Another Study Challenging the 4% SWIP Rule

[ERD50]

Quote:

Originally Posted by haha Does anyone still think that these "studies" prove anything? Ha

Very clearly. They prove that there are diverse set of views on the issue

Quote:

Originally Posted by Moemg If I had not used a variable withdrawal rate during the meltdown my rate would have been as high as 7% . I use a straight 3.5% of my yearly portfolio value and so for it has worked out fine . I also have a lot of padding in my budget .

Quote:

Originally Posted by REWahoo Remember that the textbook measurement of WR is based the value of your portfolio the day you began to withdraw from it, not the current or mid-market-meltdown value. That said, I'm in no way advocating ignoring the harsh reality of a diminished portfolio and the prudent reaction to pull back on spending.

Yes - recall that a bunch of the 'success' lines in FIRECALC dip down to ~ 1/2 the original portfolio value, so the WR was going to 8%, and they survived.

Like REWahoo, I'm not advocating one way or the other, but that is just what has happened historically. The 4% WR is including those periods of higher withdraws until a recovery brings things back in line again.

At some point, I'd cut back too. I'm not sure what that point is, but my NW is higher than when I retired, and I haven't touched SS or pension yet (but DW has a moderate income). In 2009, NW dropped below my starting point, and that got my attention, but I didn't make any changes.

-ERD50

[Nords]

Quote:

Originally Posted by youbet The concept of spending rates vs. withdrawal rates is widely misunderstood............. The authors were pretty clear that they're talking about withdrawal rates.

Well, now that you mention it, it's certainly misunderstood by me.

Can you clarify how you differentiate the two, and why the authors had to make it significant to this study?

[youbet]

Quote:

Originally Posted by Nords Well, now that you mention it, it's certainly misunderstood by me. Can you clarify how you differentiate the two, and why the authors had to make it significant to this study?

Nords, I think of a WR as the percentage you remove from your FIRE portfolio annually. If your portfolio is $1mil and you withdraw $40k, your WR is 4%. If you spend exactly that $40k, then your spending rate is 4% also.

But wait, there's more.......

If you have another source of income, then you might withdraw at 4% but spend at a higher percentage. For example, in the above example but with $20k of SS, you could spend at a rate equal to 6% of your portfolio but while only withdrawing at 4%. That is, you could spend the $40k (4%) portfolio withdrawal and the $20k SS. To your FIRE portfolio, it's the actual withdrawal from the portfolio that matters as far as "survival" goes. Other sources of spending money such as SS, pensions, rents, part time work, spouse's income, annuities, gambling winnings, gifts, inheritances, etc., don't matter.

The FireCalc calculator inadvertently adds to the confusion because it goes beyond testing WR's and includes other sources of income and tests against spending levels (making the assumption you spend all that you withdraw).

The spreadsheet version of the calculator on the Retire Early site tests only WR's and sometimes makes things clearer for me since other sources of income are not considered.

Anyway..... talking about survival rates with a 4% inflation adjusted WR and then mixing in miscellaneous other sources of income makes comparisons impossible. And determining if your FIRE portfolio will support you in retirement vs your FIRE portfolio + some other sources of income supporting you (surviving) are two different questions.

In the example above with a $1mil portfolio and $20k of SS, to spend $60k you'd withdraw $40k (4%), add your SS and have $60k to spend. But you're not withdrawing at 6%. You're spending at 6%. You're withdrawing at 4%.

[Nords]

Quote:

Originally Posted by youbet Nords, I think of a WR as the percentage you remove from your FIRE portfolio annually. If your portfolio is $1mil and you withdraw $40k, your WR is 4%. If you spend exactly that $40k, then your spending rate is 4% also. But wait, there's more....... If you have another source of income, then you might withdraw at 4% but spend at a higher percentage. For example, in the above example but with $20k of SS, you could spend at a rate equal to 6% of your portfolio but while only withdrawing at 4%. That is, you could spend the $40k (4%) portfolio withdrawal and the $20k SS. To your FIRE portfolio, it's the actual withdrawal from the portfolio that matters as far as "survival" goes. Other sources of spending money such as SS, pensions, rents, part time work, spouse's income, annuities, gambling winnings, gifts, inheritances, etc., don't matter. Anyway..... talking about survival rates with a 4% inflation adjusted WR and then mixing in miscellaneous other sources of income makes comparisons impossible. And determining if your FIRE portfolio will support you in retirement vs your FIRE portfolio + some other sources of income supporting you (surviving) are two different questions.

This board continues to amaze me with our ability to continue to come up with new definitions for the same subjects. It probably simultaneously makes new members think "This is way too hard!" while answering their other question of "What will I do all day?"

I understand the distinction now, and I agree that mixing in other sources of income makes comparisons difficult. However the authors are missing the point. They're chasing down a mathematical analysis that's really only of value to their plans for tenure the other six people sharing their niche for stuffing random stock market returns into some sort of model that they can assume represents reality.

An analysis that would have been much more useful to a financial advisor, let alone more realistic, would have been to include the effects of Social Security. At the very least an advisor would include SS in a discussion of whether the retiree wannabe had enough assets, has sufficient margin against a bear market in their first few years of retirement, and their factors in their decision on when to start disbursements.

Heaven forbid that a financial "analyst" will someday come up with a variable-spending retirement calculator that builds on Bob Clyatt's work.

[youbet]

Quote:

Originally Posted by Nords This board continues to amaze me with our ability to continue to come up with new definitions for the same subjects.

I don't think it's so much "new definitions for the same subjects" as it is clarifying terms. Communications are tough when folks have various understandings of terms. Here, the free interchange of terms such as SWR and WR, spending rate and withdrawal rate survival and stability, etc., sometimes lead to folks talking past one another.

Quote:

It probably simultaneously makes new members think "This is way too hard!" while answering their other question of "What will I do all day?"

Could be. I think the biggest source of confusion and "things seeming hard" is when the concept of SWR is freely interchanged with the the concept of retirement budgets vs. sources of income. Dollars withdrawn from a portfolio are only one source of retirement income available to offset retirement expenses.

Quote:

An analysis that would have been much more useful to a financial advisor, let alone more realistic, would have been to include the effects of Social Security. At the very least an advisor would include SS in a discussion of whether the retiree wannabe had enough assets, has sufficient margin against a bear market in their first few years of retirement, and their factors in their decision on when to start disbursements.

Again, the authors were talking about WR's, not total income available to some retirees including SS, pensions, spouse's income, etc.

I'm not defending the authors' work. It's just a paper presenting yet another study projecting hypothetical "safe" withdrawal rates given some set of assumptions and parameters. They didn't try to consider total sources of retirement income, just portfolio withdrawals. All I took away from it was the notion that if future investment returns and inflation results are less favorable than past, the inflation adjusted withdrawals you can safely make will be smaller. What a surprise...........

[Independent]

Quote:

Originally Posted by arebelspy I'd love to see links to any posts you have made on this, especially regarding your methodology. I'm curious what similar rule would correspond to a 3% SWR.

It's easier to write it out than find a link. Again, the idea is that before I retire I decide on a "floor" withdrawal rate. I plan to vary my spending depending on investment returns. I am happy spending more than my floor, but I want don't ever want to spend less than the floor.

I tried a rule which is "The greater of 6% of current balance or 3% (inflation adjusted) of original balance."

I originally tried that on a Monte Carlo system of my own, but that's kind of a black box. So I tried using the FireCalc data. I ran the standard FireCalc (all assumptions at the defaults) and clicked the "provide data and formulas in spreadsheet format" box on the Investigate page.

I moved the data to an Excel spreadsheet and extracted the annual inflation rates and investment returns. (Now that I'm remembering, I think it's easier to get these factors if I set the withdrawal rate to nearly zero.)

Then I set up a worksheet with 108 rows that ran the standard 4% withdrawals using FireCalc's formulas and spot checked the results.

Then I changed the withdrawal rule to the 6%/3% above.

My result was that no scenario ran out of money, but I did find a fair number where the withdrawals did drop to 3%. I tried to summarize the results by looking at statistics like this

Withdrawal-Amount	Yr-5	Yr-10	Yr-15	Yr-20	Yr-25	Yr-30
$45,000+	44	42	43	38	32	24
$30,000-$45,000	52	40	32	33	30	33
$22,501-$30,000	12	14	12	14	18	16
Exactly-$22,500	0	12	21	23	28	35
$00,000-$22,500	0	0	0	0	0	0

I was using FireCalc's initial fund of $750k. So 4% = $30k, 6% = $45k, and 3% = $22,500.

The table says, for example, that I looked at the 108 possible starting years in FireCalc. When I checked the withdrawal amounts in the 15th year in each of those 108 possibilities, I found that in 43 of them I was withdrawing over $45k, in 32 I was withdrawing beteen $30k and $45k, etc. Each column should total to 108.

It looks like the 6% is a little aggressive, based on what I'd like to see. I haven't tried lower numbers, but I'm guessing it wouldn't take much to get a slightly more level payout.

[Animorph]

One of the Merriman papers describes a simple 4% withdrawal rate (4% of portfolio each year, not the usual fixed from the start + inflation) versus a simple 5% withdrawal rate. It took something like 5 to 7 years on average, IIRC, for the 4% rate to start giving withdrawal amounts larger than the 5% rate due to better portfolio growth. So if you can wait that long, a lower rate can let you generate more cash.

[chinaco]

You can tell when you have a bunch of engineers in the room!

[chinaco]

Speaking of spending pattern (rate) and how it relates to the WR... I believe many people will not actually begin at some % and stick with it (nominal or COLA adjusted)

The reason is many have other income sources (e.g., pension, SPIA, SS).

We intend to create a base income using guaranteed income sources (as many of you know from previous posts a SPIA will be used with other sources). Those sources of income will be available at different times (i.e., ages). Assuming our plan does not change for some reason... our WR% and WR$ will reduce from 55 till 70. It is in the 3% range year 1 and drops to around 0% over 15 years (stepped fashion). After that depending on the inflation rate, the WR% will increase. My projections (which could be inaccurate... who knows what will happen) using 3.5% inflation show a rate at age 95 of around 2% (assuming inflation adjusted spending of the planned amount from day 1). In the mortality scenarios (one spouse dies), the spending is adjusted down a little because of reduced expenses. But so do the guaranteed income sources (e.g., Survivor pension amount and 1 SS). IN that scenario the WR% is different than both survive.... but it is still in the low 2% range.

[rescueme]

Quote:

Originally Posted by chinaco You can tell when you have a bunch of engineers in the room!

And the train is at risk of running off the tracks ...

(yes, I did understand your comment - not my reference to a train engineer ).