Kitces Discusses Safe Withdrawal Rate on Mad Fientist Podcast

[walkinwood]

An interesting exercise. I can't see myself giving 1% to an FA. But, if DW is left by herself, she may have to hire one.

 

[Tailgate]

An interesting exercise. I can't see myself giving 1% to an FA. But, if DW is left by herself, she may have to hire one.

That's my situation exactly. DW has zero appetite for $$$ management.

 

[marko]

... if someone pays a 1% fee to an adviser, their SWR goes from 4% to 3%.

Wait a minute.
Isn't your "4%" calculated from your portfolio balance? (That's how I do it).
If you have a balance and you spent X% in fees or whatever, you calculate your SWR from that net number, not the pre-fee number. Or am I missing something?
My MF takes X% as an expense each year. I don't calculate my SWR from the pre-expense number, I calculate it from the net.

 

[mathjak107]

yep , no different than fund expenses . it is already in the net balance when you calculate . you could say if you had that fee money your balance would be that much higher .

but then again maybe your return was higher with the higher expenses and mgmt fees than it would have been without the mgmt .. you really can't do a what if unless everything was exactly the same comparison .

 

[pb4uski]

Wait a minute.
Isn't your "4%" calculated from your portfolio balance? (That's how I do it).
If you have a balance and you spent X% in fees or whatever, you calculate your SWR from that net number, not the pre-fee number. Or am I missing something?
My MF takes X% as an expense each year. I don't calculate my SWR from the pre-expense number, I calculate it from the net.

But, the premise is that with 4% withdrawals and market returns that at 4% you will not run out of money over a defined period of time. Those market returns are either historical like in the case of FIRECalc or random as in the case of Monte Carlo. If you reduce those market returns by a 1% AUM advisor fee then the returns are 1% lower which increases the risk of running out of money unless you make commensurate adjustments to the withdrawl rate.

Also, the 4% is of the beginning portfolio value and withdrawals increase each year for inflation... so if returns exceed inflation as they normally would the withdrawal in subsequent years would be less than 4% of the portfolio balance at the beginning of the year. See example below with 3% inflation and 8% returns:

	Withdrawal w/3% inf	Earnings w/8% return	Balance	% of BOY balance
0			100.00
1	4.00	7.84	103.84	4.00%
2	4.12	8.14	107.86	3.97%
3	4.24	8.46	112.08	3.93%
4	4.37	8.79	116.50	3.90%
5	4.50	9.14	121.14	3.86%

That said, while many people advocate 4% of the balance, that is not what the 4% "rule" is based on. However, while you will never run out of money under that approach, you could end up withdrawing a lot less than you need to live on.

 

[ERD50]

.... (I asked FireCalc to investigate 94.9, but this is what it gave me).

I had the same issue and could not find any way to get it closer to 94.9%.

Spending = 35,800... success rate = 95.7%
Spending = 35,900.... success rate = 94.0%

Right, because of the quantization of success rates (there are only integers in "Success" and "Fail"). The algorithm gets close, but that could be plus or minus either side of the target.

A better way is to tweak the spending for each scenario until you get as close to $zero dollars remaining at the end of the period (just barely 100% success), then compare the spending in % terms (and using $1,000,000 as the default makes this easy, just shift the decimal place).

But the bottom line explanation is that for the failure path, since the portfolio is declining, fees decline as well. So the fees aren't as much a drag as they were initially (1% of $1M) compared to when the portfolio has dropped. But your initial 4% spending is the same amount (in buying power). So the effect of fees drops as the portfolio drops.

But the initial view sounds right in the more typical case of the portfolio roughly maintaining its buying power over 30 years. That's a full 1% that goes to fees, leaving you 3% of a 4% withdraw.

-ERD50

 

[pb4uski]

No... 3% is too low, it is probably more in the 3.5-3.6% range because the 1% AUM affects returns so the impact on WR is a second order effect.

 

[alaska55]

An interesting exercise. I can't see myself giving 1% to an FA. But, if DW is left by herself, she may have to hire one.

That's my situation exactly. DW has zero appetite for $$$ management.

That's what Target Date Retirement funds are for.

 

[medved]

These analyses appear to be premised upon an assumption that an investor's investment returns will not improve at all as a result of the services rendered by an adviser who charges an AUM fee. Not much point debating that point here. But it is at least possible that, for some investors, the advisor's services may improve returns.

 

[YVRRocketSurgery]

I have been retired for 8 years. Let me be brutally honest: there has not been one single moment all this time, in which I thought, "gee, I wish I was working at a low paying part time job where I wasn't getting paid what I feel my time is worth". Not one. For those who think like me, it makes more sense to stick with the high income job a little bit longer so that we aren't stuck with the much lower income job for years and years.

Personally, I'm working a few years longer to ensure I pretty much absolutely don't need work for money.
That said, I can see ER people including myself working a low pay job but a lot of things need to fall into place. The work likely has to be flexible to your own schedule. It's got to be work that you enjoy or have a passion about. Pay ends up being the side benefit which might not be a lot. Heck, why do retired people do volunteer work?

I love sports and wouldn't mind doing some low pay job with one of the local teams. I'd probably take a low paying job where I learned some new cooking techniques.

 

[NW-Bound]

... For our purposes, failure means the portfolio was depleted before the end of the 30 years. FIRECalc found that 5 cycles failed, for a success rate of 95.7%. (I asked FireCalc to investigate 94.9, but this is what it gave me).

It's because of the quantization with only 117 cycles of 30 years to work with.

If FIRECalc increases the WR so that 6 cycles failed instead of 5, the success rate then becomes 1 - 6/117 = 94.87%.

That happens to be just slightly less than your requested 94.9%, so I think FIRECalc gives you the next higher outcome with 5 failed cycles.

 

[Running_Man]

Not really... perhaps a popular misception. You can prove it is false using Firecalc.

40k spending, $1m portfolio, 30 years.... 94.9% success rate
On Portfolio tab, change expenses from 0.18% to 1.18%... success rate declines to 82.9%
Use Investigate tab solve for spending to achieve 94.9% success rate...



So all else being equal, that 1% AUM fee reduces SWR from 4% to ~3.5%.

This seems quite incorrect, so if I were to administer my portfolio myself and charge myself 1% I could therefore from a million dollar portfolio take $35,000 as a 94.9% safe and pay myself $10,000 for a total of $45,000 and that would be the same as if I said forget the fees and just withdrew $40,000. That just seems….wrong

 

[ERD50]

This seems quite incorrect, so if I were to administer my portfolio myself and charge myself 1% I could therefore from a million dollar portfolio take $35,000 as a 94.9% safe and pay myself $10,000 and that would be the same as if I said forget the fees and just withdrew $40,000. That just seems….wrong

The apparent paradox is explained by the fact that the 1% fee isn't constant with your inflation adjusted withdrawals. If you follow a failure curve in FIRECalc, the portfolio is dropping, so the 1% of portfolio fee drops as well. So the scenario you describe is really only true for the first few years.

Regardless, a 1% fee is going to hurt in the worst case, and take even more $ from your allowed spending (or heirs/charities spending) in the cases where the portfolio does not shrink so much.

-ERD50

 

[pb4uski]

This seems quite incorrect, so if I were to administer my portfolio myself and charge myself 1% I could therefore from a million dollar portfolio take $35,000 as a 94.9% safe and pay myself $10,000 for a total of $45,000 and that would be the same as if I said forget the fees and just withdrew $40,000. That just seems….wrong

It is actually correct.... if perhaps not very intuitively obvious. You can easily prove it with a simple deterministic model. Model a 4% withdrawal rate over 30 years with 3% inflation.... to end up with a zero balance you need to have a return of 4.046%. Now reduce the return by 1% to 3.046% and solve for the WR at which you have a zero balance after 30 years... it is 3.457%... reasonably close to the FIRECalc result of 3.5%.

	Withdrawal	Balance
	3.00%	4.046%
0		1,000.00
1	40.00	1,000.46
2	41.20	999.73
3	42.44	997.75
4	43.71	994.40
5	45.02	989.61
6	46.37	983.28
7	47.76	975.30
8	49.19	965.56
9	50.67	953.96
10	52.19	940.36
11	53.76	924.65
12	55.37	906.69
13	57.03	886.34
14	58.74	863.46
15	60.50	837.89
16	62.32	809.47
17	64.19	778.03
18	66.11	743.40
19	68.10	705.38
20	70.14	663.77
21	72.24	618.38
22	74.41	568.99
23	76.64	515.37
24	78.94	457.27
25	81.31	394.46
26	83.75	326.67
27	86.26	253.62
28	88.85	175.03
29	91.52	90.60
30	94.26	(0.00)

	Withdrawal	Balance
	3.00%	3.046%
0		1,000.00
1	34.57	995.89
2	35.61	990.61
3	36.68	984.11
4	37.78	976.31
5	38.91	967.14
6	40.08	956.52
7	41.28	944.38
8	42.52	930.62
9	43.79	915.17
10	45.11	897.94
11	46.46	878.83
12	47.86	857.75
13	49.29	834.58
14	50.77	809.24
15	52.29	781.59
16	53.86	751.54
17	55.48	718.95
18	57.14	683.71
19	58.86	645.68
20	60.62	604.73
21	62.44	560.71
22	64.31	513.47
23	66.24	462.87
24	68.23	408.74
25	70.28	350.91
26	72.39	289.22
27	74.56	223.47
28	76.79	153.48
29	79.10	79.06
30	81.47	-

 

[NW-Bound]

The apparent paradox is explained by the fact that the 1% fee isn't constant with your inflation adjusted withdrawals. If you follow a failure curve in FIRECalc, the portfolio is dropping, so the 1% of portfolio fee drops as well. So the scenario you describe is really only true for the first few years.

Regardless, a 1% fee is going to hurt in the worst case, and take even more $ from your allowed spending (or heirs/charities spending) in the cases where the portfolio does not shrink so much.

-ERD50

Yes. As your portfolio drops due to "withdrawal+fee" the basic WR model in FIRECalc lets you withdraw the same amount (with inflation adjustment), but the 1% fee drops along with the portfolio. At the 30th year, you still have the same annual money to spend right before you expire, but your FA gets diddly-squat.

Look at it this way. It's going to hurt your FA more than it hurts you.

 

[Running_Man]

So I took the data from Raddr's Y2Kportfolio and adjusted for a 3.5% withdrawal and 1% fees, since this is a steady declining portfolio, I was curious to see the difference.

I only updated through 2015 as that is the last time an update was given. Through the 16 years ending 12/31/2015 774 thousand was withdrawn as the 4% Y2K withdrawal and 785 thousand from the 3.5% plus 1% fees portfolio. The withdrawals for 2016 were within 1K of each other at around 55K and the 3.5% plus 1% fees portfolio is 19K lower than the 4% Y2K portfolio at 472K vs 491K.

Both portfolios are withdrawing an excess of 11 percent so unless gains are abnormally high over the next few years both are destined to fail, perhaps at the same point as the 1% forces cutbacks to equalize the income withdrawn over the next few years. So that does confirm what is said, so an intial 4.5% withdrawal rate with 1% being based on the portfolio percentage is apparently the way to spend more money, at least in a failing cause.

 

[pb4uski]

Not sure what Raddr's Y2K portfolio has to do with the issue but if you're happy then I'm happy. I assume you now agree that 4% WR/0% AUM fee approximates 3.5% WR/1% AUM fee.

 

[flintnational]

Good podcast. Thanks for posting.

FN

 

[alaska55]

Not sure what Raddr's Y2K portfolio has to do with the issue but if you're happy then I'm happy. I assume you now agree that 4% WR/0% AUM fee approximates 3.5% WR/1% AUM fee.

He's still beating a dead horse, with the Raddr's Y2K portfolio not being able to handle a 4% withdrawal rate. A 60/40,50/50, and 40/60 Total stock and bond funds from vanguard would be doing just fine.

https://www.portfoliovisualizer.com/backtest-portfolio#analysisResults

Portfolio Returns

 

[urn2bfree]

The one thing that seems to be being glossed over about these 1% AUM fees is how ridiculously high they can become - both as an absolute number but also as a percentage of YOUR spending money.
In the first year 4% is withdrawn and 1% goes to the FA.
Thereafter you no longer are withdrawing a percentage of your portfolio, you are adjusting your withdrawal amount up by inflation...in a good returning portfolio (return above inflation) your withdrawal amount becomes an ever lower percentage. If you follow the paths where you end up with a lot more than your initial value you might only be taking out 3%. But your FA keeps taking 1% of the portfolio. So while he starts out taking 25% of YOUR withdrawal. He may end up taking 33% or more. And if you do really well you might be paying this FA HUGE sums for doing no more work.
Think about this-- You make an appointment with FA with your $1 M portfolio. He says for $10,000 (1%) he will invest 25% here 12% there, etc. etc. you are about to sign the papers to engage his services. Your phone rings. That painting you slaved over was spotted by a buyer who wants to give you $2M for your brilliance and hard work. Congratulations!
And your financial, advisor now expects you to pay him 3 X as much - $30,000 -to engage his services --though he will be doing exactly ZERO more in servicing your account than he was already prepared to be doing. Your account which only grew thanks to YOUR hard work and talents somehow ends up rewarding the FA.

Moral of the story- please shop around. You can easily find flat rate advisors who do absolutely everything these AUM thieves do without skimming so much of YOUR hard earned money. The flat rate fees become less and less of a drag on your portfolio- unlike the AUM scammers. If you are engaging a FA to actively seek out ever new investments then more money does equal more work. (* but there is no evidence that such strategies lead to better returns than a passive investing approach) If your FA is just building a portfolio on passive principles of investing in index funds at certain asset allocation percentages then the work to manage $10 million is the same as $1 million - not worth paying the FA 10 times as much.

 

[pb4uski]

I would amend that.... the moral of the story should be do-it-yourself.... read and learn.... it isn't rocket science and nobody care about your money more than you.. .. hire some help temporarily (fee-only) as you learn the ropes if you need to.

 

[NW-Bound]

In view of the fact that a fee of 1% percentage of the portfolio does not reduce a constant-dollar WR by the same amount, I thought of this idea.

Suppose your WR is composed of two parts: one is a constant dollar amount to represent basic necessities, and another component for discretionary spending and this is a percentage of the present value of the portfolio. Is this not analogous to what people do in practice? During good times, your portfolio grows and you permit yourself to spend more. During bad times, your stash shrinks, and you cut back on luxury items.

So, using the investing fee to model that variable spending (you pay yourself instead of the FA), I made several runs with FIRECalc. The portfolio is set to $1M, with an AA of 50/50. The variable spending component is set at different levels from 0% to 2%. Then, I looked for the maximum constant-dollar WR that results in 100% success over 30 years. The results are as follows.

Case A: $37,939 + 0.0%
Case B: $35,688 + 0.5%
Case C: $33,511 + 1.0%
Case D: $31,406 + 1.5%
Case E: $29,380 + 2.0%

In the last case, you will start out spending $29,380 + $20,000 = $49,380. At the end of the 30 years, in the worst case you will be broke, and the last year of your life you will be spending just $29,380. In contrast, in case A, in the worst case, you still enjoy $37,939 before you croak.

Initial / Worst case

Case A: $37,939 / $37,939
Case B: $42,939 / $35,688
Case C: $43,511 / $33,511
Case D: $46,406 / $31,406
Case E: $49,380 / $29,380

So, this is a form of variable WR. In the worst case, it looks like a milder form of Bernicke's spending model. The difference is that it adapts to the actual portfolio performance as time progresses. If the market is good, you allow yourself to spend more.

So, what happens when it is not the worst case? For Case E, FIRECalc says

The lowest and highest portfolio balance at the end of your retirement was $159 to $3,078,587, with an average at the end of $920,685.
​

This means that on average, at the end of 30 years, you are still spending ($29,380 + 0.02 x $920,685 ) = $47,794. And in the best case, you will be spending ($29,380 + 0.02 x $3,078,587 ) = $90,952.

Well, actually in that best case you can spend a whole lot more than $90,952 out of that $3,078,587 because your days are numbered. It is not likely you will be thinking much about spending at that advanced age.

What this means to me is that if you are willing to cut back your spending to only 3% of your initial portfolio value in bad times, you can spend another 2% on discretionary and will not go broke before 30 years.