Roth Conversion Guide

[AtlasShrugged]

All,

I thought I'd share another one of my papers. This one is a Roth Conversion Guide. I noticed some other threads on this topic recently and thought now would be a good time to share my work.

I've spent way too much time researching Roth conversions and trying to summarize my results clearly. Roth conversions are incredibly complicated, so while I'd like to think my guide is clear, you will have to be the judge.

My guide does not argue for or against Roth conversions. Its goal is to provide information to help you make a decision for yourself. It is divided into three sections that explain: i) Roth tax rules; ii) conversion considerations; and iii) how to make Roth conversions.

I hope you find it useful.

Here's the link to it.

https://www.dropbox.com/s/istw0kr1tsckzw8/Roth Conversion Guide.pdf?dl=0

Rick

About me: I'm a retired CPA and CFO with an MBA in Finance. I am not a tax expert. I just enjoy writing about this stuff.

 

[RunningBum]

Nice work! Another fair and balanced study, clearly presented.

 

[friar1610]

Thank you very much for this and your previous articles. I’ve found them very helpful.

 

[pb4uski]

Excellent! A couple thoughts:

An example of the communicative property might be helpful, but there is a wrinkle... a second order effect that makes keeping the assets in the tIRA suboptimal compared to converting earlier.

For example, let's say you have $100,000 in a tIRA, $22,000 in a taxable account, a 22% marginal tax rate and expect that your investments will double in 10 years (earn 7.18%).

If you convert and use the after-tax funds to pay the $22,000 in taxes due then the Roth will grow from $100,000 to $200,000 of after-tax value.

OTOH, if you do nothing, the $100,000 tIRA will double to $200,000 and when withdrawn $44,000 of taxes will be due which will be paid with the after-tax money and growth... so the end result is $200,000 of after-tax value... so at first blush it appears to be a wash.

But... the taxable account will really not grow to $44,000 because the 7.18% growth is taxed at 22% each year rather than tax-deferred like in the tIRA or tax-free in the Roth. In reality, the $22,000 taxable account will only grow to $37,931* rather than $44,000, so after withdrawing $200,000 from the tIRA and paying $44,000 in taxes, the after-tax value will be only $193,931 rather than $200,000... so there is a $6,069 benefit of converting earlier rather than later even if taxes are paid from taxable account money.

* $37,913 = $22,000*(1+7.18%*(1-22%))^10

Second, I've found the What-If Worksheet in TurboTax to be very useful in doing the marginal tax impact calculations since it can incorporate taxability of SS, phase-outs, preferenced income taxes, and.. I suspect... NITT and ACA subsidies. Obviously, it can't include IRMAA impacts because those are not part of the tax return.

One of the things that I have been amazed at is that an early retiree can have a six-figure income and pay very little tax. Let's say an early retiree is living off of taxable savings in a municipal-bond fund and for convenience we'll assume that capital gains from municipal bond fund sales to provide for spending are de minimus. A ER married couple under 65 in 2020 could do a $105,050 Roth conversion and only owe $9,241 (8.8%) in tax because some of the conversion is offset by the standard deduction, some is at 10% and the remainder at 12%... given that FAs and others often talk about tax rates of 20% or more it seems too good to be true, but it is true.

 

[AtlasShrugged]

Excellent! A couple thoughts:

An example of the communicative property might be helpful, but there is a wrinkle... a second order effect that makes keeping the assets in the tIRA suboptimal compared to converting earlier.

For example, let's say you have $100,000 in a tIRA, $22,000 in a taxable account, a 22% marginal tax rate and expect that your investments will double in 10 years (earn 7.18%).

If you convert and use the after-tax funds to pay the $22,000 in taxes due then the Roth will grow from $100,000 to $200,000 of after-tax value.

OTOH, if you do nothing, the $100,000 tIRA will double to $200,000 and when withdrawn $44,000 of taxes will be due which will be paid with the after-tax money and growth... so the end result is $200,000 of after-tax value... so at first blush it appears to be a wash.

But... the taxable account will really not grow to $44,000 because the 7.18% growth is taxed at 22% each year rather than tax-deferred like in the tIRA or tax-free in the Roth. In reality, the $22,000 taxable account will only grow to $37,931* rather than $44,000, so after withdrawing $200,000 from the tIRA and paying $44,000 in taxes, the after-tax value will be only $193,931 rather than $200,000... so there is a $6,069 benefit of converting earlier rather than later even if taxes are paid from taxable .

Thanks for your comments. I'd like to reply to your example.

If I understand it correctly, I think the $6,069 benefit is the result of paying conversion taxes with taxable funds and not necessarily due to converting early. The commutative property assumes everything else is equal. In this case, you've introduced a new variable (i.e., paying conversion taxes from a taxable account). Therefore, I think your example is another way of showing the benefits of paying conversion taxes from a taxable account.

However, you can't derive the benefit without converting early. So, it's difficult to separate the two factors. That's what makes conversions so darn complicated. Please correct me if you disagree. I think this is an important point.

 

[kaneohe]

Excellent! A couple thoughts:

...........................................

For example, let's say you have $100,000 in a tIRA, $22,000 in a taxable account, a 22% marginal tax rate and expect that your investments will double in 10 years (earn 7.18%).

If you convert and use the after-tax funds to pay the $22,000 in taxes due then the Roth will grow from $100,000 to $200,000 of after-tax value.

OTOH, if you do nothing, the $100,000 tIRA will double to $200,000 and when withdrawn $44,000 of taxes will be due which will be paid with the after-tax money and growth... so the end result is $200,000 of after-tax value... so at first blush it appears to be a wash.

But... the taxable account will really not grow to $44,000 because the 7.18% growth is taxed at 22% each year rather than tax-deferred like in the tIRA or tax-free in the Roth. In reality, the $22,000 taxable account will only grow to $37,931* rather than $44,000, so after withdrawing $200,000 from the tIRA and paying $44,000 in taxes, the after-tax value will be only $193,931 rather than $200,000... so there is a $6,069 benefit of converting earlier rather than later even if taxes are paid from taxable account money.

* $37,913 = $22,000*(1+7.18%*(1-22%))^10

......................... .

I've always been too lazy to calculate the tax drag on the taxable account so I just call it 44K- .........the minus sign reminds me it is less than the 44K. If I were more ambitious tho, I would assume that it's a taxable account in equities w/ 2% qualified dividend yld with the balance of 5.18% LTCG but being taxable only after the 10 yr period. Since the tax rates are preferential QDIV/LTCG rates and the CG is paid only after the period ends, that should boost the final return some.......tho still less than 44K.

 

[pb4uski]

Thanks for your comments. I'd like to reply to your example.

If I understand it correctly, I think the $6,069 benefit is the result of paying conversion taxes with taxable funds and not necessarily due to converting early. The commutative property assumes everything else is equal. In this case, you've introduced a new variable (i.e., paying conversion taxes from a taxable account). Therefore, I think your example is another way of showing the benefits of paying conversion taxes from a taxable account.

However, you can't derive the benefit without converting early. So, it's difficult to separate the two factors. That's what makes conversions so darn complicated. Please correct me if you disagree. I think this is an important point.

I agree and you're right in that it is hard to explain easily. I think of the $6,069 benefit as BOTH the result of paying conversion taxes with taxable funds AND converting earlier... effectively you're converting taxable money to tax-free money for the taxable funds (the whole reason why paying taxes with taxable account money is recommended) and you can't get that benefit of tax-free growth without using taxable funds... and you also can't get the benefit until you convert.

Flipping the example to one with only the $100,000 tIRA and no taxable funds is easier and the communicative property applies... convert and pay taxes and you have $78,000 in the Roth that grows to $156,000... don't convert and the $100,000 tIRA doubles to $200,000 but onlhas an after-tax value of $156,000 [$200,000 * (1-22%)].

 

[pb4uski]

I've always been too lazy to calculate the tax drag on the taxable account so I just call it 44K- .........the minus sign reminds me it is less than the 44K. If I were more ambitious tho, I would assume that it's a taxable account in equities w/ 2% qualified dividend yld with the balance of 5.18% LTCG but being taxable only after the 10 yr period. Since the tax rates are preferential QDIV/LTCG rates and the CG is paid only after the period ends, that should boost the final return some.......tho still less than 44K.

The tax drag is actually a fairly easy calculation of the future value with and without income taxes:

[X*(1+i)^n] - [X*(1+i*(1-t))^n] =
[22,000*(1+7.177%)^10] - [22,000*(1+7.177%*(1-22%))^10] =

$6,069

 

[kaneohe]

The tax drag is actually a fairly easy calculation of the future value with and without income taxes:

[X*(1+i)^n] - [X*(1+i*(1-t))^n] =
[22,000*(1+7.177%)^10] - [22,000*(1+7.177%*(1-22%))^10] =

$6,069

won't the numbers be somewhat different w/ the QD/LTCG rate of 15% instead of 22% and basis will increase a bit due to div reinvestment. CG won't be taxed for 10yrs (assuming efficient fund)

 

[pb4uski]

Yes, to figure that tax drag you'd have to model it out.

 

[walkinwood]

Another well written piece. Thank you.

 

[flyingaway]

I saw in many places saying that you should pay the Roth conversion tax from money in a taxable account. Do they consider the fact that the money in the taxable account was taxed before? For a fair comparison, should that tax be added back to the money in the taxable account?

 

[RunningBum]

I saw in many places saying that you should pay the Roth conversion tax from money in a taxable account. Do they consider the fact that the money in the taxable account was taxed before? For a fair comparison, should that tax be added back to the money in the taxable account?

Well, not on the tax that came before, because that was paid no matter how you use the money--for paying taxes on Roth conversion, buying a car, splurging on new dryer sheets, or whatever.

In fact you may not have even paid taxes on it before. What if it was inherited money?

Now, if you have to sell appreciated securities to pay the conversion tax, you are going to owe an additional tax. That's a case by case situation. Remember you only pay the tax on the gain.

 

[AtlasShrugged]

Well, not on the tax that came before, because that was paid no matter how you use the money--for paying taxes on Roth conversion, buying a car, splurging on new dryer sheets, or whatever.

In fact you may not have even paid taxes on it before. What if it was inherited money?

Now, if you have to sell appreciated securities to pay the conversion tax, you are going to owe an additional tax. That's a case by case situation. Remember you only pay the tax on the gain.

I agree with this answer. I have nothing to add.

 

[Time2]

Thinking out loud, back of the envelope calculations for Roth conversions.

Wife and husband have $1M
in tax deferred accounts. In 11 years both will be 72 and the $1M could easily be $2M. First year RMD is $72k, SS 45k, Div and interest is $8500, for a total income of

$125,000. These are conservative numbers.

The question, In 11 years, will $125k push us into a bracket higher than 12%?
I know that is an unanswerable question, but that's one thing that adds to the decision. The top of the 12% tax bracket $80k inflated at 3% for 11 years is $111k.
That says it is a good idea, add the fact that one spouse could die and it a no brainer.
I'm in this boat, I'm pretty sure I want to convert, because I can convert $73k and stay in the 12% bracket. Pay my taxes for 2020 and have liberated money for all my expenses for 2021.

Care to discuss it?

 

[AtlasShrugged]

Thinking out loud, back of the envelope calculations for Roth conversions.

Wife and husband have $1M
in tax deferred accounts. In 11 years both will be 72 and the $1M could easily be $2M. First year RMD is $72k, SS 45k, Div and interest is $8500, for a total income of

$125,000. These are conservative numbers.

The question, In 11 years, will $125k push us into a bracket higher than 12%?
I know that is an unanswerable question, but that's one thing that adds to the decision. The top of the 12% tax bracket $80k inflated at 3% for 11 years is $111k.
That says it is a good idea, add the fact that one spouse could die and it a no brainer.
I'm in this boat, I'm pretty sure I want to convert, because I can convert $73k and stay in the 12% bracket. Pay my taxes for 2020 and have liberated money for all my expenses for 2021.

Care to discuss it?

Time2,

I would rather not give specific individual advice.

I will say that Roth conversion decisions do not lend themselves to back of the envelope calculations. I would encourage you to take the time to do a marginal tax rate analysis on current conversions and future RMDs. There's too much money at stake to skip this important step.

Good luck

 

[Out-to-Lunch]

Have you also inflated your SS number?

 

[Time2]

Time2,

I would rather not give specific individual advice.

I will say that Roth conversion decisions do not lend themselves to back of the envelope calculations. I would encourage you to take the time to do a marginal tax rate analysis on current conversions and future RMDs. There's too much money at stake to skip this important step.

Good luck

The invitation was to the group not specifically you.
The analysis of coversions now is fairly easy for us, we can stay in the 12% bracket. The analysis of of the future is the hard part. For us it is 1 years until RMDs, a lot can change. Here are some of the decisions.
What rate will you money grow in the future years, both tax deferred and taxable accounts? I plan to live out of my taxable account so I need to pick a growth rate for them and then each year subtract my living expenses, and repeat that every year. When 72 rolls around I will have dividends and interest coming out of that taxable account to add to our SS, RMDs. (Will Div and Int stay at about 1.8%? VTI)
How fast will our SS checks grow? Will RMDs have the same required Withdrawal Rates?
I'm posting for debate, but also so I will start creating a framework for an Excel spreadsheet.

 

[SecondCor521]

@Timo2, the only way I know to do it is to model it out in a spreadsheet.

When building my spreadsheets, I try to use the "most likely" assumption so I get sort of the middle-of-the-road picture, not one that is too much on the optimistic side or on the pessimistic side.

To do this, I try to use long term historical averages from authoritative sources, rather than people's predictions.

I also do everything in nominal dollars. Since things inflate and change at various rates, that seems to me to be the only way to do it. (For example, your portfolio rate of increase, SS rate of increase, and tax bracket rate of increase are all likely different.)

I personally also assume status quo indefinitely. This may not be reasonable to some, but I think trying to include potential future changes quickly makes the analysis too complex, and complex models, in my experience, are more likely to be both unwieldy to maintain and more likely to be wrong due to the greater chance for errors in the calculations.

Of course, if you can set up your spreadsheet so that the assumptions are inputs to the model, then that makes it easier to do sensitivity analyses and what if's. Sometimes doing this can also increase complexity (for example, if you had IRS filing status as an input, that would make the bracket calculations more complex), so there's a tradeoff.

Finally, for most analyses that I have done, either there is a clear and significant difference to one path or another, in which case the modeling should clearly point this out and it is worth it to do so. For me, trying to minimize my age 72 tax torpedo seems to be one such decision. Or it really doesn't matter much, in which case the modeling is more likely to be ambiguous but the nice thing is that it doesn't matter what you do so do whatever you want (within reason). For me, when I take SS appears to be one of these decisions.

As a rule of thumb, I think most Roth conversion folks would look at a convert-to-the-top-of-12% scenario for a MFJ couple in their early 60s to be a no brainer for the reasons you've already mentioned.

 

[pb4uski]

What I would do, is using either the TurboTax What-If Worksheet or the dinkytown tax calculator is to sketch out 2020 tax assuming $0 Roth conversions. Then add Roth conversions to the top of the 12% bracket and compare the increase in tax to the amount of the conversion to get the effective rate on the conversion.

As noted, the harder part is to get an idea of your future tax rate before and after RMDs, but since SS and tax brackets grow with inflation and RMDs grow at at least the rate of inflation due to investment growth, I think you could get a rough approximation by using today's SS at the age you plan to take it (excluding COLA growth) and RMDs assuming you were now 72 (or maybe a little more).

 

[CardsFan]

What I would do, is using either the TurboTax What-If Worksheet or the dinkytown tax calculator is to sketch out 2020 tax assuming $0 Roth conversions. Then add Roth conversions to the top of the 12% bracket and compare the increase in tax to the amount of the conversion to get the effective rate on the conversion.

As noted, the harder part is to get an idea of your future tax rate before and after RMDs, but since SS and tax brackets grow with inflation and RMDs grow at at least the rate of inflation due to investment growth, I think you could get a rough approximation by using today's SS at the age you plan to take it (excluding COLA growth) and RMDs assuming you were now 72 (or maybe a little more).

I have done the same the last few years, staying in the 12% bracket. This year, I decided to convert up to the IRMAA threshold (we are 65 and on Medicare).

The marginal rate on the additional conversion was about 25%, but the rate for the entire conversion was 16%. Since I avoided taxes at the 28% rate to make the contributions, this is a win (but paying an additional $10k in taxes is still tough to swallow).

 

[JustVisitingThisPlanet]

On the suggestion to convert at end of year vs. early in year... End would be ideal unless you have an opportunity early in year you don't want to pass up, especially for those in private equity via SD-IRA and can't easily sell a holding for a new opportunity. I had one year where I did convert early in year, which paid off.

 

[schellem]

I'm in the conversion calculation dilemma, I just retired in June at 58, DW will follow in a year or so. We will have a lot of time to do conversions with little income. I've started on spreadsheet but just keeps getting more and more complex.

The latest issue to add to the spreadsheet is to figuring out the road between shooting for low nearly tax free income and getting ACA subsidies versus trying to take advantage of the 24% bracket while we have it. I think we can get subsidies of around 8-10K.

Anyone have any tips in relation to considering ACA subsidies?

 

[MarieIG]

I am going to print out this article and the thread.

Pb4uski had explained to me (on an old thread - couldn't update it) how to do a backdoor Roth; which I did. I will be rolling over my taxable 401k (costly administrative fees) into an IRA on January 2nd; and want to be prepared for Roth conversions.

 

[SecondCor521]

^ I don't have the math for you, and haven't done it myself, but it seems that most people here think that ACA subsidies are worth more than the Roth conversion tax arbitrage savings.

I think some of that opinion is probably based on the "bird in the hand" sentiment.

It's harder for married folks I think, as you have the potential for the surviving spouse to be socked with higher income tax brackets and IRMAA adjustments after the first spouse dies.