Those Who don't/didn't do Roth Conversions

[youbet]

I had only recently considered 'breakeven' for Roth conversions, but it is something to be aware of.

I did a spreadsheet, assuming 12% conversions of $50,000 for 5 years prior to RMDs on a starting $1M tIRA. Then 22% on the tIRA RMDs. I came up with a breakeven of ~ age 87.

That's because while you have reduced your tIRA balance, the RMD is only a % of that, so it takes time to get all your money back in savings. At least that's how I see it.

But then, if your heirs are at a tax rate > 12%, they will continue to see the benefit, so now what do we make of it? I'd say not much overall, I'll just stick to the basic math.

It's tempting to go into the 22% bracket for conversions, assuming we don't pass at near the same time, one of us will be pushed into the higher single rates. But that's a tough tax-bill to swallow, and it takes even more assumptions, so I'm just sticking with the almost certain 12/22 arbitrage.

And if we both pass before that break-even, who cares - it's not like we depleted our savings with that tax bill, so no real impact on quality of life. Heirs benefit, so it's all OK.

-ERD50

What are you using to account for the opportunity cost of giving Uncle money early? Say you pay $10k/yr extra taxes for 10 years while doing conversions. How are you accounting for the cost of that money being out of your control and not earning for you?

 

[pb4uski]

What are you using to account for the opportunity cost of giving Uncle money early? Say you pay $10k/yr extra taxes for 10 years while doing conversions. How are you accounting for the cost of that money being out of your control and not earning for you?

See post #231 in this thread. Calculations include the time value of money.

I think y'all are making the analysis more confusing by including the $900k that stays in the IRA... since that $900k is common to both alternatives it doesn't matter and can be excluded from each alternative.

To simplify, say one has $100,000 in an IRA and $12,000 in taxable funds, is currently in the 12% tax bracket and expects to be in the 22% bracket in 30 years. 5% growth.

If you convert then you end up with $100,000 in a Roth and $0 in taxable. The Roth grows at 5% for 30 years to $432,194 ($100,000 * (1+5%)^30) that can be spent tax-free.

If you don't convert then the $100,000 in the IRA grows to $432,194 in 30 years and when withdrawn you end up with $337,111 (($100,000 * (1+5%)^30)*(1-22%)). Meanwhile, the $12,000 in taxable grows to $43,671 ($12,000*(1+5%*(1-12%))^30). So in total at the end of 30 years you have $380,782 that has been taxed and can be spent.

The Roth conversion is $51,412 better ($432,194 - $380,782).

Most of the benefit is from the tax savings... with the IRA you pay $95,083 in tax when you withdraw the $432,194... whereas the $12,000 paid in taxes with the Roth conversion is valued at $51,863 after 30 years ($12,000*(1+5%)^30) for a difference of $43,220.

The other $8,191 is tax savings by shifting $12,000 in taxable to the tax-free Roth for 30 years. The $12,000 in the Roth grows to $51,863 ($12,000*(1+5%)^30), $8,191 more than the growth of the $12,000 in the taxable account to $43,671 ($12,000*(1+5%*(1-12%))^30).

 

[SevenUp]

What are you using to account for the opportunity cost of giving Uncle money early? Say you pay $10k/yr extra taxes for 10 years while doing conversions. How are you accounting for the cost of that money being out of your control and not earning for you?

While the money is in the traditional account, it's part yours and part Uncle's.

When you convert, you keep the part that's all yours. Best to do that when you get to keep the largest part - i.e., when the tax rate is lowest.

The commutative property of multiplication governs many conversion results.

 

[youbet]

See post #231 in this thread. Calculations include the time value of money.

Thanks.

So, sifting through all that, the answer to my question is 5%?

And ERD50, I'm specifically interested in the number you're using.

 

[youbet]

While the money is in the traditional account, it's part yours and part Uncle's.

When you convert, you keep the part that's all yours. Best to do that when you get to keep the largest part - i.e., when the tax rate is lowest.

The commutative property of multiplication governs many conversion results.

Yeah.......

But, specifically, what are you personally using for "r" as defined in the link you posted?

 

[SevenUp]

While the money is in the traditional account, it's part yours and part Uncle's.

When you convert, you keep the part that's all yours. Best to do that when you get to keep the largest part - i.e., when the tax rate is lowest.

The commutative property of multiplication governs many conversion results.

Yeah.......

But, specifically, what are you personally using for "r" as defined in the link you posted?

3%? 5%? 7%? Depends what I had for breakfast on any given day. But, as that wiki section describes, for the choice in question it doesn't matter - agreed?

 

[pb4uski]

Yeah, even 0% will be beneficial if the conversion tax rate is less than the RMD tax rate... and 10% would work too. It's all about those two tax rates whether it is 0% or 10% or 5 years or 50 years.

 

[m3bs]

I've run multiple scenarios, and in my current situation I've concluded that it really won't matter......if tax rates stay the same and brackets move with inflation. RMDs will push me solidly into the 22% range in a few years either way. I'm inclined to do some significant Roth conversions simply as a hedge against future tax rate increases which I believe are likely. Even if conversion bumps me into the 24% range, that may be better than the rates necessary to address budget shortfalls. I'll definitely get hit with additional IRMAA surcharges, but I need to be careful to avoid the 3.8% NIIT.

While increasing taxes may be political suicide for many elected officials, I think it has to happen for a variety of reasons. Even if congress does nothing (the likely, easy way out), the Trump tax cuts will be expiring so rates will increase in 2026.

 

[PERSonalTime]

Thanks for starting this thread. It was useful as people gave reasons why and why not although you asked for the why nots. It brought some clarity to my decision. We have done them to control ACA subsidies and stay out of Medicaid. I was not sure, if we should continue until I put together a spread sheet with modest gains and realized we will move from 12% into 22% (25%) not too long after I have to start RMDs. So, it’s a given for us for now unless something changes. We can probably delay that higher tax bracket for a number of years depending on the conversions. I think to know why not you have to have the why as well. Thanks again for starting this very informative discussion.

It's great when we can all learn from each other. [emoji846]

 

[Rianne]

Question about tIRA balances when RMDs start. If you're laddered in CDs and treasuries, what does the IRS consider the balance of your portfolio? For instance, coupon interest might drop in the settlement account twice a year. Do they consider the value of your tIRA at the end of the year based on EOY balances? If you have $100k locked in a CD for 10 years, is that considered as $100k or $100k + interest accrued that year?

Pb4uski - you're an accountant and are 100% in CDs, corporates, and treasuries, right?

 

[target2019]

RMD's are based on prior year ending balance.

 

[pb4uski]

Question about tIRA balances when RMDs start. If you're laddered in CDs and treasuries, what does the IRS consider the balance of your portfolio? For instance, coupon interest might drop in the settlement account twice a year. Do they consider the value of your tIRA at the end of the year based on EOY balances? If you have $100k locked in a CD for 10 years, is that considered as $100k or $100k + interest accrued that year?

Pb4uski - you're an accountant and are 100% in CDs, corporates, and treasuries, right?

+1 with target2019, RMDs are based on the fair value of your account as of 12/31/py including settlement funds.

On the second part, yes, I am an accountant. But not 100% in CDs, corporates, and treasuries... 87.3% in CDs, agency and corporate bonds right now... no Treasuries at this time other than 5.8% in I-Bonds that I am in redemption mode (I look at Treasuries but they haven't been attractive to me)... and 4.2% in high quality preferreds to add a little spice (Allstate, Citigroup, MetLife and on the lookout for others) and 2.7% in MMF dry powder.

 

[ERD50]

See post #231 in this thread. Calculations include the time value of money.

The other $8,191 is tax savings by shifting $12,000 in taxable to the tax-free Roth for 30 years. The $12,000 in the Roth grows to $51,863 ($12,000*(1+5%)^30), $8,191 more than the growth of the $12,000 in the taxable account to $43,671 ($12,000*(1+5%*(1-12%))^30).

Hmmmm, am I missing something?

How do you figure the "shifting $12,000 in taxable to the tax-free Roth for 30 years."?

In your example, I have a tIRA of $100K. If I convert to Roth, I pay the 12% tax ($12K) from taxable, taking my taxable to zero. I have $100K in the Roth now, and Uncle Sam has my $12K. Seems I 'shifted' $12K to Uncle Sam, not to my Roth?

This also is complicated a bit as we don't know what the cost basis of the $12K is/was (minor after 30 years though), but if I had to sell stocks to free up $12K in cash flow to pay taxes, it's 12% only on the gain?

I'm reviewing this to be able to answer youbet's Qs.

-ERD50

 

[pb4uski]

Let's assume that the $12k is cash so there is no basis difference or even you sold somehting at a capital gain to generate the $12k that the capital gains tax would be nil if you are in the 12% tax bracket for ordinary income.

If the given was that there was no taxable account and only $100k tIRA, I think we would agree that if a Roth conversion was done that it would end up with $88k in the Roth, right? In that scenario, you would have no choice but to withhold $12k from the Roth conversion.

In that scenario of no taxable funds available to pay the $12k of tax, if there was no difference in rates then one would be indifferent to converting or not. OTOH, with a 10% difference in rates (22% vs 12%) then the $43,220 benefit would be very straight forward... the difference between the Roth balance after 30 years of $380,331 (=88000*(1+5%)^30) and the net proceeds from the tIRA in 30 years of $337,111...$432,194 tIRA balance (=100000*(1+5%)^30) less 22% tax of $95,083.

If there is $12k of taxable account money to pay the tax, then the Roth grows to $432,194 if converted and taxes paid from taxable account. If no conversion, then the tIRA also grows to $432,194, and $95,083 of taxes are due when the $432,194 is withdrawn and the taxable account grows to $43,671; for a net of $380,782 and a benefit of $51,412.

So with the $12k taxable funds the benefit is $51,412 and without it is $43,220... so an additional $8,192 benefit because one could pay the taxes out of taxable funds.

What causes this $8,192 additional benefit? It is because the $12k ends up in the Roth and the earnings on that $12k are tax-free for 30 years.

Looking at the value of the $12k after 30 years in isolation:

With conversion: $51,863 = 12000*(1+5%)^30
No conversion: $43,671 = 12000*(1+5%*(1-12%))^30
Taxes avoided as a result of conversions: $8,192

If the $12k didn't effectively end up in the Roth then there would be no additional benefit.
Get it now?

 

[dknighd]

It much simpler than that.
If I take out money today and pay X percent tax, or wait a couple months, then pay Y percent.

If X is less then Y, then I ill take my money now. If X=Y it does not matter,


All we have to do now is predict the future

 

[Koolau]

Let's assume that the $12k is cash so there is no basis difference or even you sold somehting at a capital gain to generate the $12k that the capital gains tax would be nil if you are in the 12% tax bracket for ordinary income.

If the given was that there was no taxable account and only $100k tIRA, I think we would agree that if a Roth conversion was done that it would end up with $88k in the Roth, right? In that scenario, you would have no choice but to withhold $12k from the Roth conversion.

In that scenario of no taxable funds available to pay the $12k of tax, if there was no difference in rates then one would be indifferent to converting or not. OTOH, with a 10% difference in rates (22% vs 12%) then the $43,220 benefit would be very straight forward... the difference between the Roth balance after 30 years of $380,331 (=88000*(1+5%)^30) and the net proceeds from the tIRA in 30 years of $337,111...$432,194 tIRA balance (=100000*(1+5%)^30) less 22% tax of $95,083.

If there is $12k of taxable account money to pay the tax, then the Roth grows to $432,194 if converted and taxes paid from taxable account. If no conversion, then the tIRA also grows to $432,194, and $95,083 of taxes are due when the $432,194 is withdrawn and the taxable account grows to $43,671; for a net of $380,782 and a benefit of $51,412.

So with the $12k taxable funds the benefit is $51,412 and without it is $43,220... so an additional $8,192 benefit because one could pay the taxes out of taxable funds.

What causes this $8,192 additional benefit? It is because the $12k ends up in the Roth and the earnings on that $12k are tax-free for 30 years.

Looking at the value of the $12k after 30 years in isolation:

With conversion: $51,863 = 12000*(1+5%)^30
No conversion: $43,671 = 12000*(1+5%*(1-12%))^30
Taxes avoided as a result of conversions: $8,192

If the $12k didn't effectively end up in the Roth then there would be no additional benefit.
Get it now?

Yeah, you state it much more eloquently (elegantly) than I ever have. Thanks from all of us.

 

[all4j]

What else am I missing?

Be sure to consider the difference your personal situation brings & difference in tax laws between taxable & tax deferred. For example, consider your asset allocation. What if you have a $1m taxable & a $1m tira with a 60/40 split….& implemented as tira with $800k bonds & $200k stocks. Taxable account is all stock. Do you really expect to get the same return between the ira & the taxable? If you have both as $600k/$400k, the returns will be similar but taxed differently. You’ll also be able to do a few things such as loss harvesting in taxable that you can’t in iras.

 

[LateToFIRE]

Let's assume that the $12k is cash so there is no basis difference or even you sold somehting at a capital gain to generate the $12k that the capital gains tax would be nil if you are in the 12% tax bracket for ordinary income.

If the given was that there was no taxable account and only $100k tIRA, I think we would agree that if a Roth conversion was done that it would end up with $88k in the Roth, right? In that scenario, you would have no choice but to withhold $12k from the Roth conversion.

In that scenario of no taxable funds available to pay the $12k of tax, if there was no difference in rates then one would be indifferent to converting or not. OTOH, with a 10% difference in rates (22% vs 12%) then the $43,220 benefit would be very straight forward... the difference between the Roth balance after 30 years of $380,331 (=88000*(1+5%)^30) and the net proceeds from the tIRA in 30 years of $337,111...$432,194 tIRA balance (=100000*(1+5%)^30) less 22% tax of $95,083.

If there is $12k of taxable account money to pay the tax, then the Roth grows to $432,194 if converted and taxes paid from taxable account. If no conversion, then the tIRA also grows to $432,194, and $95,083 of taxes are due when the $432,194 is withdrawn and the taxable account grows to $43,671; for a net of $380,782 and a benefit of $51,412.

So with the $12k taxable funds the benefit is $51,412 and without it is $43,220... so an additional $8,192 benefit because one could pay the taxes out of taxable funds.

What causes this $8,192 additional benefit? It is because the $12k ends up in the Roth and the earnings on that $12k are tax-free for 30 years.

Looking at the value of the $12k after 30 years in isolation:

With conversion: $51,863 = 12000*(1+5%)^30
No conversion: $43,671 = 12000*(1+5%*(1-12%))^30
Taxes avoided as a result of conversions: $8,192

If the $12k didn't effectively end up in the Roth then there would be no additional benefit.
Get it now?

I follow your math, but this idea of the $12K ending up in the Roth if you pull it out of taxable to pay the conversion taxes, is just messing with my corp finance guy head. To me there is an opportunity cost even if the $12K were coming from cash - because it could have been put to some other productive use and now you've lost the use of it. But, I can agree that the differences in how we might think about this specific aspect of the analysis do not change the fundamentals of the conversion decision. Thanks for laying it out for us.

 

[LateToFIRE]

Be sure to consider the difference your personal situation brings & difference in tax laws between taxable & tax deferred. For example, consider your asset allocation. What if you have a $1m taxable & a $1m tira with a 60/40 split….& implemented as tira with $800k bonds & $200k stocks. Taxable account is all stock. Do you really expect to get the same return between the ira & the taxable? If you have both as $600k/$400k, the returns will be similar but taxed differently. You’ll also be able to do a few things such as loss harvesting in taxable that you can’t in iras.

thanks - had not thought about AA between the different types of accounts

 

[all4j]

thanks - had not thought about AA between the different types of accounts

May seem like a small thing, but just as an illustration, apply it to immediately preceding post where you are looking at the one specific example. The 'no conversion' scenario for the $12k was calculated at $43671 (I didn't check, but assume it matches explanation). That was based on the taxable account having the same return as the ira, but also being all taxed at ordinary income rate. IF that was allocated to stock, there would be no tax per the assumptions. (Well, yeah there could be some stcg, but...).

 

[pb4uski]

May seem like a small thing, but just as an illustration, apply it to immediately preceding post where you are looking at the one specific example. The 'no conversion' scenario for the $12k was calculated at $43671 (I didn't check, but assume it matches explanation). That was based on the taxable account having the same return as the ira, but also being all taxed at ordinary income rate. IF that was allocated to stock, there would be no tax per the assumptions. (Well, yeah there could be some stcg, but...).

Fair point and in that case the benefit would be entirely driven by the difference between tax on the Roth conversion and tax on the RMD.

 

[Z3Dreamer]

Another retired CPA. Have NOT done Roths because I have been doing the ACA game. Quite successfully. Have done the Roth conversion math way too many times. Getting to the top of the 12% bracket is the goal. Tax torpedo (social security taxable up to 85%) complicates things. Getting into the next bracket in case I or DW go soon is not a strong argument for us. This will be my first year of Roth conversions.

 

[Montecfo]

Same thoughts as RunningBum's.



Roth conversions are significantly different from the usual "payback time" scenario. E.g., a company builds a new factory and earns money on the factory output. In that case Payback Period is the time it takes the company to recoup the money it paid for all the materials and labor to build the factory.



With a Roth conversion, it's just a question of the "access cost" to get at the money in the traditional account. If you can access the money by paying 12%, that's better than having to pay 22% to access the money, regardless of how many years elapse between having the 12% option vs. being forced to pay 22%.

Payback or breakeven analysis is in fact valid for considering Roth conversions. Some people just wonder why they should volunteer to pay taxes early that they may never otherwise owe in their lifetime.

But of course like the factory build, it is not the only tool to analyze your return.

 

[pb4uski]

I guess that I look at it too simply. The funds that I have converted over the last 10 years were converted at about 10% on average ( a mix of 0%, 10% and 12%).

If I live long, I'll be paying a mix of 12% and 22% on RMDs, probably averaging about 18%.

If I die early and DW lives long, she will be paying more than that because she'll be filing as a single unless she remarries.

If we both die early, then our kids will be paying the tax. One kid has good income and is already in the 22% tax bracket and the other kid is already in the 12% tax bracket.

So for example, on my 2023 conversion I'll pay about 11.5%. Given all of the above, how can we (and our heirs) possibly not benefit?

It seems to me that no matter what the benefit will be positive, it is just a matter of degree. In some cases the benefit will be small and in others it might be substantial, only time will tell. It seems like a chance worth taking and I don't need no stinking payback analysis.

 

[SevenUp]

Same thoughts as RunningBum's.

Roth conversions are significantly different from the usual "payback time" scenario. E.g., a company builds a new factory and earns money on the factory output. In that case Payback Period is the time it takes the company to recoup the money it paid for all the materials and labor to build the factory.

With a Roth conversion, it's just a question of the "access cost" to get at the money in the traditional account. If you can access the money by paying 12%, that's better than having to pay 22% to access the money, regardless of how many years elapse between having the 12% option vs. being forced to pay 22%.

Payback or breakeven analysis is in fact valid for considering Roth conversions. Some people just wonder why they should volunteer to pay taxes early that they may never otherwise owe in their lifetime.

Agreed that people who expect to pay a lower tax rate in the future shouldn't be doing Roth conversions now.

What do you consider the payback period to be for someone who can convert this year at 12% and expects next and all subsequent years to have at least a 22% marginal tax rate?