Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Non-investing personal finance issues including insurance, credit, real estate, taxes, employment and legal issues such as trusts and wills
User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Tue May 28, 2019 7:21 pm

trueblueky wrote:
Tue May 28, 2019 7:18 pm
Lee_WSP wrote:
Tue May 28, 2019 6:43 pm
fyre4ce wrote:
Tue May 28, 2019 2:37 pm
Where are we with regard to reaching consensus on a definition of marginal? I still favor the "small" change definition, with "average" applying to rates on larger incremental contributions as in the Irregular Marginal Rates example. I believe FiveK and gilgamesh at one time favored the "any size" definition - do they or anyone else still disagree that we should use the "small" definition?

To be clear, when I say small I'm specifically excluding rounding errors and tax table quantization. As has been pointed out, it's not practically possible to optimize taxes around these effects, and over a small range they average out to the "local" marginal rate.
The definition of marginal according to the vast majority of dictionaries is the "additional unit". Each dollar of income has it's own marginal rate. That there are bands of several thousand or tens of thousands of dollars that has the exact same marginal cost/rate/whatever does not affect the base definition, just how we think about it.
In the tax tables, the next "unit" is $50.
Novices need to understand that they cannot add $1 in Turbotax to determine their marginal rate.
That deserves an explanation, not a definition change.

Topic Author
fyre4ce
Posts: 222
Joined: Sun Aug 06, 2017 11:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by fyre4ce » Tue May 28, 2019 7:21 pm

gilgamesh wrote:
Tue May 28, 2019 7:13 pm

Can you post a couple of sources that explicitly says “small” change definition?
Sure. This is what my economics textbook from college says about it:

https://drive.google.com/file/d/1G9iQuZ ... sp=sharing
https://drive.google.com/file/d/1OB0nl6 ... sp=sharing
https://drive.google.com/file/d/1GRXBB2 ... sp=sharing
https://drive.google.com/file/d/1V3wpW1 ... sp=sharing

Here's an online source:

https://www.investopedia.com/terms/m/ma ... axrate.asp

User avatar
FiveK
Posts: 7005
Joined: Sun Mar 16, 2014 2:43 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by FiveK » Tue May 28, 2019 7:22 pm

Lee_WSP wrote:
Tue May 28, 2019 7:14 pm
Are we really turning to blogs to define words?
We are trying to follow your advice and provide operative definitions up front, so when the corresponding terms are used their meaning is unambiguous. Is that ok?

Lee_WSP
Posts: 391
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by Lee_WSP » Tue May 28, 2019 7:23 pm

FiveK wrote:
Tue May 28, 2019 7:22 pm
Lee_WSP wrote:
Tue May 28, 2019 7:14 pm
Are we really turning to blogs to define words?
We are trying to follow your advice and provide operative definitions up front, so when the corresponding terms are used their meaning is unambiguous. Is that ok?
Not when you use a blog as the definitive source when there are actual dictionaries or other more reliable/authoritarian sources that define the term.

User avatar
FiveK
Posts: 7005
Joined: Sun Mar 16, 2014 2:43 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by FiveK » Tue May 28, 2019 7:25 pm

fyre4ce wrote:
Tue May 28, 2019 7:21 pm
Here's an online source:

https://www.investopedia.com/terms/m/ma ... axrate.asp
That source also says "As income increases, what is earned will be taxed at a higher rate than the first dollar earned."

That is not correct, and thus a good example of when "the amount in question" should instead be the denominator term.

User avatar
FiveK
Posts: 7005
Joined: Sun Mar 16, 2014 2:43 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by FiveK » Tue May 28, 2019 7:27 pm

Does that book have examples from the tax code, or any other non-monotonic function?

User avatar
FiveK
Posts: 7005
Joined: Sun Mar 16, 2014 2:43 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by FiveK » Tue May 28, 2019 7:33 pm

Lee_WSP wrote:
Tue May 28, 2019 7:23 pm
FiveK wrote:
Tue May 28, 2019 7:22 pm
Lee_WSP wrote:
Tue May 28, 2019 7:14 pm
Are we really turning to blogs to define words?
We are trying to follow your advice and provide operative definitions up front, so when the corresponding terms are used their meaning is unambiguous. Is that ok?
Not when you use a blog as the definitive source when there are actual dictionaries or other more reliable/authoritarian sources that define the term.
Unfortunately there is no single authoritative definition. E.g.,
- some say "last dollar", while others say "next dollar"
- taking either of those literally, between $3,000 and $100,000 of taxable income 98% of single dollars could have a marginal rate of 0%, while the other 2% have marginal rates of 500%, 600%, 1100%, or 1200%.
- Academic references are most often (always?) in the context of smooth functions whose first derivatives never change sign. This is not the case with taxes.

It may be a fine distinction, but while custom definitions are not useful, operational definitions can be very useful.

That's how the T vs. R and marginal rate wiki articles were combined 3 years ago: the term "marginal tax rate" was used near the top of the T vs. R article, linked to the marginal tax rate article. The use of Kitces' definition made everything consistent.

Striving for practical usefulness is appropriate. As the saying goes, all models are wrong but some [the good ones] are useful.

Topic Author
fyre4ce
Posts: 222
Joined: Sun Aug 06, 2017 11:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by fyre4ce » Tue May 28, 2019 7:38 pm

FiveK wrote:
Tue May 28, 2019 7:16 pm
fyre4ce wrote:
Tue May 28, 2019 4:35 pm
Evidently we still don’t agree. In my view there are two reasons for supporting the “small” definition over “any amount”. One is that it’s by far the most commonly used definition in other sources. The other is that it’s actually a different thing than the rate using “small” changes. Do me a favor and read my post “Buying Bananas” earlier in this thread. That explains by analogy why I think “any amount” is actually a different rate and shouldn’t be confused with the “small” marginal rate.
What makes you think I haven't read it? I have. :)

I agree that “any amount” may be a different rate, but find it to be the "useful" rate for t vs. R purposes.

If you'll return the favor, how do you define "withdrawal_tax_rate" and "contribution_tax_rate" in the simple explanation of how traditional and Roth work?
Sure. If we are talking about a "simple case" where the marginal tax rate is constant over that range, then the tax rates you mention are the marginal rates on tax-deferred savings today and taxable withdrawals in the future.

If we are talking about a "complex case" like the example I worked under the Irregular Marginal Rates section, then the tax rates are different for each set of dollars contributed. For the first $1500 the tax rate is 22%, for the next $500 the tax rate is 43%, etc. I agree that the way to find the optimum contribution involves looking for the highest average contribution_tax_rate, but that doesn't change the fact that the rate is, (literally, in a mathematical sense) an average of the rates in the different bands that make it up. I might be more sympathetic to your proposed definition if contributions could only be made in increments of $6,000 or $0 (or $19/25k for a 401k), but they can be broken up, and each dollar or small set of dollars has its own rate.

Lee_WSP
Posts: 391
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by Lee_WSP » Tue May 28, 2019 7:43 pm

FiveK wrote:
Tue May 28, 2019 7:33 pm
Unfortunately there is no single authoritative definition. E.g.,
- some say "last dollar", while others say "next dollar"
It depends on how the phrase is used. Whether it is looking backwards or forwards. If you've already bought or produced a unit, then it's the "last" unit because that's the unit you're referring to. If you're contemplating purchasing an additional unit, then it's the "next" additional unit.


- taking either of those literally, between $3,000 and $100,000 of taxable income 98% of single dollars could have a marginal rate of 0%, while the other 2% have marginal rates of 500%, 600%, 1100%, or 1200%.
How are you possibly coming to that conclusion? You're using the term wrong.

Perhaps the confusion lies in how people are trying to use the term. The term can only be used when the rate on the additional chunk of dollars is the same. If it is not the same, you have to break up that chunk of dollars into pieces where the marginal rates are the same.

User avatar
FiveK
Posts: 7005
Joined: Sun Mar 16, 2014 2:43 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by FiveK » Tue May 28, 2019 7:46 pm

gilgamesh wrote:
Tue May 28, 2019 6:25 pm
Will the definition be the same for both articles you are writing?
Great question. Can't see how we would choose to do anything other than the same.

Unless we plan to change what the t vs. R wiki has always said, "The main reason to prefer one type of account over the other is the comparison of marginal tax rates," we should use a definition of "marginal tax rate" that makes the t vs. R statement universally true.

Topic Author
fyre4ce
Posts: 222
Joined: Sun Aug 06, 2017 11:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by fyre4ce » Tue May 28, 2019 7:49 pm

FiveK wrote:
Tue May 28, 2019 7:27 pm
Does that book have examples from the tax code, or any other non-monotonic function?
I'm not sure, I'll check later tonight. The book doesn't deal with taxes much; it's about supply, demand, etc. But it's definitely true that in economics, the term "marginal" basically means derivative - the local slope of the curve.

So far as I can tell, no one is denying that the optimum approach when dealing with irregular rates is to find the highest average contribution tax rate and use that in the comparison. What we're contesting is that that rate should be called "marginal" - mathematically, it's not, it's an average of the marginal rates that make it up. I refer any readers who disagree with this statement to my banana example.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Tue May 28, 2019 7:55 pm

Just to remind everyone, what the current boglehead wiki says about the definition

“Marginal tax rate is a tax rate, at the margin, on the last dollar of income.”
https://www.bogleheads.org/wiki/Marginal_tax_rate
It’s reference, a Kitces article :D

The current boglehead wiki on R vs T, always left me with the impression that it was incomplete. Now, I think everyone agrees that it is...
Last edited by gilgamesh on Tue May 28, 2019 7:56 pm, edited 2 times in total.

Topic Author
fyre4ce
Posts: 222
Joined: Sun Aug 06, 2017 11:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by fyre4ce » Tue May 28, 2019 7:55 pm

FiveK wrote:
Tue May 28, 2019 7:46 pm
gilgamesh wrote:
Tue May 28, 2019 6:25 pm
Will the definition be the same for both articles you are writing?
Great question. Can't see how we would choose to do anything other than the same.

Unless we plan to change what the t vs. R wiki has always said, "The main reason to prefer one type of account over the other is the comparison of marginal tax rates," we should use a definition of "marginal tax rate" that makes the t vs. R statement universally true.
I have to point out that this would not be universally true even under your proposed definition. Dealing with maxed out accounts, differences in investment fees/performance, employer matching, tax diversification etc make the decision more complicated.

I think it's equally correct, and more consistent with other sources, to say that when dealing with irregular rates, the average tax/savings rate should be used instead of the marginal. Do you at least agree that this approach leads to the mathematically correct strategy, even if you don't like the terminology?

User avatar
FiveK
Posts: 7005
Joined: Sun Mar 16, 2014 2:43 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by FiveK » Tue May 28, 2019 7:57 pm

Lee_WSP wrote:
Tue May 28, 2019 7:43 pm
- taking either of those literally, between $3,000 and $100,000 of taxable income 98% of single dollars could have a marginal rate of 0%, while the other 2% have marginal rates of 500%, 600%, 1100%, or 1200%.
How are you possibly coming to that conclusion? You're using the term wrong.
Are you familiar with how the Tax Tables work?
Perhaps the confusion lies in how people are trying to use the term. The term can only be used when the rate on the additional chunk of dollars is the same. If it is not the same, you have to break up that chunk of dollars into pieces where the marginal rates are the same.
What do you suggest when the instantaneous marginal rates are non-monotonic?

User avatar
FiveK
Posts: 7005
Joined: Sun Mar 16, 2014 2:43 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by FiveK » Tue May 28, 2019 8:05 pm

fyre4ce wrote:
Tue May 28, 2019 7:55 pm
FiveK wrote:
Tue May 28, 2019 7:46 pm
gilgamesh wrote:
Tue May 28, 2019 6:25 pm
Will the definition be the same for both articles you are writing?
Great question. Can't see how we would choose to do anything other than the same.

Unless we plan to change what the t vs. R wiki has always said, "The main reason to prefer one type of account over the other is the comparison of marginal tax rates," we should use a definition of "marginal tax rate" that makes the t vs. R statement universally true.
I have to point out that this would not be universally true even under your proposed definition. Dealing with maxed out accounts, differences in investment fees/performance, employer matching, tax diversification etc make the decision more complicated.
It holds true with maxed out accounts. The breakeven rate may change, but "the main reason to prefer..." does not change.
There should be no difference in investment fees/performance when the only difference is the type of account from the same provider.
There may be the occasional employer who refuses to match a Roth contribution, so that is true.
Tax diversification is a subjective term, regardless of how one defines marginal rate.
...leads to the mathematically correct strategy, even if you don't like the terminology?
I thought we all agreed on this phrase. :sharebeer

Topic Author
fyre4ce
Posts: 222
Joined: Sun Aug 06, 2017 11:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by fyre4ce » Tue May 28, 2019 8:10 pm

FiveK wrote:
Tue May 28, 2019 7:57 pm
Lee_WSP wrote:
Tue May 28, 2019 7:43 pm
- taking either of those literally, between $3,000 and $100,000 of taxable income 98% of single dollars could have a marginal rate of 0%, while the other 2% have marginal rates of 500%, 600%, 1100%, or 1200%.
How are you possibly coming to that conclusion? You're using the term wrong.
Are you familiar with how the Tax Tables work?
With all due respect, I think tax tables are a red herring; I haven't seen anyone here seriously propose that the quanta of the tax tables be accounted for when calculating a marginal rate. The steps are small enough that it's not practically possible to optimize around them. A little bit more or less distributions from a taxable account, or bank interest, will foil a strategy to get a few extra dollars tax-free. And the potential benefit from such a strategy is very low.

Then again, the steps in the Irregular Marginal Rates example are $500 up to thousands of dollars; these are possible to optimize around. The idea behind "marginal" is to capture "local" effects, and "local" is context-dependent. In this case, my contention is that "local" should mean above the level of single-dollar rounding and tax table quanta, but below the level of distinct brackets and phase-out ranges. If you're in the 12% bracket with no phase-outs, "small" changes in income will cause an increase in tax liability of 12% of the change. When looking at an irregular case, I think it's fine to say you switch to using an average rate.

User avatar
FiveK
Posts: 7005
Joined: Sun Mar 16, 2014 2:43 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by FiveK » Tue May 28, 2019 8:11 pm

gilgamesh wrote:
Tue May 28, 2019 7:55 pm
Just to remind everyone, what the current boglehead wiki says about the definition

“Marginal tax rate is a tax rate, at the margin, on the last dollar of income.”
https://www.bogleheads.org/wiki/Marginal_tax_rate
It’s reference, a Kitces article :D

The current boglehead wiki on R vs T, always left me with the impression that it was incomplete. Now, I think everyone agrees that it is...
That's what the current wiki says, only because rkhusky didn't follow LadyGeek's request "to edit [the draft] pages as we work through the discrepancies."

Go one version back to see how it has read for several years.
Last edited by FiveK on Tue May 28, 2019 8:14 pm, edited 1 time in total.


User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Tue May 28, 2019 8:13 pm

FiveK wrote:
Tue May 28, 2019 8:11 pm
gilgamesh wrote:
Tue May 28, 2019 7:55 pm
Just to remind everyone, what the current boglehead wiki says about the definition

“Marginal tax rate is a tax rate, at the margin, on the last dollar of income.”
https://www.bogleheads.org/wiki/Marginal_tax_rate
It’s reference, a Kitces article :D

The current boglehead wiki on R vs T, always left me with the impression that it was incomplete. Now, I think everyone agrees that it is...
That's what the current wiki says, only because rkhusky didn't follow LadyGeek's request] "to edit [the draft] pages as we work through the discrepancies."

Go one version back to see how it has read for several years.
I’m unable to look back, I cannot login with my forum credentials...can you cut and paste it here?

User avatar
FiveK
Posts: 7005
Joined: Sun Mar 16, 2014 2:43 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by FiveK » Tue May 28, 2019 8:16 pm

gilgamesh wrote:
Tue May 28, 2019 8:13 pm
FiveK wrote:
Tue May 28, 2019 8:11 pm
gilgamesh wrote:
Tue May 28, 2019 7:55 pm
Just to remind everyone, what the current boglehead wiki says about the definition

“Marginal tax rate is a tax rate, at the margin, on the last dollar of income.”
https://www.bogleheads.org/wiki/Marginal_tax_rate
It’s reference, a Kitces article :D

The current boglehead wiki on R vs T, always left me with the impression that it was incomplete. Now, I think everyone agrees that it is...
That's what the current wiki says, only because rkhusky didn't follow LadyGeek's request] "to edit [the draft] pages as we work through the discrepancies."

Go one version back to see how it has read for several years.
I’m unable to look back, I cannot login with my forum credentials...can you cut and paste it here?
"Marginal tax rate is the tax rate that will apply to the next marginal – or incremental – amount of income (or deductions). It is calculated by dividing the amount of additional taxes that will be due (or reduced) by the amount of income involved."

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Tue May 28, 2019 8:20 pm

DEFINITION (first draft)
Marginal tax rate is the tax rate that will apply to the next marginal – or incremental – amount of income (or deductions). It is calculated by dividing the amount of additional taxes that will be due (or reduced) by the amount of income involved.

Marginal tax rates are used to measure how a person’s tax obligation will change based on some change in strategy. For Roth vs Traditional strategy one should generally compare marginal rates on each dollar of contribution versus the marginal rate upon withdrawal of that dollar.

There are a few instances where the optimal solution requires considering marginal rates for a larger contribution. The contribution amount is up to the maximum of what’s available for that strategy. If the optimal contribution amount have a constant marginal rate then the conventional application of marginal rates on a range still applies.

However the optimal contribution amount may traverse multiple marginal rates. In those instances, the term ‘Marginal tax rate’ is still used to measure at the margin the impact of making a certain change or implementing a certain strategy, by evaluating the tax obligation in scenario A, and scenario B, and the relative difference between the two. Application of the term ‘Marginal Rate’ in these instances is a misnomer as it is not a rate (first order derivative) but a measure of change in rate (second order derivative), but is consistent with the operational use of the term amongst existing Literature.
Last edited by gilgamesh on Wed May 29, 2019 5:11 am, edited 3 times in total.

User avatar
LadyGeek
Site Admin
Posts: 54913
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by LadyGeek » Tue May 28, 2019 8:21 pm

gilgamesh wrote:
Tue May 28, 2019 8:13 pm
I’m unable to look back, I cannot login with my forum credentials...can you cut and paste it here?
To see prior versions, click on the View history tab in the top-right corner of every wiki page.

Note: This tab is not visible on a mobile device.

You only need login credentials to edit the wiki, which requires a separate account from the forum. If you want to be an editor, PM me.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Tue May 28, 2019 8:24 pm

LadyGeek wrote:
Tue May 28, 2019 8:21 pm
gilgamesh wrote:
Tue May 28, 2019 8:13 pm
I’m unable to look back, I cannot login with my forum credentials...can you cut and paste it here?
To see prior versions, click on the View history tab in the top-right corner of every wiki page.

Note: This tab is not visible on a mobile device.

You only need login credentials to edit the wiki, which requires a separate account from the forum. If you want to be an editor, PM me.
I am using an iPad, so that makes sense....thanks!

Topic Author
fyre4ce
Posts: 222
Joined: Sun Aug 06, 2017 11:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by fyre4ce » Tue May 28, 2019 8:25 pm

FiveK,

The biggest reason I have to support the "small" definition of marginal is that I think that's consistent with how it's used by other sources, in industry and academia. Can you tell me, for you, is it more that you think your "any size" definition is actually more consistent with other sources? Or, is it more that you think your definition is more useful, and consistency with other sources is less of a concern? I ask because it might tell us how we could possibly make some progress toward agreement.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Tue May 28, 2019 8:44 pm

Say a flat tax rate of 12% and no other credits etc., my income is $24,231...that’s a tax of $2907.72...but, taxes are only reported in whole numbers. Should we change the definition of the flat tax ‘rate’ now, because the answer had to be rounded off?

Rounding errors are part of any tax calculations. When marginal rates are calculated in $1 increments, it’s understood the margin of error is at least $1. No need to change the definition for that...

Tax table quantization is a larger amount. Imagine the tax table changing to $100, will the definition of marginal rates change too? What about for those who don’t follow the quantization table?, will they have their own definition of marginal rates? Is the definition of marginal rate different from USA to Canada?

These two arguments makes no sense to me.

rkhusky
Posts: 6753
Joined: Thu Aug 18, 2011 8:09 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by rkhusky » Tue May 28, 2019 8:49 pm

Interestingly, a few days ago I was thumbing through the book Math with Bad Drawings: Illuminating the Ideas That Shape Our Reality by Ben Orlin (https://www.amazon.com/Math-Bad-Drawing ... 0316509035), where I found a section talking about marginal tax rates. The author uses the next dollar definition of marginal tax rate, like most everyone else.

Reminder again that the wiki is not supposed to come up with new definitions, but to reference existing terminology.

Lee_WSP
Posts: 391
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by Lee_WSP » Tue May 28, 2019 8:55 pm

FiveK wrote:
Tue May 28, 2019 7:57 pm
Lee_WSP wrote:
Tue May 28, 2019 7:43 pm
- taking either of those literally, between $3,000 and $100,000 of taxable income 98% of single dollars could have a marginal rate of 0%, while the other 2% have marginal rates of 500%, 600%, 1100%, or 1200%.
How are you possibly coming to that conclusion? You're using the term wrong.
Are you familiar with how the Tax Tables work?
Perhaps I'm not, but your still looking at the term incorrectly. The marginal tax applies to the dollar being taxed, not the rate on the table. The tax table would be the equivalent of the cost curve and the marginal cost is the dollar point on the cost curve you are at.

Perhaps the confusion lies in how people are trying to use the term. The term can only be used when the rate on the additional chunk of dollars is the same. If it is not the same, you have to break up that chunk of dollars into pieces where the marginal rates are the same.
What do you suggest when the instantaneous marginal rates are non-monotonic?

That the cost for an additional unit is different from the last or previous unit is exactly why the term was coined.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Wed May 29, 2019 5:10 am

gilgamesh wrote:
Tue May 28, 2019 8:20 pm
DEFINITION (first draft)
Marginal tax rate is the tax rate that will apply to the next marginal – or incremental – amount of income (or deductions). It is calculated by dividing the amount of additional taxes that will be due (or reduced) by the amount of income involved.

Marginal tax rates are used to measure how a person’s tax obligation will change based on some change in strategy. For Roth vs Traditional strategy one should generally compare marginal rates on each dollar of contribution versus the marginal rate upon withdrawal of that dollar.

There are a few instances where the optimal solution requires considering marginal rates for a larger contribution. The contribution amount is up to the maximum of what’s available for that strategy. If the optimal contribution amount have a constant marginal rate then the conventional application of marginal rates on a range still applies.

However the optimal contribution amount may traverse multiple marginal rates. In those instances, the term ‘Marginal tax rate’ is still used to measure at the margin the impact of making a certain change or implementing a certain strategy, by evaluating the tax obligation in scenario A, and scenario B, and the relative difference between the two. Application of the term ‘Marginal Rate’ in these instances is a misnomer as it is not a rate (first order derivative) but a measure of change in rate (second order derivative), but is consistent with the operational use of the term amongst existing Literature.
Is anything wrong with what I wrote above? Anyone want to give their own, around this framework or similar.

The final draft obviously needs examples of each situation so everyone would understand.

rkhusky
Posts: 6753
Joined: Thu Aug 18, 2011 8:09 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by rkhusky » Wed May 29, 2019 6:18 am

gilgamesh wrote:
Wed May 29, 2019 5:10 am
Is anything wrong with what I wrote above? Anyone want to give their own, around this framework or similar.

The final draft obviously needs examples of each situation so everyone would understand.
How about something simple:

DEFINITION
The marginal tax rate is the tax rate that would be paid on the next dollar of income.

or

DEFINITION
The marginal tax rate is the tax rate that was paid on the last dollar of income.

Later one can show how the marginal rate is calculated in practice, along with outside references that support those calculations.
Last edited by rkhusky on Wed May 29, 2019 6:29 am, edited 1 time in total.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Wed May 29, 2019 6:28 am

rkhusky wrote:
Wed May 29, 2019 6:18 am
gilgamesh wrote:
Wed May 29, 2019 5:10 am
Is anything wrong with what I wrote above? Anyone want to give their own, around this framework or similar.

The final draft obviously needs examples of each situation so everyone would understand.
How about something simple:

DEFINITION
The marginal tax rate is the tax rate paid on the next dollar of income.
Because it’s not encompassing (as much as possible) of the definitions out there. I prefer the definition that has been there all along ... “Marginal tax rate is the tax rate that will apply to the next marginal – or incremental – amount of income (or deductions). It is calculated by dividing the amount of additional taxes that will be due (or reduced) by the amount of income involved.”....This is appropriate for the first wiki page that just deals with the definition of Marginal tax rate.

The lengthy ‘definition’ is just an explanation on how this works when considering R vs T. R v T will alway include a larger amount.

The next dollar definition is not embraced by many...incremental is a bit more inclusive as it includes $1, and the ‘small’ definition.

The issue with using words like ‘average’ and ‘effective’, as far as I understand the issue is, they typically apply to a person’s entire income - so, if we are sticking g to conventional usage one cannot use ‘average/effective’ to parts of one’s income...however , mathematically for R vs T, if the contribution amount crosses marginal tax rates, we are in fact measuring ‘effective marginal tax’, however there is no such word and we can’t invent definitions. Therefore, we are forced to stick with what’s out there - This requires a lengthy explanation.

rkhusky
Posts: 6753
Joined: Thu Aug 18, 2011 8:09 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by rkhusky » Wed May 29, 2019 6:30 am

gilgamesh wrote:
Wed May 29, 2019 6:28 am
Because it’s not encompassing (as much as possible) of the definitions out there.
Please provide some references for the other definitions out there.
gilgamesh wrote:
Wed May 29, 2019 6:28 am
The next dollar definition is not embraced by many.
The next dollar definition is used by 99%+ of references out there.


Calculation of the marginal rate should not be part of the definition.
gilgamesh wrote:
Wed May 29, 2019 6:28 am
The issue with using words like ‘average’ and ‘effective’, as far as I understand the issue is, they typically apply to a person’s entire income - so, if we are sticking g to conventional usage one cannot use ‘average/effective’ to parts of one’s income...however , mathematically for R vs T, if the contribution amount crosses marginal tax rates, we are in fact measuring ‘effective marginal tax’, however there is no such word and we can’t invent definitions.
Only when the dollar amount is not provided, e.g. "My average tax rate is 7.6%"

It is clear what is meant by "The average tax rate on an additional $10K of income would be 18.9%"

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Wed May 29, 2019 7:26 am

rkhusky wrote:
Wed May 29, 2019 6:30 am
gilgamesh wrote:
Wed May 29, 2019 6:28 am
The issue with using words like ‘average’ and ‘effective’, as far as I understand the issue is, they typically apply to a person’s entire income - so, if we are sticking g to conventional usage one cannot use ‘average/effective’ to parts of one’s income...however , mathematically for R vs T, if the contribution amount crosses marginal tax rates, we are in fact measuring ‘effective marginal tax’, however there is no such word and we can’t invent definitions.
Only when the dollar amount is not provided, e.g. "My average tax rate is 7.6%"

It is clear what is meant by "The average tax rate on an additional $10K of income would be 18.9%"
I’ll get back to the parts about existing definition and sources later...for now, would like to resolve small parts.

Can we leave the $1 vs ‘small’ vs ‘larger amounts’ definitions for now...and see whether there is consensus on using ‘average’ or effective to parts of one’s entire income?

If everyone agrees we can use ‘average/effective’ for parts of the income, then we cannot say R v T is all about comparing marginal rates. Then go about figuring out the rest.

This will be a small step forward.

P.S: I don’t have a problem with this, other than the general uneasiness of taking a U turn from the boglehead definition that’s been there for years, when it came to R vs T
Last edited by gilgamesh on Wed May 29, 2019 7:52 am, edited 1 time in total.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Wed May 29, 2019 7:41 am

rkhusky wrote:
Wed May 29, 2019 6:30 am
gilgamesh wrote:
Wed May 29, 2019 6:28 am
Because it’s not encompassing (as much as possible) of the definitions out there.
Please provide some references for the other definitions out there.
https://www.cbo.gov/publication/54911
“The marginal federal tax rate on labor income is the percentage of additional income an individual earns that is paid in federal income taxes and payroll taxes.”

The following lists many sources
https://financial-dictionary.thefreedic ... l+tax+rate
“A percentage of one's income that one must pay in taxes.‘
“The percentage of extra income received that must be paid in taxes.‘
“the fraction of the last pound of a person's income that is paid in TAX‘
“marginal tax rateThe percentage of income that must be paid to the IRS for a particular range of incomes,”

https://www.merriam-webster.com/diction ... tax%20rate
“the rate of additional federal income tax to be paid on additional income‘

Kitces (I think you know the link)

https://www.bankrate.com/glossary/m/marginal-tax-rate/
“The marginal tax rate is the highest percentage of income tax someone pays in a system that assigns tax burdens to citizens according to each one’s individual income”
“Under the progressive tax model, the tax rate increases as income rises and pushes individuals into a higher tax bracket. Each marginal tax rate only applies to all taxable income earned within that bracket, and each dollar earned beyond that range is taxed at the next highest tax rate.”

I only said, “encompassing” of many definitions out there...I think the definition of marginal tax rate of $1 and others is at least equally distributed. Does anyone still disagree with this?

There was one text book reference to it being a “small” change....two that clearly had $1 definition. All the above have range, with a couple implying g a $1 is fine as well.
Last edited by gilgamesh on Wed May 29, 2019 7:55 am, edited 1 time in total.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Wed May 29, 2019 7:50 am

Now I’ll wait for responses before proceeding further (the way this thread has progressed, I’m not holding on to much hope here :D )...would like to know what others think of the usage of average/effective rate for parts of one’s income.

Secondly whether anyone still thinks one of ‘$1 vs small vs larger range but constant MR’ definitions of Marginal rates is solely represented out there? If not solely represented among credible sources, then none takes precedence and we should define Marginal tax rate to encompass as much of the definitions as possible. (Which IMO, the existing Boglehead wiki definition does....not R vs T, but the page just on Marginal tax rate)

P.S: I am sympathetic towards the OP, as his intent to improve the wiki pages may have stemmed from the fact that he desires the “small” increment definition.

The only thing that I am concerned with the old definitions was when it came to R vs T. It’s obvious it needed better explanation than just saying compare marginal rates.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Wed May 29, 2019 8:37 am

If everyone is ok with using effective/average for parts of one’s income, and If, there is no overwhelming reasons, based on current usage to change the current boglehead wiki definition... then this is my second draft of the definition trying to be as amicable to Boglehead wiki older definition. .I am only interested in clarifying R vs T, for those who are interested in ‘small’ etc., I’d encourage them to put their version out.

DEFINITION FOR R VS T PAGE (second draft)
Marginal tax rate is the tax rate that will apply to the next marginal – or incremental – amount of income (or deductions). It is calculated by dividing the amount of additional taxes that will be due (or reduced) by the amount of income involved.

Marginal tax rates are generally used to measure how a person’s tax obligation will change based on some change in strategy. For Roth vs Traditional strategy one should generally compare marginal rates of contribution versus the marginal rate upon withdrawal of that contribution. There are a few instances where the optimal solution requires considering marginal rates for a larger contribution. The contribution amount is up to the maximum of what’s available for that strategy. If the optimal contribution amount involves a constant marginal rate then the conventional application of comparing marginal rates applies to a larger range as well. However the optimal contribution amount may traverse multiple marginal rates. In those instances, comparing just one marginal rate is not adequate and one needs to look at the averages of all marginal rates involved.

P.S: This gets the job done!

Lee_WSP
Posts: 391
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by Lee_WSP » Wed May 29, 2019 9:45 am

gilgamesh wrote:
Wed May 29, 2019 5:10 am
However the optimal contribution amount may traverse multiple marginal rates. In those instances, the term ‘Marginal tax rate’ is still used to measure at the margin the impact of making a certain change or implementing a certain strategy, by evaluating the tax obligation in scenario A, and scenario B, and the relative difference between the two. Application of the term ‘Marginal Rate’ in these instances is a misnomer as it is not a rate (first order derivative) but a measure of change in rate (second order derivative), but is consistent with the operational use of the term amongst existing Literature.
This is a poor use of the term marginal & underlines the confusion everyone is having about it.

You cannot use the term marginal if it spans multiple brackets. You must break up the chunks into bands where the additional cost is the same. Otherwise you need to use a term such as average rate. Or you can say that the marginal rate is an average of rates calculated by ($x * y% + $z + a% + $b * c%).
Last edited by Lee_WSP on Wed May 29, 2019 9:49 am, edited 1 time in total.

User avatar
FiveK
Posts: 7005
Joined: Sun Mar 16, 2014 2:43 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by FiveK » Wed May 29, 2019 9:49 am

fyre4ce wrote:
Tue May 28, 2019 8:25 pm
FiveK,

The biggest reason I have to support the "small" definition of marginal is that I think that's consistent with how it's used by other sources, in industry and academia. Can you tell me, for you, is it more that you think your "any size" definition is actually more consistent with other sources? Or, is it more that you think your definition is more useful, and consistency with other sources is less of a concern? I ask because it might tell us how we could possibly make some progress toward agreement.
Sure. It's actually not my definition: LadyGeek brought the Kitces article into the discussion in this post.

Yes, that definition is more useful than a "single dollar" or first derivative one, in the context of the tax code. People can quote Economics 101 textbooks all they want, but unless those quotes include evaluation of non-monotonic functions they aren't generally applicable to what we are discussing here.

As you have observed just yesterday in Understanding the Wiki - Traditional versus Roth - Bogleheads.org, people still don't understand why "marginal vs. effective" is not correct. Helping people understand that is the lens through which I look at this discussion on marginal rate. If we can provide that help, great!

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Wed May 29, 2019 9:50 am

Lee_WSP wrote:
Wed May 29, 2019 9:45 am
gilgamesh wrote:
Wed May 29, 2019 5:10 am
However the optimal contribution amount may traverse multiple marginal rates. In those instances, the term ‘Marginal tax rate’ is still used to measure at the margin the impact of making a certain change or implementing a certain strategy, by evaluating the tax obligation in scenario A, and scenario B, and the relative difference between the two. Application of the term ‘Marginal Rate’ in these instances is a misnomer as it is not a rate (first order derivative) but a measure of change in rate (second order derivative), but is consistent with the operational use of the term amongst existing Literature.
This is a poor use of the term marginal & underlines the confusion everyone is having about it.

You cannot use the term marginal if it spans multiple brackets. You must break up the chunks into bands where the additional cost is the same. Otherwise you need to use a term such as average rate. Or you can say that the marginal rate is $x * y% + $z + a% + $b * c%
True! Thus the reason for me to mention that it’s a misnomer - this part, “Application of the term ‘Marginal Rate’ in these instances is a misnomer as it is not a rate (first order derivative) but a measure of change in rate (second order derivative), but is consistent with the operational use of the term amongst existing Literature.”

Anyhow, how about the later version...the relevant part of that

“There are a few instances where the optimal solution requires considering marginal rates for a larger contribution. The contribution amount is up to the maximum of what’s available for that strategy. If the optimal contribution amount involves a constant marginal rate then the conventional application of comparing marginal rates applies to a larger range as well. However the optimal contribution amount may traverse multiple marginal rates. In those instances, comparing just one marginal rate is not adequate and one needs to look at the averages of all marginal rates involved.‘

rkhusky
Posts: 6753
Joined: Thu Aug 18, 2011 8:09 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by rkhusky » Wed May 29, 2019 9:50 am

gilgamesh wrote:
Wed May 29, 2019 7:41 am
I think the definition of marginal tax rate of $1 and others is at least equally distributed. Does anyone still disagree with this?
I get 14K hits on Google for "marginal tax rate" "next dollar" and 17.6K hits for "marginal tax rate" "last dollar". I'll concede equal distribution if you can approach those numbers with other definitions.

I find any definition, which doesn't put a bound on the amount of income added, to be mostly useless. With the amount undefined, I could start with $0 and add my whole income and call the resulting tax rate my marginal tax rate. Or I could add $1M to my existing income and call that my marginal tax rate.

$1 is the amount that is in the vast majority of definitions. Even in "practical definitions", the added amount is something like $100 or $1000. Unfortunately, you are unlikely to find references that state that the added amount should be $100 or $1000. Whenever a dollar amount is used in a definition, the amount is $1.

The article could state multiple definitions, providing references to each of them.

Incidentally, the definition that states that the marginal amount is the highest tax rate is contradictory to almost all other definitions and is only consistent with other definitions for monotonically increasing tax rates.

rkhusky
Posts: 6753
Joined: Thu Aug 18, 2011 8:09 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by rkhusky » Wed May 29, 2019 9:55 am

FiveK wrote:
Wed May 29, 2019 9:49 am

Yes, that definition is more useful than a "single dollar" or first derivative one, in the context of the tax code. People can quote Economics 101 textbooks all they want, but unless those quotes include evaluation of non-monotonic functions they aren't generally applicable to what we are discussing here.
Computing the average tax rate over a non-monotonic function is not hampered at all by using the $1 definition of marginal tax rates. Calling the average tax rate in that situation "marginal tax rate" is very confusing.

Lee_WSP
Posts: 391
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by Lee_WSP » Wed May 29, 2019 10:01 am

gilgamesh wrote:
Wed May 29, 2019 9:50 am
Lee_WSP wrote:
Wed May 29, 2019 9:45 am
gilgamesh wrote:
Wed May 29, 2019 5:10 am
However the optimal contribution amount may traverse multiple marginal rates. In those instances, the term ‘Marginal tax rate’ is still used to measure at the margin the impact of making a certain change or implementing a certain strategy, by evaluating the tax obligation in scenario A, and scenario B, and the relative difference between the two. Application of the term ‘Marginal Rate’ in these instances is a misnomer as it is not a rate (first order derivative) but a measure of change in rate (second order derivative), but is consistent with the operational use of the term amongst existing Literature.
This is a poor use of the term marginal & underlines the confusion everyone is having about it.

You cannot use the term marginal if it spans multiple brackets. You must break up the chunks into bands where the additional cost is the same. Otherwise you need to use a term such as average rate. Or you can say that the marginal rate is $x * y% + $z + a% + $b * c%
True! Thus the reason for me to mention that it’s a misnomer - this part, “Application of the term ‘Marginal Rate’ in these instances is a misnomer as it is not a rate (first order derivative) but a measure of change in rate (second order derivative), but is consistent with the operational use of the term amongst existing Literature.”

Anyhow, how about the later version...the relevant part of that

“There are a few instances where the optimal solution requires considering marginal rates for a larger contribution. The contribution amount is up to the maximum of what’s available for that strategy. If the optimal contribution amount involves a constant marginal rate then the conventional application of comparing marginal rates applies to a larger range as well. However the optimal contribution amount may traverse multiple marginal rates. In those instances, comparing just one marginal rate is not adequate and one needs to look at the averages of all marginal rates involved.‘
While that is a true statement, it would be much more accurate to just break up the large chunk into the constituent bands. For a practical example (barring any subsidy removals):

Say you're earning $45,000.

The marginal rate at which the last dollar was taxed was 22% for single.
If talking about the last $10,000 put in, then $5,525 was taxed at 22%, but the rest was taxed at 12%.

If one is considering whether to Roth the whole $10k, then one needs to break it up into the two chunks.
$5,525 taxed at 22% and the additional $4,475 taxed at the much lower 12%.

So, $5,500 won't net you a lot of savings, but that extra $4500 locks in that much much lower 12%.

That's how you need to do the analysis. You can average it out afterwards if it makes it easier to comprehend the total impact.

Lee_WSP
Posts: 391
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by Lee_WSP » Wed May 29, 2019 10:02 am

rkhusky wrote:
Wed May 29, 2019 9:55 am
FiveK wrote:
Wed May 29, 2019 9:49 am

Yes, that definition is more useful than a "single dollar" or first derivative one, in the context of the tax code. People can quote Economics 101 textbooks all they want, but unless those quotes include evaluation of non-monotonic functions they aren't generally applicable to what we are discussing here.
Computing the average tax rate over a non-monotonic function is not hampered at all by using the $1 definition of marginal tax rates. Calling the average tax rate in that situation "marginal tax rate" is very confusing.
I agree. I do not understand FiveK's use of non-monotonic. If there is a 10,000 unit band, the marginal rate on each additional unit from unit 1 through 10,000 is the same. But once you get to 10,001, the marginal rate changes. It's not that difficult of a concept, IMO.

rkhusky
Posts: 6753
Joined: Thu Aug 18, 2011 8:09 pm

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by rkhusky » Wed May 29, 2019 10:31 am

Lee_WSP wrote:
Wed May 29, 2019 10:02 am

I agree. I do not understand FiveK's use of non-monotonic. If there is a 10,000 unit band, the marginal rate on each additional unit from unit 1 through 10,000 is the same. But once you get to 10,001, the marginal rate changes. It's not that difficult of a concept, IMO.
It refers to situations where the marginal rate goes up and then goes back down. This can happen with taxation of Social Security benefits, for instance. Here one should compare average tax rates rather than marginal tax rates, when making the Roth versus Traditional decision.

Lee_WSP
Posts: 391
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by Lee_WSP » Wed May 29, 2019 10:59 am

rkhusky wrote:
Wed May 29, 2019 10:31 am
Lee_WSP wrote:
Wed May 29, 2019 10:02 am

I agree. I do not understand FiveK's use of non-monotonic. If there is a 10,000 unit band, the marginal rate on each additional unit from unit 1 through 10,000 is the same. But once you get to 10,001, the marginal rate changes. It's not that difficult of a concept, IMO.
It refers to situations where the marginal rate goes up and then goes back down. This can happen with taxation of Social Security benefits, for instance. Here one should compare average tax rates rather than marginal tax rates, when making the Roth versus Traditional decision.
Yes, using the average rate is the only way you can talk about a total quantity when it crosses brackets.

For everyone else who is still confused, the term is usually used when a purchasing manager has to decide on how many widgets to order. Manufacturers typically offer price breaks based on quantity.

$1/unit for 10+ units
$0.9/unit for 50+ units
$.75/unit for 200+ units
and so forth

So the marginal cost of that 51st unit is actually
(.9 * 51 - 1 * 50) = -$4.1 (apparently in my example it is actually cheaper to purchase 51 units than 50 units, but the example still works; that's how margins work)
Last edited by Lee_WSP on Wed May 29, 2019 1:23 pm, edited 1 time in total.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Wed May 29, 2019 11:33 am

Lee_WSP wrote:
Wed May 29, 2019 10:01 am
gilgamesh wrote:
Wed May 29, 2019 9:50 am
Lee_WSP wrote:
Wed May 29, 2019 9:45 am
gilgamesh wrote:
Wed May 29, 2019 5:10 am
However the optimal contribution amount may traverse multiple marginal rates. In those instances, the term ‘Marginal tax rate’ is still used to measure at the margin the impact of making a certain change or implementing a certain strategy, by evaluating the tax obligation in scenario A, and scenario B, and the relative difference between the two. Application of the term ‘Marginal Rate’ in these instances is a misnomer as it is not a rate (first order derivative) but a measure of change in rate (second order derivative), but is consistent with the operational use of the term amongst existing Literature.
This is a poor use of the term marginal & underlines the confusion everyone is having about it.

You cannot use the term marginal if it spans multiple brackets. You must break up the chunks into bands where the additional cost is the same. Otherwise you need to use a term such as average rate. Or you can say that the marginal rate is $x * y% + $z + a% + $b * c%
True! Thus the reason for me to mention that it’s a misnomer - this part, “Application of the term ‘Marginal Rate’ in these instances is a misnomer as it is not a rate (first order derivative) but a measure of change in rate (second order derivative), but is consistent with the operational use of the term amongst existing Literature.”

Anyhow, how about the later version...the relevant part of that

“There are a few instances where the optimal solution requires considering marginal rates for a larger contribution. The contribution amount is up to the maximum of what’s available for that strategy. If the optimal contribution amount involves a constant marginal rate then the conventional application of comparing marginal rates applies to a larger range as well. However the optimal contribution amount may traverse multiple marginal rates. In those instances, comparing just one marginal rate is not adequate and one needs to look at the averages of all marginal rates involved.‘
While that is a true statement, it would be much more accurate to just break up the large chunk into the constituent bands. For a practical example (barring any subsidy removals):

Say you're earning $45,000.

The marginal rate at which the last dollar was taxed was 22% for single.
If talking about the last $10,000 put in, then $5,525 was taxed at 22%, but the rest was taxed at 12%.

If one is considering whether to Roth the whole $10k, then one needs to break it up into the two chunks.
$5,525 taxed at 22% and the additional $4,475 taxed at the much lower 12%.

So, $5,500 won't net you a lot of savings, but that extra $4500 locks in that much much lower 12%.

That's how you need to do the analysis. You can average it out afterwards if it makes it easier to comprehend the total impact.
Yes! R vs T Analysis is done in chunks. Not only chunks now, but those chunks have to mesh with chunks in retirement that has a constant marginal rate in retirement as well.

The reason to average them out is to simplify the definition as much as possible. It is tedious as it is, without going into chunks.

P.S: Typo in the previous post “cheaper to purchase 50 units than 51 units”, need to switch the numbers.
Last edited by gilgamesh on Wed May 29, 2019 12:20 pm, edited 1 time in total.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Wed May 29, 2019 11:35 am

rkhusky wrote:
Wed May 29, 2019 9:50 am
gilgamesh wrote:
Wed May 29, 2019 7:41 am
I think the definition of marginal tax rate of $1 and others is at least equally distributed. Does anyone still disagree with this?
I get 14K hits on Google for "marginal tax rate" "next dollar" and 17.6K hits for "marginal tax rate" "last dollar". I'll concede equal distribution if you can approach those numbers with other definitions.

I find any definition, which doesn't put a bound on the amount of income added, to be mostly useless. With the amount undefined, I could start with $0 and add my whole income and call the resulting tax rate my marginal tax rate. Or I could add $1M to my existing income and call that my marginal tax rate.

$1 is the amount that is in the vast majority of definitions. Even in "practical definitions", the added amount is something like $100 or $1000. Unfortunately, you are unlikely to find references that state that the added amount should be $100 or $1000. Whenever a dollar amount is used in a definition, the amount is $1.

The article could state multiple definitions, providing references to each of them.

Incidentally, the definition that states that the marginal amount is the highest tax rate is contradictory to almost all other definitions and is only consistent with other definitions for monotonically increasing tax rates.
I’m fine with stating multiple definitions.

P.S: I didn’t try google hits, as I wouldn’t put much weight on the results. I prefer to click on the results and see.

Lee_WSP
Posts: 391
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by Lee_WSP » Wed May 29, 2019 1:23 pm

gilgamesh wrote:
Wed May 29, 2019 11:33 am
The reason to average them out is to simplify the definition as much as possible. It is tedious as it is, without going into chunks.
I'm not convinced averaging the number is the right way of thinking about how much to contribute to a Roth vs Traditional account.

I think the correct way to think about it is in chunks. If the first $10,000 is more advantageous in a traditional, and the next $10,000 is more advantageous in a Roth, the average isn't going to get you to that conclusion. (This would happen if reducing one's taxable income by 10k gets you into the lower tax bracket (ie the extra 10k would've been taxed at a higher rate without the deferral)).

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Wed May 29, 2019 1:59 pm

Lee_WSP wrote:
Wed May 29, 2019 1:23 pm
gilgamesh wrote:
Wed May 29, 2019 11:33 am
The reason to average them out is to simplify the definition as much as possible. It is tedious as it is, without going into chunks.
I'm not convinced averaging the number is the right way of thinking about how much to contribute to a Roth vs Traditional account.

I think the correct way to think about it is in chunks. If the first $10,000 is more advantageous in a traditional, and the next $10,000 is more advantageous in a Roth, the average isn't going to get you to that conclusion. (This would happen if reducing one's taxable income by 10k gets you into the lower tax bracket (ie the extra 10k would've been taxed at a higher rate without the deferral)).
It’s averages of the appropriate range. Average will get to the conclusion.

This is not a major sticking point, but to get it out of the way, I think it would be best to think of high marginal tax rate spikes followed by lower marginal rates. The chunk of high spike means nothing - for that chunk it’s a bad idea to do Roth, however subsequent lower marginal rates chunks make the overall Roth better. This clearly shows a chunk at a time is not it, one has to look at averages.

User avatar
gilgamesh
Posts: 1500
Joined: Sun Jan 10, 2016 9:29 am

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by gilgamesh » Wed May 29, 2019 2:09 pm

I personally prefer to leave the wiki pages as it is, rather than going with $100 or $1000 marginal rate definition, exclusively.

I’m fine with $1, I’m fine with appropriate range. But, limiting it based on rounding errors and tax tables of USA sounds so wrong (for reasons mentioned previously).

Of the two I prefer the $1 definition, but when it comes to R vs T, it needs to be either the definition needs to be more than just $1 and include range or do not introduce MR at all into making that choice of R vs T

Or else give all three definitions (may be this is the best option).

Lee_WSP
Posts: 391
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: Seeking comment on proposed revisions to Marginal Tax Rate and Traditional v. Roth wiki articles

Post by Lee_WSP » Wed May 29, 2019 2:40 pm

I still don't agree that averages is the correct way to describe the choices, but perhaps you can change my mind.

I'm not seeing how the average rate is accurate. Can you give several examples where it works and is better than the marginal definition?

Post Reply