Roth conversion confusion

[donheff]

Hello engineers and financial math gurus. I am having a hard time wrapping my head around a Roth conversion optimization analysis and need help. This is a long, confused post, so don't bother with it unless you enjoy sorting out these kinds of issues.

In the past I ignored Roth conversions because I had the vague impression that they were only beneficial if you could take advantage of low tax brackets when you make the conversion and/or higher brackets later when the funds come out. I ran a few TurboTax scenarios using simplistic assumptions of equal tax levels now and later and they came out with it being largely a wash. Intuitively, that makes sense. So I went back to sleep.

Recent posts here by people taking advantage of higher brackets for conversions and the impact of the Secure Act on RMDs of inherited IRAs led me to take a second look. I began to conclude that large conversions might be a good thing. That drew my attention to a change in DC’s estate tax that complicates everything and makes conversions more compelling.

DC recently dropped it’s estate tax limit to levels DW and I are about to bump against. The tax is 16% of assets above the estate limit (so $160K for each $1M over the limit). And, unlike the Federal estate tax, in DC the kids inheriting the IRA pay the full state income tax on the IRA with no deduction for the DC estate tax already paid. Also, the DC exclusion amount is not portable to the surviving spouse like it is at the Federal level. The first to die needs to get their exclusion out of the estate or it is lost. DW and I can deal with that by moving some assets like the house and joint taxable brokerage accounts into a disclaimer trust when the first of us dies. The rest could be excluded by disclaiming sufficient IRAs to reach the DC disclaimer amount and passing them directly to the kids. But that adds to the attractiveness of Roths since conversions drop the size of the estate in general (reducing the amount that exceeds the limit) and our kids are in high tax brackets so inherited Roths would be much better than tIRAs.

I have seen some online calculators that seem to confirm that even large conversions will be ok at long as the tax rates at the end period are close to the rates at conversion. These calculators all assume that the taxes are paid from outside the tIRA but, in our case, I would deduct Fed and State taxes directly from the tIRAs. My taxable funds all have capital gains of about 60% so I don’t want to sell them to pay the tax. And those funds are not sufficient to cover the conversions anyway. Still, the value equivalence seems to roughly hold when paying the tax from the IRA funds. If I have two $1M tIRAs and convert one to a Roth paying $300k in tax, I have a $700K Roth and a $1M tIRA. If I die tomorrow and cash out both, they are worth the same $700k each after tax. If I leave them both in the account to appreciate at the same rate, they will remain equivalent over time, provided the tax rates are the same when they come out. There are lots of things that can happen to change outcomes but the underlying equivalence if the tax rates stay the same seems intuitive.

If the above is fair, it seems that, if I am correct that my heirs’ income tax rates will be the same or higher when the IRAs are inherited, converting as much as possible would make sense, regardless of bracket. The outcome should be a wash or, hopefully, better since the smaller estate size would reduce the DC estate tax. If, in fact, their income tax rates turn out to be significantly lower they could lose out. Does this sound about right, or am I missing something obvious?
.
I don’t want to lay my financials out in detail, but suffice it to say that DW and I are currently right at the beginning of the 24% Federal tax bracket and will remain there for three years at which point large RMDs will kick in. The RMDs will fill the 24% bracket and spill over a bit into the 32% bracket. Over time, with good market growth, the RMDs could eventually kick us right up toward the top. The amounts we are interested in converting far exceed the room available in lower than 37% brackets over the first three years. Assuming we want to wrest the most out of the lower brackets it seems like we should fill them, but what about more? If we need another $2M to help with DC estate tax, should we just bite the bullet and convert a large part of that in year 1 at 37%? Is there a substantial down side to that that I am missing?

I can’t quite wrap my head around this. Is it reasonable to think of all conversions as a wash if you and your heirs will be at high brackets? Does that make a conversion into 37% neutral? If so, the potential advantages to offset DC estate taxes make it attractive. But I fear I am missing something that would make those large immediate conversions less than a wash. I will consult with a CPA and maybe an FA before doing anything but I would like to noodle this around first.

First world problem isn't it? Thoughts?

 

[SecondCor521]

Someone will be paying taxes on that traditional IRA money. Maybe it's you now with Roth conversions. Maybe it's you later with RMDs. Or maybe it's your kids over 10 years when they inherit. (Unless you give some of it to charity now via QCDs or when you die via a charitable bequest of some sort.)

The general idea is to pay taxes at the lowest rate possible. 37% is the top bracket, and at that level of income you may also subject your investment income to NIIT. In general, I would model it in tax software to make sure that the rate you think you're paying is the rate you're actually paying. In other words, consider all of the tax effects of that income. Besides NIIT, there's SS taxability, phaseouts of various things, ACA subsidies (although those are probably long gone), state income taxes, etc.

The only reason I can see that converting to 37% might make sense is if you can actually pay 37% and you think your kids would pay more (either due to additional tax effects mentioned above, or if you think top brackets will move higher). But if brackets drop in general, or your kids drop brackets, you're right in that it would be a losing move.

Another thing that can happen is the value of the investments you have might drop. Converting something worth $1M now at 37% when you might end up converting something worth $500K at 37% later is not a good idea.

But yeah, in general moving something over via Roth conversions sooner gets that compounding out of your traditional IRA / RMD situation. If one of the two of you is noticeably older, converting the older person's IRA might make sense.

You could consider moving to a state which doesn't have such an onerous estate/inheritance tax. Virginia is nice and I think they treat estates and inheritances better than what you described.

Also, if you're sure you'll have enough for yourselves, gifting money sooner rather than later also gets that compounding out of your estate and into your kids' estate, where it will become their problem ;-) You can give $16K per donor/donee combination per year. You can also pay certain education and medical expenses directly above that $16K exclusion.

 

[Montecfo]

Had similar thought to above. Estate taxes saved could finance a profitable move to a more tax-friendly location such as Alexandria or Arlington. My guess (from dealing with DC taxes from the corporate side) is income taxes would also favor such a move.

Just a thought.

 

[ChrisC]

Donheff,

I glanced over your post to get the gist of it. All I can say, is that I'm in a very similar tax bracket to you, with an oversized CSRS pension like you, with TDAs probably higher than you, and I've been converting my wife's TDAs since 2013 resulting in us being in the top of the 24% tax bracket the last 4 years. The net result of these conversions is that her entire TDAs (formerly tIRAs, 401a and 457b) are now all in a Roth IRA -- we did this with the anticipation that she would hit RMDs this year at 70.5 years old -- this was before the Secured Act changed the landscape for RMD threshold age and the availabilty of stretch RMDs for legacy planning.

Starting last year, I started my own conversions from TSP funds (as well as using TSP as a conduit for 401K funds to be rolled into TSP) to Roth IRA). With my conversions, we're again up to the top of the 24% tax bracket, but we don't have to worry about state income taxes for my conversions and when we're both gone, our state does not have an estate tax; We probably have to be concerned with Federal estate taxes when the thresold levels are lowered and if our retirement portfolio gets good growth, which should be the case over 20 plus years for a portfolio in 95 percent equities in Roths!

I think you owe it to yourself to go to the Bogelheads forum and research the numerous past threads and posts there about Roth conversions, especially those generated from the academic research done over the last several months by Professor Edward McQuarrie. There's quite a bit discussion and useful information there that might help guide you. A number of forum posters there have developed spreadsheet tools to assist folks. Some say, you can only get the most optimal decision in gauging conversion outcomes by doing detailed spreadsheet driven analysis. For me, there are simply too many variables to precisely predict optimal outcomes, so I just try to arrive at a decent result.

 

[CardsFan]

Had similar thought to above. Estate taxes saved could finance a profitable move to a more tax-friendly location such as Alexandria or Arlington. My guess (from dealing with DC taxes from the corporate side) is income taxes would also favor such a move.

Just a thought.

See Bold.

That was actually my first thought as I read through the post.

 

[ChrisC]

Had similar thought to above. Estate taxes saved could finance a profitable move to a more tax-friendly location such as Alexandria or Arlington. My guess (from dealing with DC taxes from the corporate side) is income taxes would also favor such a move.

Just a thought.

Many people in DC have multiple properties in DC, MD and VA so if you are fortunate to be one of those people, you might not even have to physically move -- just give up the DC residence as your domicile. That's really a good problem to have as well as the Roth conversion issues.

 

[RunningBum]

Been offline for close to 24 hours so I'm not looking for any nuances. I'll let you reply to the suggestions about losing the DC residency.

I don’t want to lay my financials out in detail, but suffice it to say that DW and I are currently right at the beginning of the 24% Federal tax bracket and will remain there for three years at which point large RMDs will kick in. The RMDs will fill the 24% bracket and spill over a bit into the 32% bracket. Over time, with good market growth, the RMDs could eventually kick us right up toward the top.

So if you only spill over a bit into 32%, it seems that a large RMD (not necessarily to 37%) would keep you from spilling over at all. Maybe filling the 24% bracket for the next 3 years is enough? Then RMDs would fill up 24% without overflowing?

But there's growth. Is most of your money in your tIRA, or do you have significant amounts in Roths and/or taxable? What I'm looking for is keeping your fixed income allocation that you have to reduce volatility in your tIRA, and the higher flyers in taxable and Roth. That would keep your tIRA from growing too fast, while also putting tax inefficient fixed income there as well.

If most of your wealth is in your tIRA, then I suggest you create a spreadsheet that lets you see how various conversions work with various growth rates and to find the best conversion rate using the assumptions you want to go with.

Don't forget to factor in IRMAA tiers. You may find that keeping just under the next IRMAA tier is the right thing especially if other results point to converting in the middle of a tax bracket.

 

[ChrisC]

Been offline for close to 24 hours so I'm not looking for any nuances. I'll let you reply to the suggestions about losing the DC residency.


So if you only spill over a bit into 32%, it seems that a large RMD (not necessarily to 37%) would keep you from spilling over at all. Maybe filling the 24% bracket for the next 3 years is enough? Then RMDs would fill up 24% without overflowing?

But there's growth. Is most of your money in your tIRA, or do you have significant amounts in Roths and/or taxable? What I'm looking for is keeping your fixed income allocation that you have to reduce volatility in your tIRA, and the higher flyers in taxable and Roth. That would keep your tIRA from growing too fast, while also putting tax inefficient fixed income there as well.

If most of your wealth is in your tIRA, then I suggest you create a spreadsheet that lets you see how various conversions work with various growth rates and to find the best conversion rate using the assumptions you want to go with.

Don't forget to factor in IRMAA tiers. You may find that keeping just under the next IRMAA tier is the right thing especially if other results point to converting in the middle of a tax bracket.

The OP might not have Medicare Part B. Many Federal annuitants forego Medicare Part B and rely solely on FEHB health insurance. I only have FEHB coverage though my wife has Medicare Part B and we’re thinking of dropping Part B and IRMAA tier 3 premiums.

 

[Exchme]

I agree that hiring a one-time-fee financial planner may make sense(as opposed to a money manager that takes a percentage each year, those are just rip-offs). Note that not all CPA's do financial or estate planning, for instance mine does a good job with taxes, but would not be good for planning.

If you want to do a DIY analysis, the tool I use is Pralana Gold. It's a paid Excel spreadsheet, but very flexible and handles lots of situations and has an excellent tax package. For instance, to simulate paying conversion taxes out of your t-IRA, you would enter a value for an additional withdrawal from the t-IRA of whatever amount you want in whichever year or years. If you have tried free tools, you may note that some of them do a poor job on taxes that affect higher incomes, many miss NIIT, 20% capital gains bracket, AMT or deduction phaseouts for instance. For Roth conversion simulation in Pralana, you can enter a tax rate for your heirs to withdraw the t-IRA or another trick is to add 10 years to your longest expected life span and then enter additional withdrawals from the IRA during those years to zero out the account.

You may ultimately decide that the local estate tax dominates and you just need to convert what you can in order to avoid that tax on the embedded income tax liability present in the t-IRA. Alternately, you may decide that generosity is the best plan and do QCDs starting at age 70.5 or set up a DAF or maybe throw in some gifts to the kids of highly appreciated shares - even if they don't need it, it might be good to move the money and its future growth out of your estate.

Congratulations on having put enough away for a comfortable retirement.

 

[Animorph]

If the tax rate stays constant, then yes, Roth converting and paying taxes from the tIRA withdrawal is a mathematical wash.

However if you do that, you are now eligible to contribute to the Roth the difference between what you withdrew from the tIRA and what you rolled over into the Roth. You need to decide what that's worth to you, tax free growth on previously taxable investments for some number of years minus the cost of paying capital gains a little early. If your equity sales will come in the next few years and you may never withdraw from the Roth, that could be a nice benefit.

 

[pb4uski]

Had similar thought to above. Estate taxes saved could finance a profitable move to a more tax-friendly location such as Alexandria or Arlington. My guess (from dealing with DC taxes from the corporate side) is income taxes would also favor such a move.

Just a thought.

I had the exact same thought as I read the OP. DC isn't that big so I'm guessing you wouldn't have to move very far to escape DC estate taxes.

 

[Dash man]

Other things to consider:
The TCJA sunsets in 2026, reverting to the old tax rates. This will likely result in higher taxes in the future.
If one spouse were to die sooner than expected, the survivor will be filing as an individual moving them into the higher tax and IRMAA brackets sooner.
Federal Estate taxes will also revert in 2026, and that will put us on the bubble.

I use Pralana Gold and it indicates in my situation to fully convert over the next few years before we hit age 70 and DW begins her SS. I’ve also done projections on spreadsheets and come to the same conclusion. We live in PA that doesn’t tax retirement accounts and we’re only concerned with the 4.5% Inheritance tax for our sons.

If you hire a tax advisor, make sure they specialize in retirement tax planning.

 

[Magus]

I think its unlikely the TCJA tax rates go away - maybe the top 1% bracket goes back up - but politicians haven't actually directly increased middle class and upper middle classes taxes in a very, very long time. If anything, I think tax brackets go down. Don't forget as you work on forecasting taxes out to increase the brackets for inflation.

One other point to consider depending on your net worth is that the tax free status of the Roth can be revoked at any point. 2021 was the first time politicians seriously broached doing that and you can bet it will come up again. It will undoubtably start with incredibly high balances (much like income taxes started out as 1% on millionaire equiv incomes) but will very rapidly be adjusted.

 

[Dash man]

I think its unlikely the TCJA tax rates go away - maybe the top 1% bracket goes back up.

One other point to consider depending on your net worth is that the tax free status of the Roth can be revoked at any point.

I’m a believer in planning using the tax laws currently in place. There are all sorts of proposals made that never become law.

Chances of Roth IRAs becoming taxable are slim because politicians love them. They love Roth conversions because it helps them get our tax money out of 401k and tIRA accounts sooner.

 

[Magus]

I’m a believer in planning using the tax laws currently in place. There are all sorts of proposals made that never become law.

Chances of Roth IRAs becoming taxable are slim because politicians love them. They love Roth conversions because it helps them get our tax money out of 401k and tIRA accounts sooner.

1) We have a very long history of never letting tax rates sunset for < 99% of the US. In fact, the history of tax rates in the US for the < 99% has been straight down for the last 60 years, which is why half of Americans no longer pay any taxes and even most upper middle class and lower upper class have a low teen effective rate (and sometimes single digit). In fact, I would bet if anything as we get closer to this sunsetting, the politicians will be fighting about "lowering middle class taxes" further rather than increasing them.
2) We were 2 senators away just months ago from having a Roth tax for the first time.

 

[target2019]

I put your numbers into Flexible Retirement Planner. Below is your tIRA and Roth, with an effective tax rate of 30%. No SSA, no pension.

There are two observations that I have from the many discussions.
1) I am trying to reduce the impact of tax torpedo at RMD, and then another when I perish and surviving spouse takes over.
2) If I can reduce the entire tIRA, I lessen the tax impact to some extent.
3) There isn't a single number I can come up with to say go or no go with a conversion. My final analysis was to do something with conversion to help my surviving spouse. I cannot make all taxes go away, but in the end I can say I did my best.

This picture doesn't capture all of your situation, but I think it highlights the effect. YMMV.

 

[Midpack]

I think its unlikely the TCJA tax rates go away - maybe the top 1% bracket goes back up - but politicians haven't actually directly increased middle class and upper middle classes taxes in a very, very long time. If anything, I think tax brackets go down. Don't forget as you work on forecasting taxes out to increase the brackets for inflation.

One other point to consider depending on your net worth is that the tax free status of the Roth can be revoked at any point. 2021 was the first time politicians seriously broached doing that and you can bet it will come up again. It will undoubtably start with incredibly high balances (much like income taxes started out as 1% on millionaire equiv incomes) but will very rapidly be adjusted.

You may well be right, but I don’t see how we can keep running higher and higher deficits/debt. So I am assuming tax rates have to increase one way or another, and there aren’t enough 1%ers to make up the difference. Increasing tax rates on corporations isn’t the answer either, as those costs ultimately go to customers/individual taxpayers anyway. Most of us are going to have to pay more, or accept less from government which seems more unlikely.

All this is a big factor in whether Roth conversions make sense…

 

[Magus]

You may well be right, but I don’t see how we can keep running higher and higher deficits/debt. So I am assuming tax rates have to increase one way or another, and there aren’t enough 1%ers to make up the difference.

It helps with inflation near 10% and your average debt cost is 2%

I'm guessing it will be a combination of going after top 1% income (maybe top 3%), inflating our way out, taxing net worth to some degree (eg: High balance Roths) and not just income, probably raising or even eliminating the max social security tax cap and some import tariffs and gasps - perhaps marginally cut spending.

Either way, I'd bet $10k at 5:1 odds there is a higher chance of high net worth's roth's accounts being taxed in 7 years than those in the 22-24% marginal rate today having their taxes increased.

 

[donheff]

Thanks for the suggestions. They have given me some good thoughts to mull over.

Just a couple of responses to suggestions;

IRMAA is not an issue since we don't have Part B

Move to VA: Very attractive, but DW and I love Capitol Hill so we are staying place - screw the kids. DW is very committed to aging in place and we are well situated to do so. In the document I am working on providing guidance for DW and the kids to deal with things if I am not competent, I tell the kids that if DW dies and I am declining, pack me away in a CCRC in VA and move into the house or rent it.

Pralana Gold: thanks. I will look at at.

Flexible Retirement Planner: I will look at that one also.

Just as a follow-up thought. I stated at the outset that I was seeing a wash because current and exit tax rates are about the same. I ran some TurboTax scenarios and the wash principle works for my son. He would probably be as good or better if we just converted the whole lot over a few years and damn the brackets. But DD is a different matter. She is a teacher and her husband is currently in a similar paying job. For them a slower approach that keeps conversions out of 37% and maybe lower is better.

I will keep playing with the numbers until I am comfortable but we will clearly be converting through 24% and probably higher over the next three years. We should have been doing it for the past few years while we could have. But that will only get a portion of the IRAs converted. If we convert thru 32% for three years, DW's RMDs will still fill most of the 24% bracket. I didn't mention it in the post above but the tIRAs constitute about 80% of our portfolio and are almost all DWs. She was a lawyer.

 

[donheff]

I agree that hiring a one-time-fee financial planner may make sense...

If you want to do a DIY analysis, the tool I use is Pralana Gold.

Wow. I got Pralana Gold and it is amazing. It basically confirmed my intuitive conclusion that the pay not or pay later washes out. If you live a few years the end results are almost identical including total estate and total taxes paid. The difference is the mix between tax differed, taxable, and Roths.

I plan to run a lot more simulations with this and then talk to both a CPA and a fee only FA to discuss whether there are gotchas I am missing. But big conversion(s) near-term are probably in our future.

Thanks for the recommendation.

 

[ncbill]

Wow. I got Pralana Gold and it is amazing. It basically confirmed my intuitive conclusion that the pay not or pay later washes out. If you live a few years the end results are almost identical including total estate and total taxes paid. The difference is the mix between tax differed, taxable, and Roths.

I plan to run a lot more simulations with this and then talk to both a CPA and a fee only FA to discuss whether there are gotchas I am missing. But big conversion(s) near-term are probably in our future.

Thanks for the recommendation.

Roth gives your heirs another decade of tax-free growth...

 

[RockyMtn]

question for you Roth Conversion experts...when converting to the top of the tax bracket do you use effective tax rate % or tax bracket % to decide whether or not to do?

For instance, I am in 35% tax bracket but after deductions, donations, etc my effective tax rate is closer to 16%. So, would I convert up to the top of the 22% bracket or up to top of 35% bracket?

 

[Magus]

question for you Roth Conversion experts...when converting to the top of the tax bracket do you use effective tax rate % or tax bracket % to decide whether or not to do?

For instance, I am in 35% tax bracket but after deductions, donations, etc my effective tax rate is closer to 16%. So, would I convert up to the top of the 22% bracket or up to top of 35% bracket?

For roth conversions, you don't use effective rate, its marginal rate that matters. That's why for most people t401k are better because you start top down at peak earnings years while working for the deduction and bottom up in retirement. If you are in the marginal 22% rate, could be worth doing a bit. But if you are in the marginal 35% fed rate, I would not convert anything.

 

[RockyMtn]

For roth conversions, you don't use effective rate, its marginal rate that matters. That's why for most people t401k are better because you start top down at peak earnings years while working for the deduction and bottom up in retirement. If you are in the marginal 22% rate, could be worth doing a bit. But if you are in the marginal 35% fed rate, I would not convert anything.

Not sure I totally understand your comment.

In essence, when you put away income in a traditional pre-tax account and take the deduction now, you are saving at the marginal tax rate. If you are in the 35% tax bracket, you save paying 35 cents for every dollar you defer.

Later, when withdrawing funds, you pay taxes at your effective tax rate which in my case is half of my marginal rate.

 

[Out-to-Lunch]

Not sure I totally understand your comment.

In essence, when you put away income in a traditional pre-tax account and take the deduction now, you are saving at the marginal tax rate. If you are in the 35% tax bracket, you save paying 35 cents for every dollar you defer.

Later, when withdrawing funds, you pay taxes at your effective tax rate which in my case is half of my marginal rate.

Your argument may hold water for the FIRST time you make a contribution, but, if you have successfuly saved for retirement, it becomes irrelevant. By then, your bottom brackets will be filled no matter what, so the future effective rate isn't the one to focus on.

See: https://www.bogleheads.org/wiki/Traditional_versus_Roth#Common_misconceptions