What tax bracket should I stop doing Roth Conversions

To each there own. I'm seeing some good reasons to do Roth Conversions.
1) My wife will be in a higher bracket when I die.

To me, this is one of the most compelling reasons. In my family, the males usually pass 15-20 years before the women. Between the pensions, SS, and the RMDs she will go from our current 24% to the 35%. To me, this is reason enough to at least do some.
 
Distribution timeline for Roth (which were converted from tIRA) and tIRA are the same, or nearly so, correct? See post #46.

This means there is no material difference to heirs as it relates to inherited distributions, whether from a Roth or a tIRA, correct?

Therefore there is no advantage, from the point of view of an heir, to inherit a Roth or a tIRA from the decedent, correct?

Wrong!

As an example:

Say I have $1,000,000 in a tIRA, and convert that at a 15% effective tax rate over a period of time. The heir would get $850,000 TAX FREE (plus what ever growth), and could let it grow TAX FREE for up to 10 more years.

If the heir inherited a $1,000,000 tIRA they would have to pay taxes at their incremental tax rate on every withdrawal. Say they elect to empty the account over 10 years. If they are working, the incremental rate could easily be 22% and maybe 24% or 32%. So they would net less than $800,000 THEN, if they don't need the money, and invest it, they would need to pay taxes on any earnings.

I am not going to bother to calculate it, but the net effect is much greater when you take the growth into account.

I am not sure why this is so hard to grasp.
 
... Therefore there is no advantage, from the point of view of an heir, to inherit a Roth or a tIRA from the decedent, correct?
Not that this isn't complicated enough, I'll give you an edge case that we actually have. Three grands, all getting their money from via trusts, one of them a special needs beneficiary. The two get their money from a tIRA over ten years, never taxed at the trust level (35% IIRC) because any realized gains will be distributed during the tax year and taxed at the beneficiary's rate.

The special needs trust, though needs to hold assets longer than ten years, so it will be funded by Roths. It is still subject to the 10 year rule, but will hold the distributions rather than paying them out immediately. The result if they were tIRA distributions would be paying income tax at the high trust rate. Since the distributions will come from Roths, however, no taxes will be due except to the extent that the trust has investment profits after the 10 year point.

Said another way, the practical lifetime of a trust funded with tIRAs is ten years.
 
Distribution timeline for Roth (which were converted from tIRA) and tIRA are the same, or nearly so, correct? See post #46.

This means there is no material difference to heirs as it relates to inherited distributions, whether from a Roth or a tIRA, correct?

Therefore there is no advantage, from the point of view of an heir, to inherit a Roth or a tIRA from the decedent, correct?

Your first statement is correct.

However, distributions from a traditional IRA are generally taxable income to the heir. With the new 10-year rule and considering the typical age at death, most heirs will probably have their 10 year period overlap with the last 10 years of their working career, when their income is probably the highest. Adding a distribution from a traditional IRA on top of that means that those distributions will be taxed at a high marginal rate.

Distributions from a Roth IRA are generally tax free to the heir and do not involve the same tax complication as in the previous paragraph.

So even though the timeline for distribution is the same, the tax treatment of the two are quite different. Which means that the answers to your second and third questions are, IMHO, "No".

(Whether one is better than the other and the impact of that on whether to do Roth conversions is a separate point.)
 
Spending is different than managing taxes though. If I can pay taxes on deferred income at a lower rate by converting when I'm in ER before I start SS, I can withdraw that money tax-free from my Roth IRA to live on.
Yeah, but there isn't a point to dip into a higher tax bracket now than you'll ever likely have to touch later once RMDs hit which is why I'd stay in the 12% given their numbers. If you go into the 22% now, you're assuming that you'll be in the 24% bracket later on which seems unlikely with $1.2m in a tIRA. A lot of people assume that RMDs will force you to withdraw massive amounts of money right away and pay lots of taxes once they hit which isn't the case.
 
Last edited:
Yeah, but there isn't a point to dip into a higher tax bracket now than you'll ever likely have to touch later once RMDs hit which is why I'd stay in the 12% given their numbers. If you go into the 22% now, you're assuming that you'll be in the 24% bracket later on which seems unlikely with $1.2m in a tIRA. A lot of people assume that RMDs will force you to withdraw massive amounts of money right away and pay lots of taxes once they hit which isn't the case.
Which is why I said to do some conversion now if you can do it at a lower rate. Converting at the same rate is perfectly fine too, and an advantage if you can pay the taxes out of a taxable account.

Nobody should assume what RMDs are going to be. One should estimate what the tax deferred balance will be at 72 and what the RMD amount will be based on the RMD table. Where are you seeing people assume the RMD will be massive? I hadn't seen that. But when combined with SS and possibly a pension, there usually is an income increase.
 
New study from Edward McQuarrie, emeritus professor at Santa Clara:
https://www.marketwatch.com/story/to-roth-or-not-to-roth-11623431970

Link to his blog that includes link to his spreadsheet:

Roth Conversions – Essays, Opinion & Advice

Long paper, I haven't gone through it all yet. His basic point seems to be that Roth conversions gain less than people think.

Factors that he did not examine that would make them better are death of a spouse, leaving the other in a high tax bracket and passing inheritances that take the annual gain of a Roth and compound for it for another decade. Those are factors that I'm interested in, so not sure the conclusions apply completely for me (still reading the paper).

Still I'm glad for some academic interest in a real world topic.
 
New study from Edward McQuarrie, emeritus professor at Santa Clara:
https://www.marketwatch.com/story/to-roth-or-not-to-roth-11623431970

Link to his blog that includes link to his spreadsheet:

Roth Conversions – Essays, Opinion & Advice

Long paper, I haven't gone through it all yet. His basic point seems to be that Roth conversions gain less than people think.

Factors that he did not examine that would make them better are death of a spouse, leaving the other in a high tax bracket and passing inheritances that take the annual gain of a Roth and compound for it for another decade. Those are factors that I'm interested in, so not sure the conclusions apply completely for me (still reading the paper).

Still I'm glad for some academic interest in a real world topic.

@Exchme Thanks, I agree that Roth conversions "gain less", or rather don't benefit as much as, people think. I'll work through the paper. Thanks for posting.

My view is that Roth conversions don't gain many people anything at all. But rather they are subtractive to net worth during the planning horizon, until demise of the principals.
 
+1. After exhaustive research, including detailed analysis using Income Strategy software, I believe the above is exactly right. The goals isn’t one tax bracket, it’s keeping your tax bracket constant before and after RMDs. ....

I agree Midpack. My plan assumes no rate changes but would have an effective rate that is within 1.2% each year from now until I'm 90.

One caveat that might apply in some cases.... where the heirs are very financially successful and in a higher tax bracket it might be optimal to do some conversions higher the the IRA owner's ultimate tax bracket since any heirs would need to drain the tIRA sometime in 10-11 years after inheriting the tIRA and those distributions may be at a higher tax rate than what the owner was paying.

I'm not sure how frequently this situation will apply but it will in some situations. I probably wouldn't start even considering this until I'm in my late 70s or early 80s.
 
Last edited:
Roth conversions are essentially tax rate arbitrage where you don't know the future tax rate. Stop when you get to the point where your tax rate equals your expected future tax rate. If you are right, you win. Conversely ...

This is absent other considerations like ACA and IRMAA which effectively raise your current tax rate and RMDs, which may raise your future tax rate. Having fun yet?

Another +1 for this comment. Trying to decide what tax bracket to stop in is not the right question.

Also, the people who talk about higher taxes in the future don't often consider that taxes can be raised in different ways. What if the big future tax increase is not an income tax increase, but the implementation of a national sales tax? Maybe income taxes actually go down. Now your Roth conversions haven't helped at all.

It's difficult to make predictions, especially about the future.
 
New study from Edward McQuarrie, emeritus professor at Santa Clara:
https://www.marketwatch.com/story/to-roth-or-not-to-roth-11623431970

Link to his blog that includes link to his spreadsheet:

Roth Conversions – Essays, Opinion & Advice

Long paper, I haven't gone through it all yet. His basic point seems to be that Roth conversions gain less than people think.

Factors that he did not examine that would make them better are death of a spouse, leaving the other in a high tax bracket and passing inheritances that take the annual gain of a Roth and compound for it for another decade. Those are factors that I'm interested in, so not sure the conclusions apply completely for me (still reading the paper).

Still I'm glad for some academic interest in a real world topic.
Good find. I look forward to reading this. Hopefully he quantifies what he means by "gains less" or what he thinks people think the gain will be. I'm certain doing conversions is not going to be a 6-figure gainer for me, but I'm pretty sure it will be easily into 5-figures. I'll refrain from any other comments until I've hard a chance to read this thoroughly.

This seems worthy of a thread of its own, rather than an existing thread for an individual situation.
 
Also, the people who talk about higher taxes in the future don't often consider that taxes can be raised in different ways. What if the big future tax increase is not an income tax increase, but the implementation of a national sales tax? Maybe income taxes actually go down. Now your Roth conversions haven't helped at all.
Yes, it is possible. But you'd be speculating at there being a major change in taxation in the US. Taking the bird in the hand by converting now seems less risky to me. Just my view.
 
I use net worth as the singular parameter to maximize in my Excel model. Accelerating taxes by forcing income through Roth conversions results in a lower net work for me across each year of my life expectancy. Who cares if I bump into a higher tax bracket at age 72+? Net worth until "end of plan" is what matters to me. I am MFJ and assume my wife will outlive me, and her life expectancy is "end of plan".

Chassis, you have obviously spent a lot of time thinking about this, and I would like to learn more about your thought process. (I have also spent a lot of time; I come out in favor of Roth conversions in my case, but I agree that the effect is not huge.)

But my particular question is on your calculation of "net worth." In your calculation, do you account for the (differing) value of dollars in Roth vs. tax-deferred accounts? It seems obvious to me that, in comparing my net worth with or without conversions, I need to discount the value of the tax-deferred accounts for taxes yet to be paid. Do you do this in your case?
 
+1 if you're going to use net worth as a metric, then you need to include a deferred tax liability for your tax-deferred accounts times the tax rate that you expect to pay on withdrawals and RMDs. Measuring it isn't difficult at all, the difficulty determing what tax rate or rates to use.

If you do this, then when you do a conversion your net worth doesn't go down.

Example: $100 cash, $1,000 tIRA, 20% tax rate and $200 Roth conversion

With no DTL: Net worth before Roth conversion is $1,100 ($100 cash, $1,000 tIRA)
...................Net worth after Roth conversions is $1,060 ($60 cash, $800 tIRA, $200 Roth)

With DTL: Net worth before Roth conversion is $900 ($100 cash, $1,000 tIRA, $200 DTL)
...............Net worth after Roth conversion is $900 ($60 cash, $800 tRIA, $200 Roth, $160 DTL)
 
Last edited:
@Out-to-Lunch and @pb4uski Great comments.

Net worth = Assets - Liabilities

Assets for me are tangible and financial, valued in nominal dollars. I account for inflation in my model, and everything I calculate is in nominal dollars.

I don't see that putting deferred tax liabilities on the balance sheet is relevant.

In my Roth convert/do not convert calculator, I measure net worth with and without Roth conversions. Federal income tax is paid in the current year, in nominal dollars, according to current tax law. RMDs are distributed, and taxed, as required by current tax law.

Some points where my view of the world may differ from others:

- I have strong views of market/portfolio return based on my investing style
- Income in excess of living expenses is reinvested, for example after-tax RMD income or converted Roth dollars
- I will deal with future tax rate changes, if they come and to the extent they affect me, when they happen

@Out-to-Lunch what do you mean by "differing value"? Are you talking about pre- vs after-tax or are you talking about inflated/nominal vs real? Or something else?

Do any of us need to qualify ourselves, relative to education or professional experience on the topic? I have hesitated to do this. At times I feel like it would move things forward, other times I feel like it would raise resentments and more discord. I am divided on this question.
 
Last edited:
Thanks for writing that up.

@Out-to-Lunch what do you mean by "differing value"? Are you talking about pre- vs after-tax or are you talking about inflated/nominal vs real? Or something else?

Oh, I only meant due to the tax liability on one, and not on the other. (Nothing to do with inflation or anything else.)

I just meant that if I have $100 in a Roth, I can spend $100, so it is "worth" $100 to me. If I have $100 in a tIRA, I can only spend about $75 of it, so it is worth less to me than $100 in a Roth (or taxable).
 
I generally agree that many get little benefit. roth conversions

But I expect many can get benefit from roth conversions.

Take a couple that did that did roth conversions and the the first spouse dies with no effect on income. Taxes go up due to single filer.

The conversion itself may not may not be the whole picture.
 
I generally agree that many get little benefit. roth conversions

But I expect many can get benefit from roth conversions.

Take a couple that did that did roth conversions and the the first spouse dies with no effect on income. Taxes go up due to single filer.

The conversion itself may not may not be the whole picture.

@bingybear I see the mechanics of how it would work.

Spouse 1 dies.
Spouse 2 loses MFJ status and now claims single. Effective tax rate goes up.

Case a: with Roth conversions, and a fully depleted tIRA, single spouse has no taxable income from the portfolio. footnote: SS and other income (e.g. pension) need to be treated, see below.

Case b: without Roth conversions, with without sufficient after tax cash, tIRA withdrawals are needed and are taxed at the higher single filer rate. And RMDs are in play for this case.

But this all needs to get pencilled out in a model. How many of you are doing this? The world does indeed include Social Security, pensions for some, and other sources of income and net worth.

@bingybear do you have a thorough and detailed model that shows you the effect of Roth conversions, Social Security and your other financial moving pieces?

I have such a model for myself. It's not 100% perfect and I add detail and fidelity as I go. I trust it for the big picture, and for many of the fine details. Roth conversions don't interest me based on the model.

There are so many claimed reasons for doing Roth conversions, yet when a detailed model is built, none of the reasons hold water for me. At least in my case. However my case is not unique. The one case I haven't fully sorted is the impact to heirs when the estate is passed on. I may not care about this because it's too far into the future to worry about.
 
Last edited:
Roth conversions are essentially tax rate arbitrage where you don't know the future tax rate. Stop when you get to the point where your tax rate equals your expected future tax rate. If you are right, you win. Conversely ...

This is absent other considerations like ACA and IRMAA which effectively raise your current tax rate and RMDs, which may raise your future tax rate. Having fun yet?

That's an interesting perspective. Many of us are maxing out are pre-tax retirement accounts so are using Roths to invest after-tax money that would be invested in a regular after-tax account.

How can it hurt putting after-tax money into a Roth that would have been invested in an after-tax account anyway?
 
New study from Edward McQuarrie, emeritus professor at Santa Clara:
https://www.marketwatch.com/story/to-roth-or-not-to-roth-11623431970

Link to his blog that includes link to his spreadsheet:

Roth Conversions – Essays, Opinion & Advice

Long paper, I haven't gone through it all yet. His basic point seems to be that Roth conversions gain less than people think.

Factors that he did not examine that would make them better are death of a spouse, leaving the other in a high tax bracket and passing inheritances that take the annual gain of a Roth and compound for it for another decade. Those are factors that I'm interested in, so not sure the conclusions apply completely for me (still reading the paper).

Still I'm glad for some academic interest in a real world topic.

Those who have achieved ER can do what we'll be doing:

Both of us over 50, me not currently working, spouse deferring nearly all their "leftover" pay (after deductions for SS/Medicare, HSA, employee health insurance premium) to their employer's traditional tax-deferred retirement plan means we will have very little taxable income for the next several years.

Given the above there's plenty of room for me to convert my rollover IRA to Roth, up through the top of the 12% bracket.
When spouse retires (probably age 60) they'll rollover their plan assets to a traditional IRA & then do the same.

So we'll end up with our retirement accounts converted to Roth years before having to worry about RMDs or whether our SS is taxable.
 
Last edited:
...I have such a model for myself. It's not 100% perfect and I add detail and fidelity as I go. I trust it for the big picture, and for many of the fine details. Roth conversions don't interest me based on the model.

There are so many claimed reasons for doing Roth conversions, yet when a detailed model is built, none of the reasons hold water for me. At least in my case. However my case is not unique. The one case I haven't fully sorted is the impact to heirs when the estate is passed on. I may not care about this because it's too far into the future to worry about.

I think we've discussed this before. If you're playing the 12% tax bracket vs the 22% tax bracket then there are benefits, but if you're playing the 22% tax bracket vs the 24% tax bracket then it's pretty much a nothing burger.

I do have a pretty comprehensive model from my current age until I'm 90. The model includes a simplified tax return calculation to get the tax for each year.

If someone inherits an tIRA they have to drain it by December 31st of the 10th year subsequent to the owners death, so that money will effectively be taxed at the heirs marginal tax rates. That's another good reason for me to do Roth conversions as the tax rates that I'm converting at are significantly lower than the marginal tax rates of my kids.
 
... How can it hurt putting after-tax money into a Roth that would have been invested in an after-tax account anyway?
It can't. My point relates strictly to looking at conversions, which is a decision whether to accelerate payment of the income tax or not.

If you have cash where the tax has already been paid and you have the option of putting it in a Roth, that seems like a no-brainer for almost any situation.
 
Last edited:
I think we've discussed this before. If you're playing the 12% tax bracket vs the 22% tax bracket then there are benefits, but if you're playing the 22% tax bracket vs the 24% tax bracket then it's pretty much a nothing burger.

I do have a pretty comprehensive model from my current age until I'm 90. The model includes a simplified tax return calculation to get the tax for each year.

If someone inherits an tIRA they have to drain it by December 31st of the 10th year subsequent to the owners death, so that money will effectively be taxed at the heirs marginal tax rates. That's another good reason for me to do Roth conversions as the tax rates that I'm converting at are significantly lower than the marginal tax rates of my kids.

I follow pretty much the same path.

We will be in the 22-24% bracket (or what ever it is changed to) without question, when we hit RMD's. But one target I have is potentially staying below the IRMMA brackets in the future. So there is a minor potential benefit there. even converting into the 22% bracket.

I don't have a comprehensive model, and not sure I really need one for this decision.

My biggest driver is your last point: Heirs. Or in our case, just one heir. No matter what I do, he will inherit a significant tIRA (unless one of us lives to 109:D). What I CAN do is pre-pay taxes at a lower rate to leave him, and DDIL, more money, after taxes. The new 10 year depletion requirement makes this a much bigger factor. Even converting at a 24% effective rate is no worse than a wash.

I guess, if you are single, with no one you really want to leave the money to, it just doesn't matter what you do.
 
If someone considers $100K in a tIRA to have the same value as $100K in a Roth, there's just no sense in discussing conversions with them. You cannot spend a dollar out of your tIRA unless you pay the tax on it. That money is not all yours, just like your gross pay when you were working was not what you took home.
 
Thanks for writing that up.



Oh, I only meant due to the tax liability on one, and not on the other. (Nothing to do with inflation or anything else.)

I just meant that if I have $100 in a Roth, I can spend $100, so it is "worth" $100 to me. If I have $100 in a tIRA, I can only spend about $75 of it, so it is worth less to me than $100 in a Roth (or taxable).

Sure, you can only spend the part that the the IRS doesn't take first. Naturally the IRS' share is somewhat uncertain as many things can happen, but it should be pretty easy for most people to make an estimate. Once you know your general tax rates that RMDs/Roth conversions and your heir's conversions will cause, then you can fill up the brackets, evaluate ACA subsidies and whatever else makes sense. At least in my case, the benefits hit a broad, relatively flat optimum as long as I'm in the general vicinity.

Chassis is right that having a year by year model is a really good idea, I don't know how people do it without. Of the model packages I've looked at, from simplest to most comprehensive, I-orp (extended version), Boglehead's Retiree Portfolio model or Pralana Gold (costs money) are good packages.
 
Back
Top Bottom