Back during the runup to the 2020 election it looked like Biden might be elected with control of the legislature and it looked to me as though, for my circumstances, federal taxes might get a lot worse. I did a lot of modeling and published some articles. In the end, I paid a modest price to Roth convert everything. I thought of that decision as buying insurance, not as seeking a financial gain.
I think the headline is that Roth conversions are overrated in some circumstances and are brilliant in others. As we look at retiring early, we're considering using the lower income years before ss to reduce rmd’s and the tax burden we'll pass to our kids on those accounts. It seems the best strategies are a mixture of the pretax, no tax and taxable long term capital gains buckets to leverage the tax code as best as you can. Still researching while we're accumulating though.
That seems like the right way to go with this Christina. I was mostly trying to encourage folks to figure out for themselves what the real lifetime benefit is to them - and to heirs. I go to many financial dinners, for my research, to see what these financial groups are pushing. This is a topic where I constantly saw how these groups push it as a justification to hire them.
Maybe others on here know what products work best. I've heard good things about Pralana, but also I understand very complex to use. I created my own tool, but it is currently not suitable for public use (i.e. complex, lots of inputs, no user manual, no support).
You’d be surprised how good the LLM’s like perplexity are for this. Caveat is to ensure your prompt is very specific and thorough, you consult others like ChatGPT and you do sanity checks along the way. I might add we have a complex situation and it came through with excellent advice I’m acting on.
Thank you for this article. A few additional points. The Roth hypers tend to act like the positives (for conversions) will happen: taxes go up and you will live a long time. Neither of which is necessarily true. Also, this is a big one to me, can anyone ever see the government changing the rules? Like oh… these Roth’s worked way too well let’s double the RMD on them (by the way I would support this as a tax policy) or something like the max amount allowed in a Roth is 2MM. (If you have more your RMD is the amount to get you down to that number. After all, the government already did change the rules once from lifetime RMD’s to 10 year cash out for beneficiaries.
Thanks for the response Scott. I wasn't going to try and predict how the rules may change in the future - that would be pure speculation. More I was simply trying to see what benefit individuals could get from this in regards to maximizing lifetime wealth assuming current rules. The reason I wrote the article is some of my retired friends made such a big deal about how great the conversations are - but never providing me any numbers... And then all the financial seminars I've been to (as part of my research), where this always came up as a BIG reason to hire that financial advisor! Some people read my Roth article and assumed I was saying not to do it. Absolutely not true. I was more trying to encourage folks to figure out the real benefit to them and put it in a perspective of other key decsions.
I am a financial advisor. I lost a client 10 years ago because she insisted on doing a Roth IRA contribution rather than a traditional IRA contribution when her tax rate was 22% , retirement was 5 years away and her tax rate in retirement was going to be at most 12% with some years in the zero tax bracket. I wouldn’t do it so she went somewhere else; which is fine by me. FWIW I am 45 and do backdoor Roth’s. The main thing is to use the tool (as you did a great job laying out) that is best for your situation not being dogmatic to one thing like Roth’s. Also, don’t go to those dinners one of them is going to end up being thee most expensive free
"...Roth's worked way too well let's double the RMD on them..."
Roths don't have RMDs, but ok......(unless what you really mean is that the govt may start requiring RMDs on Roths as well as traditional IRAs? seems relatively unlikely to me that an RMD requirement would be added to Roths - at least for past contributions - or that's the way I would evaluate it in doing the analysis)
The lack of an RMD on Roths can be a meaningful potential benefit for those who are in a position where RMDs may push them into a higher tax bracket.
When a Roth is inherited it has an RMD. Many people converted to Roth’s in let’s say mid 2010’s or so using calculators that showed inherited Roth’s being stretched out for a long time by beneficiaries but the secure act blew that up by mandating they must be cashed in within ten years. My point is that the government can change the rules…. as they already have done….
An interesting read. As a fee-only financial planner who charges less than 1%, we use 2 different Roth conversion modeling tools: Right Capital and Holistiplan. Both are powerful yet complicated, and we invested dozens of hours learning them and now have years of experience using them with clients. As you note in the comments, every case deserves its own analysis. And there are unknowables - changes in tax laws and rates, when you and your spouse will die (change in filing status), beneficiary's financial circumstances at the time of inheritance, for example. We have recommended and implemented everything from no Roth conversions to $2M+ of total Roth conversions over several years. It depends. And it rarely involves converting 100% of pre-tax accounts. Having diverse tax statuses is similar to asset allocation diversity. If you could predict the future, neither would be necessary.
A modest benefit can be worth it for clients who have kids inheriting funds. Every client is different but pulling $1m inherited IRA out in 10 years vs a Roth over 10 years is a killer.
Also when a spouse dies and inherits and IRA from their spouse, RMDs can be quite large and now that spouse is at a single filer.
This is a case by case basis and what the client’s goals are.
Yup agree Alex - case-by-case. I encourage anyone considering Roth conversions to run the numbers for comparison using a good program. (I've heard good things about Pralana - except not that easy to use).
While I know this may apply to a limited audience, for those blessed with sufficient income, a traditional IRA will impose income taxes via the RMD. Roth withdrawals are purely voluntary, so if your goal is to leave $ to your heirs, spend money from your taxable accounts and let the Roth grow. The net tax rate in that scenario is 0%.
I now view my traditional IRA as my least favorite asset class-a vehicle to turn capital gains into ordinary income.
Thanks for the article. It appears that your models assume that the taxes are paid out of the traditional 401k/IRA account as they are converted to Roth. Is that the case? How do your models work if taxes are paid using funds from non-retirement accounts? That would allow the dollars not used to pay those taxes to grow in the Roth account and be tax free upon withdrawal.
Good question Ken. My detialed model does not assume taxes are paid from the IRA. By default, conversion taxes are paid from a taxable (non-retirement) account, which allows the full converted amount to go into the Roth and grow tax-free.
It also tracks the impact of reducing the taxable account to fund those taxes, so the tradeoff is fully captured. I can also run scenarios where taxes are paid from the IRA or split between sources, but the results I showed already reflect the “pay taxes from outside” case.
Hi Bryon. Thanks for clarifying. I couldn't tell from your charts because it looked like the "Tax Hit (upfront)" is happening in the Roth account itself since the Roth amount dips down that amount. I understand now that you are netting out the potential loss in the taxable account as well.
One additional point for those who need it, if you have a disabled child who has a special needs trust, a Roth may be especially helpful. If they are an Eligible Designated Beneficiary, they can spread the distribution over their remaining lifespan, not 10 years. In addition, if the Roth is beneficiary-designated to the Trust, there should be no tax on the appreciation within the Roth once inside the Trust, while I believe there would be with a traditional IRA.
I did a lot of modeling and for me, it did not make a compelling financial case. Plus, the thought of writing a check to the government when you really don’t have to was equally important.
what's good for the goose, is not good for the gander... this goose has a gander who will likely live longer so the effective tax rate goes from married to single. This scenario is to presented in any of your examples, but it is an interesting idea.
Like all tax strategies, it all depends. If an advisor has clients who possess meaningful wealth beyond their qualified accounts and will not have to depend on living off withdrawals from their retirement accounts, paying taxes on RMD withdrawals are taxes that can be avoided by moving assets out of these tax deferred accounts over to Roth IRAs that their heirs can inherit without the oneous of being forced to pay taxes on the value of those accounts over 10 years. So, paying taxes on $X in their 50s and 60s and allowing the Roth monies to compound for the next 20-30 or more years to $X++++ before they are eventually inherited by heirs who will never pay taxes on these appreciated funds makes a lot of sense. Even though the heirs must distribute all of the Roth monies by the 10th year, those distributions are not taxable.
Interesting article. But one confusing thing is the graphs show the traditional IRA taxes only being paid at age 90, but actually taxable RMDs would start years before that. Those RMDs would also reduce to some extent the IRA balance that could still compound. More generally, I think comparing the ending balances between the trad. and Roth options is not the best way to compare them. If you have enough in your IRA, the RMDs can end up keeping you in the higher tax brackets, meaning higher medicare premiums, higher social security taxes, etc. The government will consider you "rich" and make you pay your "fair share" for the rest of your life, just when your health expenses may be going up. (And that "fair share" is definitely going up in the years to come.) I think it's better to compare how much after-tax income you have during your retirement years, after the RMDs start.
Thanks for the comment William. Please note the first 3 graphs are very simplistic trying to show the main effect of the starting tax bracket and ending tax bracket. These do not account for RMDs, etc. The point of these was simply to show the main factor in "tax arbitrage" for readers to better understand what is mainly going on when trying the get a benefit out of the conversion.
But note the 4th graph came out of my very complex Roth calculator for one example persona. This calculator includes all the many parameters involved including: Starting taxable, trad and Roth account amounts, effects of RMDs, conversion fill tax rate max, account spend order, IRMAA, ACA, fed and state tax tables, spending, income, and more. The two lines are showing the sum value of all accounts over the retirement years for doing Roth conversions versus not doing the conversions.
Regarding how much you would have for spending in the latter years (e.g. if you had to do a big cash out for a medical emergency, etc.), this can be done with tools such as this one I created. I simply could not cover all the variables involved in this short article. If you use a tool like Pralana, you can run countless what if scenarios.
Great article and I appreciate the detailed reasoning. Still, isn't it the case that if you are going to have low income in the first 10 years of retirement (say less than $100K) and then RMDs from large traditional retirement accounts make your income much higher (say 7 to 10X) that incremental Roth conversion during those first 10 years will be very advantageous?
Also, can you say more about the situation that only has upside that you use but requires support from an employer plan? I am not sure what you are referring to.
Yes Dan - the biggest "tax arbitrage" effect is if you can do the conversions when your tax bracket is low over as many years as you can (and especially before RMDs hit if you have large accounts) that could result in higher tax brackets dur to the RMDs. But the tricky part is how to do those early conversions without boosting your tax bracket too high. This is why you really need a good Roth calculator to figure it out. (note: don't use the free ones from financial advisor companies!).
The situation I was referring to was a company 401k plan that allows after tax funds (if you max out your pretax). Then you can typically do what is known as a "mega Roth conversion". This is a way to take the after tax basis funds - meaning the starting amounts not incl. tax deferred gains - and converting them to a Roth account. This is basically a Free Roth conversion. I did this - it's a no brainer, because you pay no taxes and your Roth can now grow without a tax liability when you cash it out.
Tricky to publish this one because it is very complex with a large number of inputs. The need for a user manual, support, etc. I may try in this in the future but I would need to spend a lot of hours to refine it for public use with currently zero payback. Maybe there is a way to make it work.
Back during the runup to the 2020 election it looked like Biden might be elected with control of the legislature and it looked to me as though, for my circumstances, federal taxes might get a lot worse. I did a lot of modeling and published some articles. In the end, I paid a modest price to Roth convert everything. I thought of that decision as buying insurance, not as seeking a financial gain.
The real trick is to harvest enough capital losses to offset the taxes on the Roth conversion.
Use a SMA with short/long sleeves.
And the cost of conversions can be lowered or eliminated.
I think the headline is that Roth conversions are overrated in some circumstances and are brilliant in others. As we look at retiring early, we're considering using the lower income years before ss to reduce rmd’s and the tax burden we'll pass to our kids on those accounts. It seems the best strategies are a mixture of the pretax, no tax and taxable long term capital gains buckets to leverage the tax code as best as you can. Still researching while we're accumulating though.
That seems like the right way to go with this Christina. I was mostly trying to encourage folks to figure out for themselves what the real lifetime benefit is to them - and to heirs. I go to many financial dinners, for my research, to see what these financial groups are pushing. This is a topic where I constantly saw how these groups push it as a justification to hire them.
What commercial product gives the best overall clarity to the issue of Roth conversions?
Maybe others on here know what products work best. I've heard good things about Pralana, but also I understand very complex to use. I created my own tool, but it is currently not suitable for public use (i.e. complex, lots of inputs, no user manual, no support).
You’d be surprised how good the LLM’s like perplexity are for this. Caveat is to ensure your prompt is very specific and thorough, you consult others like ChatGPT and you do sanity checks along the way. I might add we have a complex situation and it came through with excellent advice I’m acting on.
Thank you for this article. A few additional points. The Roth hypers tend to act like the positives (for conversions) will happen: taxes go up and you will live a long time. Neither of which is necessarily true. Also, this is a big one to me, can anyone ever see the government changing the rules? Like oh… these Roth’s worked way too well let’s double the RMD on them (by the way I would support this as a tax policy) or something like the max amount allowed in a Roth is 2MM. (If you have more your RMD is the amount to get you down to that number. After all, the government already did change the rules once from lifetime RMD’s to 10 year cash out for beneficiaries.
Thanks again.
Thanks for the response Scott. I wasn't going to try and predict how the rules may change in the future - that would be pure speculation. More I was simply trying to see what benefit individuals could get from this in regards to maximizing lifetime wealth assuming current rules. The reason I wrote the article is some of my retired friends made such a big deal about how great the conversations are - but never providing me any numbers... And then all the financial seminars I've been to (as part of my research), where this always came up as a BIG reason to hire that financial advisor! Some people read my Roth article and assumed I was saying not to do it. Absolutely not true. I was more trying to encourage folks to figure out the real benefit to them and put it in a perspective of other key decsions.
I am a financial advisor. I lost a client 10 years ago because she insisted on doing a Roth IRA contribution rather than a traditional IRA contribution when her tax rate was 22% , retirement was 5 years away and her tax rate in retirement was going to be at most 12% with some years in the zero tax bracket. I wouldn’t do it so she went somewhere else; which is fine by me. FWIW I am 45 and do backdoor Roth’s. The main thing is to use the tool (as you did a great job laying out) that is best for your situation not being dogmatic to one thing like Roth’s. Also, don’t go to those dinners one of them is going to end up being thee most expensive free
Dinner you ever went to.
I go to those dinners for the awesome great dinner and research material for my newsletter!
"...Roth's worked way too well let's double the RMD on them..."
Roths don't have RMDs, but ok......(unless what you really mean is that the govt may start requiring RMDs on Roths as well as traditional IRAs? seems relatively unlikely to me that an RMD requirement would be added to Roths - at least for past contributions - or that's the way I would evaluate it in doing the analysis)
The lack of an RMD on Roths can be a meaningful potential benefit for those who are in a position where RMDs may push them into a higher tax bracket.
When a Roth is inherited it has an RMD. Many people converted to Roth’s in let’s say mid 2010’s or so using calculators that showed inherited Roth’s being stretched out for a long time by beneficiaries but the secure act blew that up by mandating they must be cashed in within ten years. My point is that the government can change the rules…. as they already have done….
An interesting read. As a fee-only financial planner who charges less than 1%, we use 2 different Roth conversion modeling tools: Right Capital and Holistiplan. Both are powerful yet complicated, and we invested dozens of hours learning them and now have years of experience using them with clients. As you note in the comments, every case deserves its own analysis. And there are unknowables - changes in tax laws and rates, when you and your spouse will die (change in filing status), beneficiary's financial circumstances at the time of inheritance, for example. We have recommended and implemented everything from no Roth conversions to $2M+ of total Roth conversions over several years. It depends. And it rarely involves converting 100% of pre-tax accounts. Having diverse tax statuses is similar to asset allocation diversity. If you could predict the future, neither would be necessary.
A modest benefit can be worth it for clients who have kids inheriting funds. Every client is different but pulling $1m inherited IRA out in 10 years vs a Roth over 10 years is a killer.
Also when a spouse dies and inherits and IRA from their spouse, RMDs can be quite large and now that spouse is at a single filer.
This is a case by case basis and what the client’s goals are.
Yup agree Alex - case-by-case. I encourage anyone considering Roth conversions to run the numbers for comparison using a good program. (I've heard good things about Pralana - except not that easy to use).
While I know this may apply to a limited audience, for those blessed with sufficient income, a traditional IRA will impose income taxes via the RMD. Roth withdrawals are purely voluntary, so if your goal is to leave $ to your heirs, spend money from your taxable accounts and let the Roth grow. The net tax rate in that scenario is 0%.
I now view my traditional IRA as my least favorite asset class-a vehicle to turn capital gains into ordinary income.
Thanks for the article. It appears that your models assume that the taxes are paid out of the traditional 401k/IRA account as they are converted to Roth. Is that the case? How do your models work if taxes are paid using funds from non-retirement accounts? That would allow the dollars not used to pay those taxes to grow in the Roth account and be tax free upon withdrawal.
Good question Ken. My detialed model does not assume taxes are paid from the IRA. By default, conversion taxes are paid from a taxable (non-retirement) account, which allows the full converted amount to go into the Roth and grow tax-free.
It also tracks the impact of reducing the taxable account to fund those taxes, so the tradeoff is fully captured. I can also run scenarios where taxes are paid from the IRA or split between sources, but the results I showed already reflect the “pay taxes from outside” case.
Hi Bryon. Thanks for clarifying. I couldn't tell from your charts because it looked like the "Tax Hit (upfront)" is happening in the Roth account itself since the Roth amount dips down that amount. I understand now that you are netting out the potential loss in the taxable account as well.
Thanks for the detailed discussion on this topic.
One additional point for those who need it, if you have a disabled child who has a special needs trust, a Roth may be especially helpful. If they are an Eligible Designated Beneficiary, they can spread the distribution over their remaining lifespan, not 10 years. In addition, if the Roth is beneficiary-designated to the Trust, there should be no tax on the appreciation within the Roth once inside the Trust, while I believe there would be with a traditional IRA.
Thanks Joey - That's good to know!
I did a lot of modeling and for me, it did not make a compelling financial case. Plus, the thought of writing a check to the government when you really don’t have to was equally important.
what's good for the goose, is not good for the gander... this goose has a gander who will likely live longer so the effective tax rate goes from married to single. This scenario is to presented in any of your examples, but it is an interesting idea.
Like all tax strategies, it all depends. If an advisor has clients who possess meaningful wealth beyond their qualified accounts and will not have to depend on living off withdrawals from their retirement accounts, paying taxes on RMD withdrawals are taxes that can be avoided by moving assets out of these tax deferred accounts over to Roth IRAs that their heirs can inherit without the oneous of being forced to pay taxes on the value of those accounts over 10 years. So, paying taxes on $X in their 50s and 60s and allowing the Roth monies to compound for the next 20-30 or more years to $X++++ before they are eventually inherited by heirs who will never pay taxes on these appreciated funds makes a lot of sense. Even though the heirs must distribute all of the Roth monies by the 10th year, those distributions are not taxable.
Agree it benefits heirs the most - especially if they in a high tax bracket.
Interesting article. But one confusing thing is the graphs show the traditional IRA taxes only being paid at age 90, but actually taxable RMDs would start years before that. Those RMDs would also reduce to some extent the IRA balance that could still compound. More generally, I think comparing the ending balances between the trad. and Roth options is not the best way to compare them. If you have enough in your IRA, the RMDs can end up keeping you in the higher tax brackets, meaning higher medicare premiums, higher social security taxes, etc. The government will consider you "rich" and make you pay your "fair share" for the rest of your life, just when your health expenses may be going up. (And that "fair share" is definitely going up in the years to come.) I think it's better to compare how much after-tax income you have during your retirement years, after the RMDs start.
Thanks for the comment William. Please note the first 3 graphs are very simplistic trying to show the main effect of the starting tax bracket and ending tax bracket. These do not account for RMDs, etc. The point of these was simply to show the main factor in "tax arbitrage" for readers to better understand what is mainly going on when trying the get a benefit out of the conversion.
But note the 4th graph came out of my very complex Roth calculator for one example persona. This calculator includes all the many parameters involved including: Starting taxable, trad and Roth account amounts, effects of RMDs, conversion fill tax rate max, account spend order, IRMAA, ACA, fed and state tax tables, spending, income, and more. The two lines are showing the sum value of all accounts over the retirement years for doing Roth conversions versus not doing the conversions.
Regarding how much you would have for spending in the latter years (e.g. if you had to do a big cash out for a medical emergency, etc.), this can be done with tools such as this one I created. I simply could not cover all the variables involved in this short article. If you use a tool like Pralana, you can run countless what if scenarios.
Great article and I appreciate the detailed reasoning. Still, isn't it the case that if you are going to have low income in the first 10 years of retirement (say less than $100K) and then RMDs from large traditional retirement accounts make your income much higher (say 7 to 10X) that incremental Roth conversion during those first 10 years will be very advantageous?
Also, can you say more about the situation that only has upside that you use but requires support from an employer plan? I am not sure what you are referring to.
Thanks for your insights
Yes Dan - the biggest "tax arbitrage" effect is if you can do the conversions when your tax bracket is low over as many years as you can (and especially before RMDs hit if you have large accounts) that could result in higher tax brackets dur to the RMDs. But the tricky part is how to do those early conversions without boosting your tax bracket too high. This is why you really need a good Roth calculator to figure it out. (note: don't use the free ones from financial advisor companies!).
The situation I was referring to was a company 401k plan that allows after tax funds (if you max out your pretax). Then you can typically do what is known as a "mega Roth conversion". This is a way to take the after tax basis funds - meaning the starting amounts not incl. tax deferred gains - and converting them to a Roth account. This is basically a Free Roth conversion. I did this - it's a no brainer, because you pay no taxes and your Roth can now grow without a tax liability when you cash it out.
Good article, interesting that solutions still focus on Excel formulas instead of providing AI scripts.
Ramy - you mean the tools out there today?
Would love to test your Roth conversion tool and help if I can.
Tricky to publish this one because it is very complex with a large number of inputs. The need for a user manual, support, etc. I may try in this in the future but I would need to spend a lot of hours to refine it for public use with currently zero payback. Maybe there is a way to make it work.