as me) with 75% of his assets in the S&P 500 index in 25% in 3 month tbills. This is is someways a worse case analysis, 75% in Vanguard Total Stock Market and say 20% in Total Bond 5% in T-bills would significantly better although probably not good enough to avoid running out of money in 30 years (much less 40 years for an early retiree)


The withdrawals were calculated to be the 4% of the end of year portfolio value, with a floor of 95% of the previous years+ inflation and ceiling of 10% of the previous years withdrawal. (I probably should add inflation)
Note that your initial start portfolio value is irrelevant to your withdrawal.

This table shows the nominal withdrawals and portfolio values.

1999   0  $1,000,000 
2000 $40,000   $909,360       
2001 $39,292   $799,375 
2002 $38,373   $637,340       
2003 $37,037   $730,118
2004 $35,995   $752,256 
2005 $35,118   $748,871
2006 $34,497   $806,886 
2007 $33,820   $813,265       
2008 $35,207   $564,870       
2009 $33,480   $637,269       
2010 $32,665   $672,320       
2011 $31,373   $647,677
2012 $30,818   $690,204 
2013 $29,804   $825,005       
2014 $32,784   $873,760

As you can see the Y2K retiree starts withdrawing less and less money before getting a pay raise in 2008, and then 2014. The low point 2013 is 26% below 40K in nominal terms..

Including inflation the 2013 withdrawal is only $22,038 which is significant drop from 40K. The portfolio value at the beginning of 2015 is $636K in 2000 $. Which is a big improvement from $390K real for the Retiree blindly following the Trinity study methodology, a withdrawal rate of >10% for the next 16 years seems very likely that blindly following the 4% SWR methodology you'd go broke in the next dozen years.

Interesting, but a couple points -

A) It's not apples to apples, as you draw significantly less with this plan - so of course your portfolio does better! Average $34,684 versus $40,000.

B) The 4% WR fails 5% of historic periods, so the speculation that the Y2K period will end up looking like one of the 5% failures isn't that unexpected. And cutting your withdraw amount obviously helps - but is a variable cut significantly better/worse than starting at a lower amount?

I'd be more interested in comparing a 100% Historically Safe Withdraw Rate (HSWR) with an adjustment method. If you went with HSWR and the 'retire again & again' plan, you would be able to increase spending if things turn around, but if this truly a very-bad-case period, probably no increase.

It just seems to me that anyone who can cut up to 25% of their budget, and cut by over 17% on average for a full ten years must have a lot of discretionary spending planned, and is willing to give up that discretionary spending for a very long time. Of course, anything above basic food, water, shelter is discretionary, so this is all a matter of degrees.

Of course the flip side is starting with the lower HSWR is taking a cut right from the start. No magic, just numbers. I personally prefer to look at the lower HSWR amount, budget for that with a higher confidence that I will likely never need to figure out what to cut. But that's just me.

-ERD50

 

[audreyh1]

Audrey, I apologize if I have asked you this before. Say at the end of the year, you spend $10k less than you withdraw. Is that $10k not counted as part of your portfolio anymore? I have always thought of the withdraw as the money that you spend, so if you spend less in a certain year, your withdraw is lower.

Once money is is withdrawn it no longer counts as part of the portfolio, regardless of whether I spend it that year.

 

[Meadbh]

I am comfortable with an occasional big drop in income. We already don't spend all we withdraw, so the excess is set aside for that rainy day or a splurge, etc.

If you set aside the excess outside your portfolio, isn't that sort of a bucket? Or at least a bowl?

I had an excess from last year. It is in a HISA and I am applying it to this year's expenses. But I still consider it part of my portfolio, just part of the cash allocation.

 

[Options]

I'm no longer as comfortable basing retirement calculations on any historical safe withdrawal rates as a result of current research such as this:

Safe Withdrawal Rates for Retirement and the Trinity Study

Yes, history is all we basically have to go on, but given the current investing environment it's a good idea to be conservative. According to Firecalc/******** I'll die very rich. Maybe. Maybe not. I'm certainly going to assume the worst case scenario in all planning.