Why ROTH conversion is a bad idea for me.

[pb4uski]

It is a. But unless one has a crystal ball that is the best that you can do and I agree with many here that in the future it is more likely that tax rates will be higher than lower, so using todays 2017 Tax Act rates is conservative. And add to that that if DW or I die before RMDs kick in that the remaining spouse will be in a much higher single tax bracket unless we remarry (won't be me!).

The arbitrage is the marginal tax rate on a Roth conversion done today... increase in tax divided by the conversion amount. For us for 2021 that will be ~11.5%.

The comparison is to the projected marginal tax on an RMD... in our case in 2027... I'm estimating 15.3% (based on current tax rates, knowing that tax rates may well be higher).

While the difference in 2021 is a modest 3.8%, on $80k of conversion that is $3k.... well worth a couple of hours work... plus the benefit is likely understated since tax rates are scheduled to increase between now and 2027.

 

[SecondCor521]

As a first order answer, I do Roth conversions for tax arbitrage. But I do that arbitrage because I think it will result in more spending money for me during my lifetime. Generally speaking, I think it will also be beneficial to my heirs, but that depends on their tax rates in the future, which is hard to suss out.

As far as the question in post #100, one has to make assumptions about the future. I assume certain growth and inflation rates for investments and inflation-adjusted tax items. Generally speaking, I assume the status quo with regards to RMD ages and divisors, as well as tax brackets.

I then look at what I predict my age 72 marginal tax rate will be. I convert each year as long as I can remain below that tax rate. I'm agnostic as to the amount, but in general I'll likely end up converting modestly while my kids are in FAFSA universities (another two years or so), then up to the 400% FPL level until Medicare age (depending on what happens with the ACA cliff), then up to the top of a certain bracket after that. (I know what that certain bracket is now, but it'll likely change by the time I get there anyway for various reasons.)

 

[Gumby]

I consider it a win if I take money out of the tIRA/401k at a marginal tax rate less than when I put it in. There is no guess work about that. A great amount of it went in at 35% + and none of it went in at less than 24%. It is coming out at 22%.

 

[RetireBy90]

Let’s just be clear, there may be several reasons to do Roth conversions depending on your situation. For me, they include, not in order of impact:

- IRMAA limits, if I don’t convert today RMDs will increase my IRMAA.
- higher income with RMDs will disallow any loss on rental property
- of course, my expectation of higher tax rates after 2025 or even sooner
- The hit from single tax rates for the one to survive the longest
- I prefer to pass any remaining funds with taxes already paid
- with Roth I can take the money at any time without a tax impact, or impact of other income limits. If I. Want to buy a beach house, I can take purchase price in one year.
- it gets the tax bill done for me.

I’m sure others can add to this list. We had a thread where many advantages were discussed.
Again, if you don’t want to do Roth conversions then that is fine with me. However, for the vast majority I think they are a great idea. Many can use them as a tool with some or all of benefits listed.

 

[REWahoo]

I consider it a win if I take money out of the tIRA/401k at a marginal tax rate less than when I put it in.

^ This.

 

[Out-to-Lunch]

Edit to add: Once again, I got caught out by the fact that I did not see the others had provided an answer (because your post was the last one on the previous page!).

It is ALL speculative. Your decision to put funds in a pre-tax account in the first place was speculation that you will, net net, be saving taxes versus when you spend the money.

The way I look at it, pre-tax contributions and Roth conversions allow you to smooth your taxable income over time to minimize taxes. If you are in a high tax bracket while working, (and you speculate that you will be in a lower one later), you can lower your taxable income by saving in a 401(k). If you are in a lower bracket now, you can do Roth conversions to realize more income now to spare it from higher taxation later.

But, yes, speculation is required.

 

[Dtail]

I consider it a win if I take money out of the tIRA/401k at a marginal tax rate less than when I put it in. There is no guess work about that. A great amount of it went in at 35% + and none of it went in at less than 24%. It is coming out at 22%.

Yup.
If monies go in at 35% and are taken out at 22%, but not doing max conversions to bring the 22% lower, not the end of the world.

 

[RunningBum]

Thanks @pb4uski.

I'm trying to separate the variables. They seem to be: estate/heir related, and tax arbitrage related. Several posters have commented that tax arbitrage is the main driver for Roth conversion.

Regarding tax arbitrage:

a. is the arbitrage today's rates vs legislatively changed future rates after RMD age, or

b. is the arbitrage a non-RMD rate vs a (higher) RMD rate, regardless if or how tax rates change in the future?

If the choice is a., I see a speculative nature to the Roth conversion decision. The speculation is whether and how future income tax rates will change.

If the choice is b., this can be sleuthed out with a calculator. So far, i-orp and my own spreadsheet have not sleuthed the "Roth conversion is good" indicator for me.

Mostly (b) for me.

Right now my Roth conversions are costing 10% Fed Tax and almost 10% loss of subsidy, so about 20% total. The subsidy loss could drop to 8.5% in 2021, if the senate passes the bill that the house has already done. I have in the past paid as high at 25% on some conversions.

In my projections for the future, I would not quite be out of the SS tax hump so the RMD cost of the first $1-2K is 49.95%. After that, I'm still pushing QDivs into being taxed so it's 27%. If I'm being too conservative on my income and I am out of the SS tax hump, then it's 27% until all of my QDivs are taxed, and then it settles down to 22%.

Clearly, this favors conversion for my case. I wonder how many other people have really tried to figure out their true marginal tax rates when they'll have to take RMDs.

If I consider (a), future tax rates returning to previous levels, I'm at 55.5%, 30% and 25% for the income levels above.

You call (a) speculative, but it's what tax rates WILL BE in 2026 unless Congress takes action to extend the tax cuts. In my opinion, and it is just an opinion, it's more speculative to assume that congress will extend the tax cuts.

My marginal state tax rate has been 5.75% all the way so there's no tax arbitrage there.

 

[RunningBum]

There is a further advantage to Roth conversions if you can pay the tax out of your taxable account rather than out of the conversions. Keep the amount of taxes you paid in the Roth by paying out of taxable allows that amount to grow tax free. Even if conversion tax rates = future RMD tax rates, the tax free growth on that portion puts you ahead.

 

[Cobra9777]

I'm mostly (b) also. We deferred this income at 28% to 33%. Without conversions, we'd pay mostly 25% at RMD time, possibly some at 28% depending on how the IRAs grow. So I'm near-giddy to settle-up today at a mix of 12% and 22%.

I do expect rates to increase in 2026, per current law. I agree that it seems more speculative to assume otherwise. Beyond that, I'd say it's more likely that rates go up than down. But that plays a very minor role in my thinking.

I do consider the surviving spouse at single rates. And the advantage of using taxable funds to pay the tax.

I am not motivated in the least by pre-paying inheritance tax for the kids. Though I concede that it might actually play out that way.

 

[daver]

There is a further advantage to Roth conversions if you can pay the tax out of your taxable account rather than out of the conversions. Keep the amount of taxes you paid in the Roth by paying out of taxable allows that amount to grow tax free. Even if conversion tax rates = future RMD tax rates, the tax free growth on that portion puts you ahead.

That's my dilemma, not enough in taxable account to cover conversion taxes through RMD age..

 

[chassis]

Thanks for the comments, this has been a great thread.

What do posters on this thread think the total lifetime benefit is by doing Roth conversions? It's a question similar to deciding when to start receiving Social Security.

What is your calculated breakeven age for Roth vs no-Roth strategy?

What is your calculated dollar, or percent of net worth, benefit for the Roth vs no-Roth strategy? This can be a NPV in today's dollars, or a future projected net worth number at an age of your choosing.

I'm not asking people to share their personal details, but rather to quantify the magnitude of impact of a Roth vs. no-Roth decision. Thanks again.

 

[Out-to-Lunch]

Thanks for the comments, this has been a great thread.

What do posters on this thread think the total lifetime benefit is by doing Roth conversions? It's a question similar to deciding when to start receiving Social Security.

What is your calculated breakeven age for Roth vs no-Roth strategy?

What is your calculated dollar, or percent of net worth, benefit for the Roth vs no-Roth strategy? This can be a NPV in today's dollars, or a future projected net worth number at an age of your choosing.

I'm not asking people to share their personal details, but rather to quantify the magnitude of impact of a Roth vs. no-Roth decision. Thanks again.

I think this question is ill-posed.

Most of the time that I have seen the matter posed as a "break-even" decision, it is framed as "I need XX years of tax-free growth to make up for the taxes." But this is the wrong way to look at it. It really is about the tax rate when the money is withdrawn (either now, by conversion, or later, for spending).

To simplify, consider $1,000 in a tIRA. Further, let's assume that you will be in the same tax bracket (say, 22%) both now and later, and that you will pay for taxes out of the tIRA funds. (More on that last assumption later.)

In scenario A, you convert it now, and have $780 in your Roth. Let's say over the next 10 years, the money doubles, there are no further taxes, and you therefore have $1560 in spendable funds.

In scenario B, you wait 10 years, and your tIRA money doubles to $2,000. You then decide to spend that, and you take it out and pay your $440 in taxes, which leaves you with (drum roll...) $1560 to spend.

Mathemeticians call this "the commutative law of multiplication." There is no break-even time period.

Obviously, if the tax rate is higher later, you are better off if you convert now; and, contrariwise, if your tax rate is lower later, you would be better off waiting.

Things can get a little more complicated if you are paying for taxes from other funds, like a taxable account. This favors converting now, because you effectively move some of your money in a non-tax-preferred savings vehicle into a tax-preferred account. You can read more about that here: https://personal.vanguard.com/pdf/ISGBETR.pdf

 

[chassis]

I think this question is ill-posed.

Most of the time that I have seen the matter posed as a "break-even" decision, it is framed as "I need XX years of tax-free growth to make up for the taxes." But this is the wrong way to look at it. It really is about the tax rate when the money is withdrawn (either now, by conversion, or later, for spending).

To simplify, consider $1,000 in a tIRA. Further, let's assume that you will be in the same tax bracket (say, 22%) both now and later, and that you will pay for taxes out of the tIRA funds. (More on that last assumption later.)

In scenario A, you convert it now, and have $780 in your Roth. Let's say over the next 10 years, the money doubles, there are no further taxes, and you therefore have $1560 in spendable funds.

In scenario B, you wait 10 years, and your tIRA money doubles to $2,000. You then decide to spend that, and you take it out and pay your $440 in taxes, which leaves you with (drum roll...) $1560 to spend.

Mathemeticians call this "the commutative law of multiplication." There is no break-even time period.

Obviously, if the tax rate is higher later, you are better off if you convert now; and, contrariwise, if your tax rate is lower later, you would be better off waiting.

Things can get a little more complicated if you are paying for taxes from other funds, like a taxable account. This favors converting now, because you effectively move some of your money in a non-tax-preferred savings vehicle into a tax-preferred account. You can read more about that here: https://personal.vanguard.com/pdf/ISGBETR.pdf

Thanks @Out-to-Lunch

Are you saying the Roth conversion strategy results in less tax dollars being paid, with the same spendable funds? $1560 in this example.

I try to analyze financial decisions as investments. Nearly all businesses have standard forms that calculate NPV and breakeven. In this case, the investment is in accelerated tax payments. I do not understand (yet), the quantified payback value in terms of dollars or time. Where's the beef? What's the rub? What's in it for me? Quantified in terms of dollars or time.

I understand the hypothetical scenario of higher future federal income tax rates. It has been agreed on this thread that an element of speculation exists on the topic of future rates.

I want to become a believer in the Roth strategy, but so far I'm not there. I haven't seen a quantified (dollars or time) benefit for this, and the rationale of tax arbitrage contains speculation that I may choose not to participate in.

In my view, driving to the grocery store contains an element of speculation - that you will arrive in the parking lot safely and that the store will have the products you wish to purchase. Most people will say that is a low and acceptable level of speculation. I invest in equities, so I participate in speculation in that example. I am not saying that a world exists with no speculation.

I'm not sure I'm on board with the speculation of higher future income tax rates, to the extent I would choose to accelerate payments to the government. I remain open to changing my view.

 

[Cobra9777]

...In this case, the investment is in accelerated tax payments. I do not understand (yet), the quantified payback value in terms of dollars or time...

There is no investment in accelerated tax payments. The "breakeven" concept does not apply here. Breakeven implies that you spent some money (temporarily worsening your financial position) with the expectation of recovering it (and much more) over some period of time. In the case of Roth conversions, you are no worse off after paying conversion tax than you were before.

Reason is: your tIRA is *NOT* all yours. From the moment you deferred the income, you created a liability which has grown alongside the part you *DO* own. It's just a question of when, and at what rate, you decide to finally settle that liability.

By converting, you simply retire the liability early, assuming that rates are lower now than if you had waited until the tax torpedo, when SS and RMDs start. I suppose you could say it's break-even from Day 1, but that's not entirely accurate either. In a sense, you are just locking in your gain vs the higher rate when you deferred the income.

Here's an old thread with lots more discussion on this specific point.

 

[Out-to-Lunch]

There is no investment in accelerated tax payments. The "breakeven" concept does not apply here. Breakeven implies that you spent some money (temporarily worsening your financial position) with the expectation of recovering it (and much more) over some period of time. In the case of Roth conversions, you are no worse off after paying conversion tax than you were before.

Reason is: your tIRA is *NOT* all yours. From the moment you deferred the income, you created a liability which has grown alongside the part you *DO* own. It's just a question of when, and at what rate, you decide to finally settle that liability.

By converting, you simply retire the liability early, assuming that rates are lower now than if you had waited until the tax torpedo, when SS and RMDs start. I suppose you could say it's break-even from Day 1, but that's not entirely accurate either. In a sense, you are just locking in your gain vs the higher rate when you deferred the income.

Here's an old thread with lots more discussion on this specific point.

I like this explanation! Nicely said.

 

[Out-to-Lunch]

Let me say just one more thing: When I say "future tax rate," there can be a lot hiding there. Things like MFJ vs. single, IRMAA tiers, NIIT, ACA subsidies, SS tax torpedo, etc. All of these could be (IMHO) lumped into "future tax rate," but obviously they all require a little more thought (and speculation!) than simply the federal and state marginal rates.

 

[pb4uski]

....What do posters on this thread think the total lifetime benefit is by doing Roth conversions? ...

I'm not sure if this helps or not.

I have a projection model from now (age 65) to age 91. My current plan is to do Roth conversions in a way that levelizes my effective tax rate from age 65-71... I'm converting slightly into the 22% tax bracket over the next 6 years... then RMDs begin at age 72.

This plan compared to if I do not do any Roth conversions:

Taxes paid: 91% of the no Roth conversions
Age 91 invested assets: 107% of no Roth conversions
Age 91 invested assets net of deferred income taxes @ 22%: 109% of no Roth conversions

So I think that the benefit of Roth conversions are substantial enough to be worth the effort.... and these numbers presume that tax rates are the same as today and that both of us live until 91.

To me, the precision isn't all that important... it shows that it is beneficial to us with conservative assumptions so that alone makes it worthwhile... the fact that the actual results will likely be better is icing on the cake.

 

[chassis]

I'm not sure if this helps or not.

I have a projection model from now (age 65) to age 91. My current plan is to do Roth conversions in a way that levelizes my effective tax rate from age 65-71... I'm converting slightly into the 22% tax bracket over the next 6 years... then RMDs begin at age 72.

This plan compared to if I do not do any Roth conversions:

Taxes paid: 91% of the no Roth conversions
Age 91 invested assets: 107% of no Roth conversions
Age 91 invested assets net of deferred income taxes @ 22%: 109% of no Roth conversions

So I think that the benefit of Roth conversions are substantial enough to be worth the effort.... and these numbers presume that tax rates are the same as today and that both of us live until 91.

To me, the precision isn't all that important... it shows that it is beneficial to us with conservative assumptions so that alone makes it worthwhile... the fact that the actual results will likely be better is icing on the cake.

Thanks @pb4uski. This is exactly what I am looking for. I'm working on tuning up my model to be able to project/forecast this.

Does your projection model tell you at what age the 107% or 109% are equal to 100%?

 

[pb4uski]

I think it would only be 100% at time zero (age 65) and it would get progressively better if I extended the projection... sort of like compounding interest.

 

[RunningBum]

Thanks @pb4uski. This is exactly what I am looking for. I'm working on tuning up my model to be able to project/forecast this.

Does your projection model tell you at what age the 107% or 109% are equal to 100%?

Are you still looking for a breakeven? It.does.not.apply.here. Some people seem to be thinking they need to justify paying the conversion tax, as if they may pay too much before they die. You could look at this another way. Since you can't spend your tIRA/401K dollars until you pay the taxes, and you died with a tIRA balance, you deferred too much. Does that sound crazy? Maybe. So does saying that you are looking for a breakeven.

Here's how I look at it. If I have $100K in a Roth, I have $100K to spend. If I have $100K in taxable, I have $100K minus whatever cap gains tax I have to pay before spending it. If I have $100K in a tIRA, I have $100K minus regular income tax due when withdrawing or converting. I make assumptions on the future tax rate, and that I won't be pulling it all out at once. That helps me see that a conversion is immediately a break even, if my assumptions were correct even if the tax rates are the same and I pay the tax out of the conversion. If I get favorable tax rate arbitrage, I'm ahead.

You earlier said that there's an element of speculation on future rates, which has been agreed to. To me, if you are not comfortable with that, it favors conversion, to pay the tax NOW, while it is at a known rate. Holding the funds in the tIRA is where you are speculating about what rates will be when you withdraw. Right? I'm locking in on the income tax rate when I convert. Maybe later I could do better, but I don't want to risk that.

You asked about measurable impact of converting. I figure that I'm paying somewhere around 5-10% less on taxes overall, giving me somewhere around $40K-$80K more money out of it in today's dollars. Some people belittle that as being a small % of their entire wealth. I think I heard 2% in this thread. I'll take that every time. Folks, there are no home runs to be had with such low risk. And forget about your entire portfolio, you only have control of your IRAs with conversions, so measure the gains on that money. I'm getting a couple of brand new cars out of this, for free! A little paper work to open the Roth, some calculations to try to make it as efficient as I can with my assumptions, and a few button clicks each year to do the transfer.

Maybe I'm underestimating the work of the calculations, because I enjoy doing this kind of thing. If you really want to understand this, you need to do some tax program runs to see what happens today if you do some conversions, and what happens in the future when you have SS income, perhaps a pension, and RMDs if you do and if you don't convert. Incorporate that into a spreadsheet which shows the value of your tIRA and Roth IRA in the lifetime of your IRA, with projections on portfolio growth.

Note I said the life of the IRA, because it lives on even if you die, so factor in what happens with your heirs. If you don't care, you might as well figure out what happens if you drain the IRA in your lifetime, because otherwise it's going to waste. If you want to leave some to charity, factor that in as an amount you don't want to convert. If you see using it as LTC self-insurance, see what happens if you start having huge medical expenses to write off income. But don't get lazy here. You likely have 10-15 or more years of RMDs until you have those LTC expenses, so run a fair scenario. I've done this for my parents, and found that withdrawing the full amount of their medical care from their tIRA does not result in 0 taxes. Not even close. It may for others. Run the numbers and see.

 

[Gumby]

I spent today building a new spreadsheet exactly along the lines you suggest in order to re-evaluate my Roth conversion plan, which is to convert to the first IRMAA threshold for the next ten years and then start RMDs. Versus the no conversion alternate, my plan puts us ahead ~2% in total value if we should live to 90. I'll take that.

 

[chassis]

I added Federal tax by bracket to my model and estimated the headspace for Roth conversions before age 72, then estimated the tax saved on the converted amount.

I just can't get excited about this. It seems much ado about nothing. It is not a home run as mentioned by @RunningBum.

@RunningBum with respect, what is your financial training? Mine is in both graduate school and as a business leader. I can't think of a financial action that does not come with an attendant analysis of payback, breakeven, NPV and IRR. This applies in my view to a Roth conversion.

If the concise business case for a Roth conversion can't be simply stated in 25 words or less, it doesn't get my attention.

We all agree there is speculation in the tax arbitrage nature of Roth conversions. Everything in life has some element of speculation. The speculation may work as we expected, or it may not. No one knows beforehand.

@RunningBum mentioned he might pay 5%-10% on taxes less overall. This is the payback I am referring to.

@pb4uski provided a form of a NPV calculation.

@Gumby provided another form of NPV calculation.

Folks, the Roth topic can and should be quantified for people to make quantified, informed decisions.

So far none of this is appetizing to me. I will continue to read threads on this site and run models to see if I can get myself to see the light.

 

[CardsFan]

I spent today building a new spreadsheet exactly along the lines you suggest in order to re-evaluate my Roth conversion plan, which is to convert to the first IRMAA threshold for the next ten years and then start RMDs. Versus the no conversion alternate, my plan puts us ahead ~2% in total value if we should live to 90. I'll take that.

2%? That is interesting.

My plan is the same as yours, though we only have 7 years to RMD's.

In the end it does not surprise me. Even the developer of I-Orp stated years ago benefits were marginal for most people (though I could not find that article last time I looked for it).

I still think the benefit to an heir could have a higher value, so that is my reason to convert.

Then again, DS and DDIL like to vote for people that will raise their taxes, so how hard should I work to pre-pay at a lower rate?

 

[Gumby]

I held the tax rates at current levels in my model. If they go up in 2026, conversion is slightly better than I have calculated. Another factor is that we have $100k per year income from COLA'd pensions and social security, so there is limited room for conversion. If we had started earlier and had less current income, conversion would also be better.

In any event, even if it is not earth-shattering, if I can pick up extra value without much effort, I'll take it.