Any good Roth conversion estimator websites ?

[ERD50]

No, in the 12% now 22% later scenario you realize the savings when the conversion occurs and the tax is paid. ...

I still do not see this. I can see where you could say you 'book' the savings, but I can't see how you realize it when the conversion occurs.

To my thinking, the only way I could realize it in the same year, is if on my tax form that year, there was one line that would debit me for the tax for the conversion ($1.2K on a $10K conversion), and a second line that credited me that same tax year for the future savings (but we don;t know what it is yet). It does not happen in the same tax year. It is not realized in the same tax year.

Yes, the future liability for the taxes is gone at the time I convert, because that money is no longer in an account that will tax it - but that's not 'realizing' it. Just like when a stock goes up, I have a gain, but I have not realized it until I sell. I don't think I 'realize' the tax savings, until I see my RMD reduced by the % of that conversion, or until I've liquidated the tIRA.

Say I converted, and a few years later my 1st RMD was $10K, but I decided to take out $15K. I get taxed on $15K that year, whether I converted some of it or not. If I keep taking more than my RMD, I don't realize any tax saving until that last year, when the amount I take out is less than it would have been had I not done a conversion. If my tax didn't change for all those years, then I certainly did NOT 'you realize the savings when the conversion occurs'. If I didn't convert that earlier year, no tax is owed that year. Paying a tax is not 'saving'.

I think we are stuck on 'booking' the savings, versus 'realizing' the savings. Am I wrong? Where?

-ERD50

 

[pb4uski]

Think of it this way... at the start of the year you have a $10k tIRA (with a basis of zero) and a deferred tax liability of $2.2k because you expect to have to pay 22% in tax when you take that money out of the tIRA. Your net assets are $7.8k. Right?

During the year, the only thing that happens is that the opportunity to settle that $2.2k deferred tax liability for $1.2k (12%) arises and you take the distribution and settle the $2.2k DTL for $1.2k.

At the end of the period you have cash (or a Roth) of $8.8k ($10k withdrawal less $1.2k of taxes paid at 12%) with the same basis so no DTA or DTL and your net assets are $8.8k.

Are you or are you not $1.0k ahead of where you were at the beginning of the period since your beginning of year net worth is $7.8k and end of period net worth is $8.8k?

The $1k gain arises from settling the DTL for less than what you expected to have to pay... 12% rather than 22%. Since you have a net income of $1k, you have realized the gain when the tax was paid.

------------------

If you have a $1,000k tIRA that you expect to pay 22% on and convert/withdraw $10k it is the same dynamic... a portion of the DTL based on 22% is settled for 12% resulting in a gain for the 10% difference times the amount converted or withdrawn.

 

[RunningBum]

^^ Yes, that's it. Seeing $10K in a tIRA and thinking that's all yours isn't right. It's just like if you have a $100K salary. You wouldn't think, hey, I've got $100K to spend or save. It's the take home part that you have available, after taxes. Something less than $100K. Same with the tIRA because that is tax-deferred income.

 

[sengsational]

Yep, RB and PB have it right. The concept of thinking you own the entire tIRA is the problem. You never owned 100% of it. I like the idea of only counting a percentage of it in net worth (but since the tax rate is not knowable precisely, that's a problem), so I include taxes paid on tIRA conversions and withdrawals in the expense budget. But that leads to the kind of break even confusion covered in this thread. DTL... I'm going to start using that acronym.

 

[ERD50]

Think of it this way... ...

Because I appreciate/respect your views, I printed out your response, and carefully went through it line by line.

I feel confident now, that I "got it" several posts back, and my opinion has not changed. We are in agreement of the arithmetic, but I just don't agree with your terminology of 'realized' in this case. It doesn't matter at this point, we can agree to disagree, I just wanted to make sure I wasn't missing anything.

So to summarize for closure (not debate), I agree that you can 'book' the effect of the conversion, based on current and assumed future tax rates. But since the future rate is unknown, and unpaid at this time, I just won't call it 'realized'.

Consider this scenario for example - (admittedly unlikely, but certainly not impossible) where I die a year before taking any RMD. My heir has a low income, and the RMDs are not taxed at all. There was no savings.

Maybe a more probable case is a single person, making conversions to the top of the 22% bracket, and their heir is married, and solidly in the 12% bracket. Do you 'realize' a loss? Where can I write off this 'loss'?

Bottom line for me, 'realized' means 'realized', not estimated.

And yes, I am making my ROTH conversions based on these assumptions. But that is all they are, 'assumptions'.

-ERD50

 

[pb4uski]

Since by definition deferred taxes are the tax impact of reversals of temporary timing differences (actually, I guess more properly book-tax basis differences but I am more old-school)... their calculation requires the use of tax rates that will be in effect and apply when those differences reverse (in our case 22%). If something happens and you can settle that liabilty for 12% you have realized a gain... that change in estimate goes through the P&L like any other change in estimate.

Similarly, when corporate tax rates changed from 35% to 21% the impact of the change in estimate went through the P&L.... and other than mark-to-market trading securities, such changes in estimate are typically considered realized and go through the P&L.

Under your definition using the $10k conversion example, I'm not sure when you would view the savings as being realized.... when would that be? If you can provide a sensible answer to that question then perhaps I can see the light.

In the example that you used where you die before any RMDs then you save 22% since YOU never had to pay the tax that you expect to be due... so the 22% DTL evaporates into thin air and you have a gain since YOU settled the DTL for $0. Your heirs don't count.. they are a different entity.

During the year, the only thing that happens is that the opportunity to settle that $2.2k deferred tax liability for $1.2k (12%) arises and you take the distribution and settle the $2.2k DTL for $1.2k.

At the end of the period you have cash (or a Roth) of $8.8k ($10k withdrawal less $1.2k of taxes paid at 12%) with the same basis so no DTA or DTL and your net assets are $8.8k.

Are you or are you not $1.0k ahead of where you were at the beginning of the period since your beginning of year net worth is $7.8k and end of period net worth is $8.8k?

The $1k gain arises from settling the DTL for less than what you expected to have to pay... 12% rather than 22%. Since you have a net income of $1k, you have realized the gain when the tax was paid.

 

[RunningBum]

Because I appreciate/respect your views, I printed out your response, and carefully went through it line by line.

I feel confident now, that I "got it" several posts back, and my opinion has not changed. We are in agreement of the arithmetic, but I just don't agree with your terminology of 'realized' in this case. It doesn't matter at this point, we can agree to disagree, I just wanted to make sure I wasn't missing anything.

So to summarize for closure (not debate), I agree that you can 'book' the effect of the conversion, based on current and assumed future tax rates. But since the future rate is unknown, and unpaid at this time, I just won't call it 'realized'.

Consider this scenario for example - (admittedly unlikely, but certainly not impossible) where I die a year before taking any RMD. My heir has a low income, and the RMDs are not taxed at all. There was no savings.

Maybe a more probable case is a single person, making conversions to the top of the 22% bracket, and their heir is married, and solidly in the 12% bracket. Do you 'realize' a loss? Where can I write off this 'loss'?

Bottom line for me, 'realized' means 'realized', not estimated.

And yes, I am making my ROTH conversions based on these assumptions. But that is all they are, 'assumptions'.

-ERD50

Back in post 10 you said:

"Assuming heirs are in a > 12% marginal tax bracket"...

That's what I was going on. But now you are changing assumptions. Of course if your heirs are probably going to be in a lower tax bracket than you, you should take that into consideration.

There are a few other things that could make conversion go bad:

- tax rates could go even lower

- SS might go away, and/or you lose all your other investments somehow, and your RMDs are you only income and are taxed at 0%

- Your tIRA drops a lot. You'd have been better off paying taxes on the lower amount later than the full amount now.

And so on.

You can't make a conversion decision that will work best 100% of the time. You can make one with assumptions of the most likely scenarios, and optimize those.

 

[pb4uski]

Yup... any conversion to benefit from lower tax rates is an educated guess based on your likely tax rates later on.

 

[ERD50]

....

Under your definition using the $10k conversion example, I'm not sure when you would view the savings as being realized.... when would that be? If you can provide a sensible answer to that question then perhaps I can see the light. .....

It was in my earlier post, but I'll repeat a simplified version here, with a total tIRA of $20,000, assume no growth or inflation for this example:


A) At age 60, I'm in a 12% bracket, and I convert $10,000 of the tIRA to a ROTH and pay 12% tax on it. My check to the FEDS is $1,200 more than if I did not convert that year. Cash flow wise (and this is the view I'm coming from), I am now 'out' $1,200.

B) At age 70, in a 22% bracket, I liquidate the remaining $10,000 of the tIRA, and I pay $2,200 in taxes. I am now 'out' a total of $3,400 in taxes.

Had I not converted in step A, I would not have paid any tax yet (no cash-flow 'hit'). If I liquidate the entire $20,000 of the tIRA at age 70, I pay a total of $4,400 in taxes. The $1,000 savings wasn't realized until age 70, and it required an upfront cash-flow of $1,200.

In the example that you used where you die before any RMDs then you save 22% since YOU never had to pay the tax that you expect to be due... so the 22% DTL evaporates into thin air and you have a gain since YOU settled the DTL for $0. Your heirs don't count.. they are a different entity.

So if we say 'heirs don't count', and I die before RMD, then if I don't convert I never pay taxes on any of it. If I convert, I pay $1,200. In my world, paying $1,200 versus paying $0 is not a savings!

Again, I think it it is two ways to look at it, neither is right/wrong, they are just different. You are looking from a balance sheet view with tax liability assumptions based on current information, I'm looking from an actual cash-out-of-my-pocket view.


Hah, I love analogies, let me try another one:

I buy a gift card for $100 worth of merchandise at a discount, I pay only $80 for it. The card is non-transferable ( 'heirs don't count'). The $80 is deducted from my checking account - it is gone. I die before buying the $100 of merchandise - did I 'save' anything? All I see is an $80 deduction!

Back in post 10 you said:

"Assuming heirs are in a > 12% marginal tax bracket"...

That's what I was going on. But now you are changing assumptions. ....

You can't make a conversion decision that will work best 100% of the time. You can make one with assumptions of the most likely scenarios, and optimize those.

Agreed. It's all based on assumptions. We just make our best guess and decide whether to act or not.

-ERD50

 

[pb4uski]

....the $1,000 savings wasn't realized until age 70, and it required an upfront cash-flow of $1,200. ...

=rate((70-60), ,-1200,2200)=6.25%... and that is an after-tax return since you never get taxed on the $1,000 benefit.

 

[Cobra9777]

=rate((70-60), ,-1200,2200)=6.25%... and that is an after-tax return since you never get taxed on the $1,000 benefit.

I wonder if you might elaborate on this formula a bit? I follow everything except the $2200 FV. I think it's the difference between $2200 actually paid at 70 vs $4400 that would have been paid with no conversion 10 years earlier. Is that right?

Also, does this mean you are warming up to ERD50's cashflow approach to evaluating the conversion decision? Or do still think it's strictly rate related? Or both?

 

[RunningBum]

Hah, I love analogies, let me try another one:

I buy a gift card for $100 worth of merchandise at a discount, I pay only $80 for it. The card is non-transferable ( 'heirs don't count'). The $80 is deducted from my checking account - it is gone. I die before buying the $100 of merchandise - did I 'save' anything? All I see is an $80 deduction!

-ERD50

No, you don't see it. You're dead. I'm serious. You don't care. It doesn't matter.

Similarly, you don't make long-term financial decisions for retirement based on the chance you might die tomorrow. You base them on the most likely case, or probably more appropriately, a reasonably likely "worst-case" of longevity.

Once again you're trying to argue by picking an outlier case and yelling "See! It doesn't work!" It's never going to work 100% of the time. You even agreed that you have to make assumptions, but here you go again with the "yeah, but what if..." scenarios.

 

[ERD50]

No, you don't see it. You're dead. I'm serious. You don't care. It doesn't matter.

Similarly, you don't make long-term financial decisions for retirement based on the chance you might die tomorrow. You base them on the most likely case, or probably more appropriately, a reasonably likely "worst-case" of longevity.

Once again you're trying to argue by picking an outlier case and yelling "See! It doesn't work!" It's never going to work 100% of the time. You even agreed that you have to make assumptions, but here you go again with the "yeah, but what if..." scenarios.

I'm actually not trying to 'argue' that either view is right or wrong. I'm only trying to say that each is a different, and valid, view. And that we should recognize the differences.

You say "You're dead. I'm serious. You don't care. It doesn't matter.", OK, fine. But I was only trying to make it an equivalent to what pb4uski said, that "heirs don't matter'. So if I don't care, I'm dead, there was no point in spending money to save later, if I never made it to the savings age.

I agree - look at the long term, I am, and I can't predict my heirs tax bracket, I can't even predict mine, since tax laws can change. So I'm making my best guess, and I am converting as much as I can in the 12% bracket. I am not arguing against converting.

But I continue to look at as, the savings is not realized until I (or my heirs) actually 'cash in' the savings. You can 'book it' on paper if you wish, but you also need to update that as conditions change. To my thinking, if something can change, it has not been "realized".

-ERD50

 

[RunningBum]

But I continue to look at as, the savings is not realized until I (or my heirs) actually 'cash in' the savings. You can 'book it' on paper if you wish, but you also need to update that as conditions change. To my thinking, if something can change, it has not been "realized".

-ERD50

If you're worried about things changing on you, then you can't count on anything. That $10,000 in your account, of any type? The market could be cut in half tomorrow, so I better not count that as $10,000.

So one way to look at it is, if I needed to access the money today, what could I get? How much of that $10,000 could I actually spend?

If it's $10,000 in a Roth, I can spend all $10,000.

If it's $10,000 in a tIRA, I can spend $8800 (12% tax). And keep in mind that later I could only spend $7800 of it.

If it's $10,000 in taxable, it might be reduced by cap gains tax.

So we're back to the concept that you never had $10K to actually spend in that tIRA. There's no "realizing" savings later--you never really had it in hand to begin with.

 

[pb4uski]

I wonder if you might elaborate on this formula a bit? I follow everything except the $2200 FV. I think it's the difference between $2200 actually paid at 70 vs $4400 that would have been paid with no conversion 10 years earlier. Is that right?

Also, does this mean you are warming up to ERD50's cashflow approach to evaluating the conversion decision? Or do still think it's strictly rate related? Or both?

By paying $1,200 today you avoid having to pay $2,200 in 10 years, so the IRR is 6.25%. ... IOW if you deposited $1,200 in an account that paid 6.25% interest in 10 years you would have $2,200. Based only on the $10k that is withdrawn/converted.

And no, not warming up to ERD50's cash flow approach at all. Not the first time that we've disagreed, and I'm guessing not the last!

 

[ERD50]

If you're worried about things changing on you, then you can't count on anything. ....

What does 'worrying' have to do with anything?

I think you made my point there - you can't count on what the future will bring, so to my view, you can't say it is 'realized' either. If you want to 'book it' to keep your balance sheet current, based on current assumptions, fine go ahead, and update it as things change. I just choose to count it when it happens. That's where the rubber meets the road.

We're going in circles, and it really doesn't matter - we both see the benefit (and risk) of a ROTH conversion, we just choose different accounting methods.The difference isn't changing my decision, that's what I was interested in, in case I saw something that would change my view of it.

If it isn't 'actionable', then it is academic, and I think I've gone through the academic exercise enough for now.

-ERD50

 

[RunningBum]

What does 'worrying' have to do with anything?

Then used "concerned with", "considering", etc. I don't think it's productive to nitpick on words, but I'll apologize for using one that may have a negative connotation.

...
We're going in circles, and it really doesn't matter - we both see the benefit (and risk) of a ROTH conversion, we just choose different accounting methods.The difference isn't changing my decision, that's what I was interested in, in case I saw something that would change my view of it.

If it isn't 'actionable', then it is academic, and I think I've gone through the academic exercise enough for now.

-ERD50

Your accounting method is looking for a payback at a later date. I say there is none. That's a pretty big difference that can change how one does conversions. It goes without saying that you're free to use whatever method you want, but I wouldn't advise it for others looking at Roth conversion.

Like you, I don't mind debating these points to see if it changes my point of view and strategy. I really have changed a few things I do based on discussion here. For this subject, I'm very convinced I'm not changing to your method. It's going nowhere at this point, so I'm also done.

 

[sengsational]

So if one holds that something can't be realized if it can't be quantified, that got me thinking (dangerous, I know)...what if the thing could be quantified within 1%? Is that close enough? 0.1%? Obviously a tax rate difference of 12 versus 22 makes it difficult for some to realize the expense, but how small would it need to be to throw in the towel.

 

[Cobra9777]

...I think you made my point there - you can't count on what the future will bring, so to my view, you can't say it is 'realized' either. If you want to 'book it' to keep your balance sheet current, based on current assumptions, fine go ahead, and update it as things change. I just choose to count it when it happens. That's where the rubber meets the road.

We're going in circles, and it really doesn't matter - we both see the benefit (and risk) of a ROTH conversion, we just choose different accounting methods.The difference isn't changing my decision, that's what I was interested in, in case I saw something that would change my view of it.

If it isn't 'actionable', then it is academic, and I think I've gone through the academic exercise enough for now.

-ERD50

Actually I think you haven't. This discussion *may* be more than just semantics and accounting methods. We don't know yet. One camp has shown lots of analysis concluding that tax rate differences are essentially all that matter in the conversion decision. You introduced the idea that cashflow timing and payback also matter, or maybe matter more. I think this *may* have some merit, probably not. Yet with the exception of PB4's TVM formula for your very simple cashflow-based example, I've seen no real quantitative analysis of this idea. I'd like to see a thorough TVM-based analysis of realistic conversion-related cashflows that validates posts like this:

...Before I gave this much thought, conversions seemed to be a "no brainer" because in my mind I was saving 10% in taxes by converting early at 12% versus RMD time in the 22% bracket. But after thinking it thru, I see that it takes time to "realize" that full savings, although I hadn't spent the time to quantify just how long it would take...

If the rest of us are missing something important, this is a great opportunity to flesh it out. Or we can just let the thread die with an uninteresting agreement to disagree. I'd do the analysis but I'm not the one who presented the idea, and I have a woodworking project going in the garage.

 

[pb4uski]

I'll give it a try.

Base assumptions: At inception $10k in tIRA and $1,200 in taxable account. Can convert today at 12% tax or in 10 years at 22% tax. 7% annual pre-tax return. 12% tax rate for all years other than end of 10th year (SS starts and puts subject into 22% tax bracket from 12% tax bracket at end of 10th year). Measure after-tax value of each alternative after 10 years.

Scenario 1: Convert. $1,200k taxable is paid to feds for tax on conversion. 10k Roth grows to $19,672 ($10k *(1+7%)^10) that is fully spendable.

Scenario 2: Don't convert. $10k in tIRA grows to $19,672. $1,200 taxable account grows to $2,182 ($1,200*(1+7%*(1-12%))^10). Tax due on withdrawal is $4,328. Total available to spend is $17,526 ($19,672 - $4,328 tax + $2,182 taxable account balance)

Difference is $2,146... principally the $1,000 of taxes saved that grows to $1,967 ($1,000*(1+7%)^10) plus the taxes on the taxable account that were avoided of $179 ($1,200-(1+7%)^10-$1,200*(1+7%*(1-12%))^10).

 

[Bongleur]

By paying $1,200 today you avoid having to pay $2,200 in 10 years, so the IRR is 6.25%. ... IOW if you deposited $1,200 in an account that paid 6.25% interest in 10 years you would have $2,200. Based only on the $10k that is withdrawn/converted.

And no, not warming up to ERD50's cash flow approach at all. Not the first time that we've disagreed, and I'm guessing not the last!

Oh my. What is the rate on a 10 year Govt Bond? Nowhere near 6.25%.
So a conservative investor should convert to ROTH, right?

So the thing to do is to calculate the IRR on every possible tax rate (every % point from 12 to 30 or so) for various time spans -- relating to how soon you will retire, start drawing, etc. That's a simple table, to printout & keep looking at. Does the IRR exceed your investment risk/return comfort zone for the long term? If it does, convert to ROTH. Right?

 

[TwoByFour]

I went through this recently and here is my calculation.
Let:
A = amount of IRA to potentially convert
G = growth rate of IRA
H = growth rate of Roth
T = tax rate at time of conversion
R = tax rate of RMD
N = number of years from conversion to use of funds

What I looked at is the after-tax value after N years of each.
Value of Roth = A * (1 - T) * (1 + H)^N
Value of IRA = A * (1 - R) * (1 + G)^ N

Difference between the two:
A * ((1 - T) * (1 + H)^N - (1 - R) * (1 + G)^N)

Simple case of G == H (i.e., the IRA investments are same as Roth investments)
Difference = (R - T) * A * (1 + G)^N

If R == T (i.e., the conversion tax rate is same as RMD tax rate)
Difference = 0. No advantage.

But if R == T and G != H (i.e., investments are different), then there is an advantage to doing a Roth even if the before and after tax rates are the same. This is the situation that interests me because we are right at the 12/22 bracket threshold and any conversions will be taxed at 22%, same as an RMD withdrawal. But, the bulk of the IRA is bonds with G = 3%, but if I converted I might switch those to equities with H = 7% (or higher, or lower).

I am still mulling over this so am wondering if anyone has any comments on it. I might just do a very small Roth conversion so I get the 5-year clock ticking (it starts with the first Roth conversion regardless of how many others you do).

 

[pb4uski]

My comment would be that if R=T, then the advantage is the $179 in the last paragraph of post #45.

I would not mix the decision analysis of conversion benefits with the change in rates from changing investment mix.

 

[SevenUp]

I am still mulling over this so am wondering if anyone has any comments on it.

As pb4uski noted, there is a difference depending on whether taxes are paid from the traditional account, or from a taxable account.

See Maxing out your retirement accounts and either of the spreadsheets linked there if you'd like too test your own situation.

 

[TwoByFour]

As pb4uski noted, there is a difference depending on whether taxes are paid from the traditional account, or from a taxable account.

See Maxing out your retirement accounts and either of the spreadsheets linked there if you'd like too test your own situation.

I would not be paying conversion tax from a taxable account. The rationale for doing that seems to be wrapped up in maximizing Roth contributions (as per the link to the Boglehead wiki) but that does not apply to my situation.

BTW, that link leads to a portfolio modeling/tracking spread sheet for anyone to use, that is pretty impressive.