Understanding SWR, SS, etc?

[Options]

Quote:

Originally Posted by sengsational I am confused by lumping after tax and traditional IRA money together...they are such different beasts. ....

ESPlanner does not lump them together, but allows you to model withdrawing after tax or tax-deferred first and see which is superior. This is what I like about that software, it allows you to model many, many different scenarios.

Regarding iorp, I'm not comfortable with calculators asking me to assume a rate of return. I have no idea, and I know of no one that does either. I will say I have taken into account Bogle's estimation that real returns for the next decade could be zero (notwithstanding p/e changes), and modeled that in ESPlanner.

I use the most pessimistic scenarios for all planning, and have found Pfau's SWR calculations (as of 4/15 according to his blog) to be the most conservative, followed by ESPlanner in monte carlo and then followed by its upside mode (what I call the armageddon scenario as it calculates your equity holdings lose all value), followed then by FIDO's "poor returns" scenario. I've found firecalc and ******** to be the most optimistic as they use historical returns.

[pb4uski]

Just because i-orp is deterministic doesn't mean that it is less relevant to developing a tax-efficient withdrawal strategy. I highly doubt that a tax-efficient withdrawal strategy that is optimal on a deterministic basis would end up being suboptimal on a stochastic basis.

[Options]

Quote:

Originally Posted by pb4uski Just because i-orp is deterministic doesn't mean that it is less relevant to developing a tax-efficient withdrawal strategy. I highly doubt that a tax-efficient withdrawal strategy that is optimal on a deterministic basis would end up being suboptimal on a stochastic basis.

It certainly would if one assumed overly aggressive returns for the deterministic model. In fact, I wouldn't know how to correlate the two: one assumes a fixed number and the other is random. How does one extrapolate to the other?

Regarding deterministic calculators, I'm just not comfortable guesstimating future returns. It's been said that "no one knows nuthin'", to which I'll add least of all me when it comes to estimating what returns in the future will look like. All calculator outputs are subject to their inputs, which are in turn subject to (behavioral) bias. One can manipulate any calculator and come up with support for a given strategy. I simply want to minimize (as much as possible) that bias.

Otar discussed his issues with MC calculators which were in fact addressed by someone (my apologies for failing to remember who!), so I'm simply personally more comfortable with MC calculators. I should clarify that while I prefer them to deterministic calculators, YMMV. As we all know, no calculator is perfect and I agree with the above recommendation to recalculate/recalibrate every year.

[dallas27]

Quote:

Originally Posted by marko Gee. I thought I had earned that money myself. Didn't see anyone else around while I was working so hard for it.

Really, no police to safeguard your travel to work? No fireman to protect your assets from fire? No roads or rail to allow you to work more than walking distance from home? No military to stabilize your government? So on...


Sent from my iPhone using Early Retirement Forum

[marko]

Quote:

Originally Posted by dallas27 Really, no police to safeguard your travel to work? No fireman to protect your assets from fire? No roads or rail to allow you to work more than walking distance from home? No military to stabilize your government? So on... Sent from my iPhone using Early Retirement Forum

I think you missed the point.

[pb4uski]

Quote:

Originally Posted by Options It certainly would if one assumed overly aggressive returns for the deterministic model. ...

Really? I never would have guessed that. You could use overly aggressive return and variability assumptions for a stochastic model too.

Your favorite stochastic calculator implicitly includes an expected average return that is also an educated guess... you could find it and then use that in i-orp for the purpose of testing different withdrawal strategies.

Since you provide i-orp with a return assumption, if you manipulate it to support a given strategy, then it's your own fault.

I prefer deterministic calculators because they are easier to understand various alternatives, but I concede a need to the supplement the plan with a stochastic analysis to assess sequence of returns risk.

[dallas27]

Quote:

Originally Posted by marko I think you missed the point.

I guess, i thought your point was that you earned all the money entirely on your own.


Sent from my iPhone using Early Retirement Forum

[marko]

Quote:

Originally Posted by dallas27 I guess, i thought your point was that you earned all the money entirely on your own. Sent from my iPhone using Early Retirement Forum

Please, I was actually there in the room at the time of the "you didn't build it" lecture, but let me spell it out for you.

My point was that I don't mind paying for police, fire, roads and military.

I do mind when it's explicitly stated that the money I've earned "never was mine in the first place" and somehow 'belongs' to someone else. I'll pay, but I'm paying with my money.

As I posted earlier, it reminds me of the local DPW worker who once got indignant and told me that "I'm not paid by you, I'm paid by the city".

Hope this clarifies it for you.

[Options]

Quote:

Originally Posted by pb4uski Really? I never would have guessed that. You could use overly aggressive return and variability assumptions for a stochastic model too. Your favorite stochastic calculator implicitly includes an expected average return that is also an educated guess... you could find it and then use that in i-orp for the purpose of testing different withdrawal strategies...

What?? No you [the user] cannot use "overly aggressive and variability assumptions" for a stochastic model. Any stochastic model, favorite or otherwise, uses randomly generated data*, See this:

Retirement calculators and spending - Bogleheads

Specifically:

Quote:

Calculator Type Abbreviations. These are used to categorize the calculator's approach to estimating future growth in retirement savings. Det = Deterministic. The calculator uses fixed, user selected future stock and bond returns. HRet = Historical Returns. The calculator uses historical stock and bond returns as a guide. HStr = Historical stress test. The calculator uses historical U.S. stock and bond returns (and sometimes the corresponding inflation) starting from one or more specific years in the past to Stress Test the retirement plan for worst case scenarios. MC = Monte Carlo. The calculator uses a Monte Carlo model of stock and bond investment returns as a guide.

Again, no calculator is perfect. The extent to which any of us seek certainty in any of these tools is the extent to which we'll never find it. Which is why I agree with previous suggestions to update your favorite calculator, spreadsheet, or other tool annually. Then trust in God, but tie up your camels.

Edit: *It's actually more detailed than this (see Otar's discussion in Unveiling the Retirement Myth).

[daylatedollarshort]

We use 0% real returns in a spreadsheet, hope for 1% and anything more will be party time. The zero real return lets us draw down up 2.5% over 40 years, though our planned retirement budget does not call for spending that much on top of our other retirement income streams. I also use the Fido RIP for the income tax projections.

For this year and probably next we are choosing ACA subsidies over Roth conversions. We just view the extra income taxes on RMDS at age 70+ as a nice problem to have. We'll still be paying less in income taxes in inflation adjusted dollars most retirement years compared to when we were both working full time.

[pb4uski]

Quote:

Originally Posted by Options What?? No you [the user] cannot use "overly aggressive and variability assumptions" for a stochastic model. Any stochastic model, favorite or otherwise, uses randomly generated data....

While it is true that most stochastic models do not allow user provided inputs some do (firecalc being an example of one that does). It then calculates a return to be applied to each projection year from those two assumptions and a random number generator.

So using firecalc as an example, if you input an expected return of 5% and 10% variability and leave all else default, you'll get a low success rate each time you hit "Submit", if you change the expected return to 10% then you'll get a high success rate each time you hit "Submit" and if you change the expected return to 15% you can hit "Submit" until you are blue in the face without getting a failure.

While in most MC simulators you can't change the MC assumptions like you can with firecalc and some others, it doesn't mean that there are not expected return or variability assumptions there, it is just that you can't see them or change them. There need to be some parameters programmed for the calculator to generate a random return to be applied to that year's cash flows.

[Options]

Quote:

Originally Posted by pb4uski While it is true that most stochastic models do not allow user provided inputs some do (firecalc being an example of one that does). It then calculates a return to be applied to each projection year from those two assumptions and a random number generator. So using firecalc as an example, if you input an expected return of 5% and 10% variability and leave all else default, you'll get a low success rate each time you hit "Submit", if you change the expected return to 10% then you'll get a high success rate each time you hit "Submit" and if you change the expected return to 15% you can hit "Submit" until you are blue in the face without getting a failure. While in most MC simulators you can't change the MC assumptions like you can with firecalc and some others, it doesn't mean that there are not expected return or variability assumptions there, it is just that you can't see them or change them. There need to be some parameters programmed for the calculator to generate a random return to be applied to that year's cash flows.

Firecalc provides for 3 types of calculations: historical, deterministic, and monte carlo, and one is not to be confused with the other. I have never used the deterministic model in FC and never would based on behavioral concerns I stated above.

Expected return and variability assumptions are not present in MC simulations. See this:

https://en.wikipedia.org/wiki/Stochastic

Specifically:

Quote:

Researchers refer to physical systems in which they are uncertain about the values of parameters, measurements, expected input and disturbances as "stochastic systems". In probability theory, a purely stochastic system is one whose state is randomly determined, having a random probability distribution or pattern that may be analyzed statistically but may not be predicted precisely. In this regard, it can be classified as non-deterministic (i.e., "random") so that the subsequent state of the system is determined probabilistically. Any system or process that must be analyzed using probability theory is stochastic at least in part.[1][2] Stochastic systems and processes play a fundamental role in mathematical models of phenomena in many fields of science, engineering, and economics.

[Emphasis added]

Edit: Perhaps my issue is with your use of the words "expected" and "assumptions". Sure, some data is basic to any calculator, but I see nothing "expected" or "assumed" in MC calculators as any data generated is, again, random (FC being an example of an exception, which again, I would not personally feel comfortable using).

[pb4uski]

Quote:

Originally Posted by Options ....Expected return and variability assumptions are not present in MC simulations. ...

The Bogleheads link that you provided earlier indicates in footnote 13 that "The Monte Carlo capability in FIRECalc 3.0 is limited. Only a mean return and corresponding standard deviation can be entered, and these are applied to the entire investment portfolio."

Is that not a Monte Carlo calculation that utilizes an expected return and variability assumption? If not, you should write to bogleheads and straighten them out because they obviously think that a Monte Carlo calculator can have expected return and variability inputs.

I think the reason they view it as limited is because one much define these for the entire portfolio rather than for asset classes.

[Options]

We are talking past each other. As I stated above, I do not use the deterministic or MC capabilities in FC, precisely because they are limited. The historical returns calculations are more informative for my purposes. YMMV.

[sengsational]

Quote:

Originally Posted by pb4uski The Bogleheads link that you provided earlier indicates in footnote 13 that "The Monte Carlo capability in FIRECalc 3.0 is limited. Only a mean return and corresponding standard deviation can be entered, and these are applied to the entire investment portfolio." Is that not a Monte Carlo calculation that utilizes an expected return and variability assumption? If not, you should write to bogleheads and straighten them out because they obviously think that a Monte Carlo calculator can have expected return and variability inputs. I think the reason they view it as limited is because one much define these for the entire portfolio rather than for asset classes.

Thanks for picking up this line of logic...you have done a better job at it than I could have.

I agree that there is an assumption of return in every model. Fircalc assumes the future return will be the same as some past set of years. And theres no magic in a monte carlo model, it just tries a whole bunch of rates, but that set of rates is defined beforehand.

[rayvt]

Quote:

Originally Posted by marko Agreed. Just found it an interesting perspective that my tax bill isn't my money and "never was".

I don't disagree with you, but that isn't the issue. The IRS disagrees with both of us, and they have the guns -- so their opinion prevails.

Try, for example, not sending in your quarterly tax payment. My wife's old company tried that. One day she went in to work and the IRS had seized the company and put a chain and padlock on the door.

In their opinion ~15%-25% of your paycheck belongs to them from the get-go and you'd better hand it over. So as a practical matter (not as a moral matter!) that part of your paycheck was never yours, it was theirs from the beginning.

[marko]

Quote:

Originally Posted by rayvt I don't disagree with you, but that isn't the issue. The IRS disagrees with both of us, and they have the guns -- so their opinion prevails. Try, for example, not sending in your quarterly tax payment. My wife's old company tried that. One day she went in to work and the IRS had seized the company and put a chain and padlock on the door. In their opinion ~15%-25% of your paycheck belongs to them from the get-go and you'd better hand it over. So as a practical matter (not as a moral matter!) that part of your paycheck was never yours, it was theirs from the beginning.

Yeah. I guess it's about context. It's the same thing either way, but I still like to think of ALL of it as "my money", a portion of which I graciously contribute toward the common good.

[pb4uski]

Quote:

Originally Posted by rayvt ....In their opinion ~15%-25% of your paycheck belongs to them from the get-go and you'd better hand it over. So as a practical matter (not as a moral matter!) that part of your paycheck was never yours, it was theirs from the beginning.

I'm not sure if you really believe the poppycock you are posting or whether you're just trying to pull our leg. I suspect it is the latter because you seem like a pretty smart guy.

Many people are exempt from withholding because their income is insufficient to result in a tax liability so in those cases there is no amount that belonged to the IRS from the beginning as you claim and 15-25% does not belong to them as you claim.

Similarly, the IRS does not have a claim on tax-deferred funds as you suggest. When you withdraw those funds in a way that is a taxable event then they expect it to be included in your tax return and pay any taxes that are due. In many cases, you income might be low enough that you own no tax on that income so any suggestion that they have a claim is just plain silly.

If you do know better, please desist your rhetoric as you will simply confuse people who don't know the difference.

[sengsational]

Yep, I always figured taxes were owed when they are due, not before. Even withholding rules can be safely broken if you ard willing to pay a penalty.

[rayvt]

Quote:

Originally Posted by pb4uski I'm not sure if you really believe the poppycock you are posting or whether you're just trying to pull our leg. I suspect it is the latter because you seem like a pretty smart guy. Many people are exempt from withholding because their income is insufficient to result in a tax liability so in those cases there is no amount that belonged to the IRS from the beginning as you claim and 15-25% does not belong to them as you claim.

Yup. Something like 42% of the people, so I hear. And approximately ZERO of them are reading about Financially Independent and Retiring Early. And they aren't mulling over Withdrawal Ratios, they are more likely to be asking how does a ratio apply to $0.

Quote:

Similarly, the IRS does not have a claim on tax-deferred funds as you suggest. When you withdraw those funds in a way that is a taxable event then they expect it to be included in your tax return and pay any taxes that are due. In many cases, you income might be low enough that you own no tax on that income so any suggestion that they have a claim is just plain silly. If you do know better, please desist your rhetoric as you will simply confuse people who don't know the difference.

I assume the audience here knows that there are may tax rates and cutoff thresholds, and they know that the rates & thresholds change all the time. After all this is the "Early Retirement & Financial Independence Community". It would be very boring and WallOfWord'ing to continually spell out "15% of your taxable income, minus deductions, on the amount between $18,150 and $73,800 if you are married and between $9,075 and $36,900 if you are single, plus 25% of the amount between $73,800 and $148,850 if you are married or between $36,900 and $89,350 if you are single, plus 28% ...." and then to describe the difference between average tax rate and marginal tax rate.

By the time the reader got to the point of the post, they'd have fallen asleep.

The common tax rates are 15% and 25% -- and people who follow financial sites know this, and they know what their rate is. There is no need to spell out all the intricate details to people who are knowledgable on the subject, to discuss a concept.

Whew!

Okay, so....I'll restate my point. The IRS has a complex formula they use to decide how much of your income belongs to them. Said formula is effectively beyond mortal ken. Nonetheless, they decide what portion of the money you get is theirs. They also decide the timing at which you'd better fork it over. Doesn't matter what you think of their opinion. They have the guns, so their opinion prevails.

One can rail at this all ones wants. But it doesn't have any practical effect. In practical terms, a portion of what you think of as "your" money is deemed by the IRS to be "their" money, which you just happen to be holding for them.

IMHO, it is psychologically easier to accept this as a fact of life which you can do nothing about, and deal with it that way.

My original post was in response to someone bemoaning his IRA being held "tax hostage". But it isn't. Some of that IRA is money that he should have paid to the IRS years ago, but they magnanimously said, "Nah, you don't have to pay up now, you can hang onto it for us in an IRA and pay us later on when you withdraw money from the IRA."

A portion of the money in your IRA is not yours, it's the governments. In fact, approximately [...insert long details about marginal tax rates, deductions, taxable income, etc. here...] 15% or 25% of it belongs to them. Just because you held on to it for a few decades, and you came to think of it as all yours doesn't change this. It was always theirs; it was never yours.

It's as if you paid the entire income tax you owed and the IRS then deposited some of it in an account of your choosing, with the proviso that you had to give it back later.