Roth Conversions - McQuarrie study

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
FiveK
Posts: 15742
Joined: Sun Mar 16, 2014 2:43 pm

Re: Roth Conversions - McQuarrie study

Post by FiveK »

hulburt1 wrote: Fri Jun 18, 2021 3:38 pm I'm 68 single with 1.1m Ira and 220000 Roth. I can easily live at 12%. But once SS kicks in and RMD at 72 I will be at top 22%. I live very nice at 25000 a year. Running all my no. (10000 pension, RMD 40000, SS 26000 and a part time job 10000) My job go's right to state and Fed. tax's. But I do not pay tax on my Ira with draw or what I put in Roth that's tax is covered by my job. Should I stay at 12% or run it up to about $80000 and put more In Roth and Take home more. Thanks :confused
Taking your numbers as given, before SS and RMDs start the marginal rate picture for your Roth conversions is here:
Image
Including a brief phase-out of the earned income credit, converting up to the top of the 12% bracket costs you ~12%.

Beyond that, IRMAA kicks in if you want to convert enough to bring your tIRA balance down (depending on what return rate you assume). You could pay 22% up to the first IRMAA tier, or ~24% if you include all the money within the 22% bracket.
Image

Once the $26K SS benefit kicks in, you'll pay at least 24% on the RMD (assuming a minimum $40K RMD), and more than that on higher withdrawals.
Image

In short, converting now to the top of the 12% bracket seems a "go do". Beyond that it's close to six of one or a half dozen of the other. With a reasonably sharp pencil, you could justify converting now up to (but not including) the second IRMAA tier.

With a very sharp pencil, assuming you have cash on hand to pay the conversion tax, and depending on what you assume about tax drag in a taxable account, you could justify converting even higher amounts.

Or you could justify not converting beyond the top of the 12% bracket. Depends on your assumptions.... ;)

See the personal finance toolbox if you don't mind using Excel to replicate the charts above.

While the numbers are specific to hulburt1's question, the analysis strategy is generic and perhaps applicable to this thread's discussion. Up to the admins if they want to leave this here or delete it, as I've copied the post to hulburt1's new thread.
DSBH
Posts: 740
Joined: Tue Jul 21, 2020 3:31 pm
Location: Texas

Re: Roth Conversions - McQuarrie study

Post by DSBH »

McQ wrote: Fri Jun 18, 2021 1:39 pm ...
Best, probably, if you adapt the spreadsheet to your needs, and summarize for the forum why aggressive conversions still make sense to you--especially if for a reason other than those in the paper, i.e., a reason other than that you have the opportunity to convert at 0%. Then I can comment, particularly if you point to remaining puzzlements / issues / quandaries.
My primary reason for Roth IRA conversion is the risk of MFJ becoming Single for tax filing purpose, since both parents of a spouse passed away in their very early 70's, so I am very interested in your future comments.

Also, I did a significant Roth IRA conversion in late March 2020 when the stocks market dropped around 30% from high, at a marginal tax rate higher than what I expect when RMD is supposed to begin for me in 9 years, thinking that whenever the market returns to the same Jan/Feb 2020 level I would come out ahead compared to the no conversion case. I would appreciate your comments on what you would do in such scenario, or what you think Rob and Sue would do, and why.

Thank you for the paper and I will study it.
John C. Bogle: "Never confuse genius with luck and a bull market".
User avatar
iceport
Posts: 6054
Joined: Sat Apr 07, 2007 4:29 pm

Re: Roth Conversions - McQuarrie study

Post by iceport »

This is slightly off topic, and I'm still chugging through the text. But I was disappointed with the discussion on tax diversification as a way to address tax uncertainty. With all due respect to someone who, I have no doubt, is both more learned and more intelligent than I am, the discussion of tax diversification on P. 17 seems pretty much entirely off-base.

One major flaw is in this logic:
A portfolio is diversified by adding a distinct asset believed to be uncorrelated, as when bonds are added to a stock portfolio. But there is only one Congress responsible for the tax laws governing retirement accounts in the US. There is no diversification away from it. No independent taxing authority stands guard over Roth provisions.
That's not the concept behind using tax diversification as a hedge against future tax changes, that there would have to be different entities forcing uncorrelated changes to various tax laws.

What the concept of tax diversification relies upon is only that different types of accounts will be taxed in different ways and at different rates than one another, and that the specific tax treatments in the future are uncertain today.

So what if the same congress were to change tax laws related to all three main types of accounts — tax-deferred, tax-free, and taxable — at the same time and in the same direction? Unless the author is asserting that all of the accounts will end up being taxed in the same manner and at the same rates, the concept of tax diversification would still be valid.

Does the author really think that future legislation will result in withdrawals from all three types of accounts being taxed identically? If not, then isn't it likely that the taxation of one type of account would still be more advantageous than another, for a given taxpayer in a given year? In that case, the taxpayer would benefit from the option to withdraw from the type of account that produces the most favorable tax treatment. That's tax diversification.

The second major flaw is the author's failure to acknowledge that changes in tax laws is only one source of tax uncertainty.

The other primary source of tax uncertainty is in the variability of an individual's tax liability from year to year. That has absolutely nothing to do with changes in tax laws.

An income spike or spending spike in a given year — large one-time purchases, emergencies, family calamities, sales of assets, etc. — could drastically change a taxpayer's tax liability from year to year. And with the potential for spending or income spikes to change the individual's tax rates, there also comes the potential for a different withdrawal strategy to become advantageous that particular year than the one used in more typical years. Such a taxpayer could benefit from having different types of accounts to choose to make withdrawals from. That's also when tax diversification becomes potentially valuable.

Assuming a taxpayer's income and spending will remain consistent from year to year when considering future tax liabilities is just as wrong as ignoring sequence of returns risk in deciding upon a prudent portfolio withdrawal rate. What happens in each individual year matters.

Having options to mitigate the effects of changing individual circumstances is part of the value tax diversification.

The rest of the paper might very well be spot-on. But the rationale for dismissing — or "exploding" — the value of tax diversification makes little sense to me, at least as I understand the concept.
"Discipline matters more than allocation.” |—| "In finance, if you’re certain of anything, you’re out of your mind." ─William Bernstein
McQ
Posts: 1425
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

lazynovice wrote: Fri Jun 18, 2021 2:03 pm
McQ wrote: Fri Jun 18, 2021 1:28 pm
jeffyscott wrote: Fri Jun 18, 2021 12:04 pm
McQ wrote: Fri Jun 18, 2021 1:04 am Bogleheads: Edward McQuarrie here, happy to answer questions directly. An example of the spreadsheet I used to compare outcomes can be downloaded here: http://www.edwardfmcquarrie.com/?p=573. If you don't think my conclusions apply to your particular case, the spreadsheet allows you to confirm that mismatch. But if you haven't set up a spreadsheet to evaluate your Roth conversion, then you may have fallen prey to what my behavioral finance colleagues call "misleading heuristics."
Yep, I have gone with this one: If a conversion today will be taxed at a rate lower than future RMDs, it is likely to be beneficial.
Though modified to some extent, in that I am ony assuming this for conversions in the 12% bracket vs. future marginal rate of 22% on RMDs, if we were to do no conversions. I'd not convert at 22% to avoid potential for RMDs to be taxed at 24%.
Next, I want to reiterate a key conclusion: "A Roth conversion, within the parameters examined, always pays off, but always slowly, and always in small amounts." The exception is conversions in the 0% bracket and neighboring brackets, which pay off early and sizably.
I assume this is simply because this typically means paying 0, 10, or 12% now vs. 22% (or more) later, so the the differential is at least 10%. The 12 to 22% jump is the biggest one, the next closest is the 8% gap between 24 and 32% (which are not brackets of concern to me).

Maybe a new consensus could be something like: If a conversion today will be taxed at a rate at least 10 pecentage points lower than future RMDs, it is likely to be beneficial. :?:
Actually, subtleties of exponential math combined with the difference between marginal and average tax rates make it impossible to build a simple rule of thumb around the arithmetic difference in present vs. future tax rate. Larger gaps do favor Roth conversions, of course; but nothing succeeds like converting in the 0%/10%/12% bracket to prevent RMDs taxed at 22% and above.
I think most on the forum will generally agree that conversions make sense for most of us up to the 12%.

The 22% is where there is a clash. It seems the more pension income you have, the higher your deferred balance the greater the age difference between spouses, and the more pessimistic you are about future rates, the more likely you are to convert at 22%.

The general question for you is how sensitive each of those variables are to the decision. For whom does it make sense to convert at 22%?
I concur that 12% or less is almost always a go. 22% will also almost always work if you have zero taxable income and are filling the 0%, 10%, 12% and 22% brackets (average rate < 15% on the conversion). If the 12% bracket is already filled, then it is more thinkers, with all the conversion taxed at 22%. Higher pension, a 457 as well as a 401k balance per an earlier comment (civil servant case), etc. make it more likely that you will be high up in the 22% bracket, which may mean IRMAA, which will make a conversion today at 22% relatively more attractive. BTW, when I set a $1.8 million TDA balance as the threshold for entering the 22% bracket seven years hence, I assumed $70,000 social security + pension for MFJ. If that is only $50,000, then you will need $2.3 million to nose into the 22% bracket.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
McQ
Posts: 1425
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

DSBH wrote: Fri Jun 18, 2021 4:51 pm
McQ wrote: Fri Jun 18, 2021 1:39 pm ...
Best, probably, if you adapt the spreadsheet to your needs, and summarize for the forum why aggressive conversions still make sense to you--especially if for a reason other than those in the paper, i.e., a reason other than that you have the opportunity to convert at 0%. Then I can comment, particularly if you point to remaining puzzlements / issues / quandaries.
My primary reason for Roth IRA conversion is the risk of MFJ becoming Single for tax filing purpose, since both parents of a spouse passed away in their very early 70's, so I am very interested in your future comments.

Also, I did a significant Roth IRA conversion in late March 2020 when the stocks market dropped around 30% from high, at a marginal tax rate higher than what I expect when RMD is supposed to begin for me in 9 years, thinking that whenever the market returns to the same Jan/Feb 2020 level I would come out ahead compared to the no conversion case. I would appreciate your comments on what you would do in such scenario, or what you think Rob and Sue would do, and why.

Thank you for the paper and I will study it.
Good market timing combined with smart asset location will always produce good results. Translation: moving high return assets into the most tax-favored account structure at just the moment when they were temporarily depreciated was a good move. But as I hope you will accept, I could not take that as a general case in the paper.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
DSBH
Posts: 740
Joined: Tue Jul 21, 2020 3:31 pm
Location: Texas

Re: Roth Conversions - McQuarrie study

Post by DSBH »

McQ wrote: Fri Jun 18, 2021 6:14 pm
DSBH wrote: Fri Jun 18, 2021 4:51 pm
McQ wrote: Fri Jun 18, 2021 1:39 pm ...
Best, probably, if you adapt the spreadsheet to your needs, and summarize for the forum why aggressive conversions still make sense to you--especially if for a reason other than those in the paper, i.e., a reason other than that you have the opportunity to convert at 0%. Then I can comment, particularly if you point to remaining puzzlements / issues / quandaries.
My primary reason for Roth IRA conversion is the risk of MFJ becoming Single for tax filing purpose, since both parents of a spouse passed away in their very early 70's, so I am very interested in your future comments.

Also, I did a significant Roth IRA conversion in late March 2020 when the stocks market dropped around 30% from high, at a marginal tax rate higher than what I expect when RMD is supposed to begin for me in 9 years, thinking that whenever the market returns to the same Jan/Feb 2020 level I would come out ahead compared to the no conversion case. I would appreciate your comments on what you would do in such scenario, or what you think Rob and Sue would do, and why.

Thank you for the paper and I will study it.
Good market timing combined with smart asset location will always produce good results. Translation: moving high return assets into the most tax-favored account structure at just the moment when they were temporarily depreciated was a good move. But as I hope you will accept, I could not take that as a general case in the paper.
Yes a 30% drop should not be a general case, but what would you think about something much smaller for instance a 5% drop?
John C. Bogle: "Never confuse genius with luck and a bull market".
cbeck
Posts: 640
Joined: Sun Jun 24, 2012 1:28 am

Re: Roth Conversions - McQuarrie study

Post by cbeck »

McQ wrote: Fri Jun 18, 2021 1:24 pm
Chip: certainly seems reasonable that Bogleheads would attract super savers. But it does not seem reasonable, in the general case, that anyone other than a Boglehead would have both maxed out their 401(k) and piled up hundreds of thousands of dollars in taxable funds + saved kid's college + lived in a nice house (e.g., 30% savings rate, where I assumed 15% was the max reasonable). Nonetheless, if you do have years of living expenses in your taxable account, then you can convert at 0%/10%/12%, year after year. That was absolutely the best Roth conversion outcome I could find among all examined. But don't convert so much that age 72 income falls out of the 22% bracket.
Might happen more than you think. Suppose at retirement age your IRAs are stuffed and you sell your home to realize a substantial gain, but you don't intend to buy another home, because you prefer renting for some reason. So, the proceeds from the sale of the home can fund a series of Roth conversions prior to the RMDs and SS at age 70.
User avatar
JAZZISCOOL
Posts: 2658
Joined: Sat May 18, 2019 11:49 am
Location: Colorado - 5,700 ft.

Re: Roth Conversions - McQuarrie stud

Post by JAZZISCOOL »

LilyFleur wrote: Thu Jun 17, 2021 4:11 pm
lazynovice wrote: Thu Jun 17, 2021 4:02 pm
celia wrote: Thu Jun 17, 2021 3:55 pm
indexlover wrote: Thu Jun 17, 2021 1:59 pm Hello BHs,

Have you seen this article ? https://www.marketwatch.com/story/to-ro ... 1623431970

and the study, “When and for Whom Are Roth Conversions Most Beneficial?,” by Edward McQuarrie, a professor emeritus at the Leavey School of Business at Santa Clara University.

https://poseidon01.ssrn.com/delivery.ph ... INDEX=TRUE
I would have loved to read it, but it now looks disabled. If anyone downloaded the .pdf file, can you search for "surviving spouse" or just "spouse" and see if it says anything about the survivor having to file as Single instead of MFJ. Also see if it mentions "IRMAA". Thanks.
Does not mention moving to single brackets but he finds tax rate changes very immaterial- see my answer above.

He takes IRMAA into account.

You can read it for free if you sign up with an email address.
Considering how many adults in the United States are single, I find it interesting how often financial articles and forums only address MFJ.

What if Rob has a heart attack and dies? What if Rob finds a wife with less mileage on her, and divorces Sue? What is Sue going to do? Or Rob, if those things happen to him?

Even in the thread present in our forum right now about how you made it to $3 million, it isn't specified whether that is $1.5 million for singles, or $3 million whether you're married or not. It's a lot harder to get to $3 million as a single. The single tax brackets are a challenge.
Agree. Also, the information in the financial press that does mention singles is often limited and lacks substantial research IMO.
international001
Posts: 2748
Joined: Thu Feb 15, 2018 6:31 pm

Re: Roth Conversions - McQuarrie study

Post by international001 »

lazynovice wrote: Thu Jun 17, 2021 4:00 pm I’d really encourage everyone to read the study itself. The author addresses many of the objections being raised here.

One finding is that if you withdraw the money from the Roth before a certain break even period, you didn’t get any benefit. The break even point is 13 to 15 years best case.
Didn't read it (yet)
CAn you explain that break even point?

I thougght if you do a conversion at a tax bracket 12% and you spend the money later when your tax bracket is 22%, then you win. Am I wrong?
lazynovice
Posts: 3370
Joined: Mon Apr 16, 2012 10:48 pm
Location: Denver area. Former Texan.

Re: Roth Conversions - McQuarrie study

Post by lazynovice »

international001 wrote: Fri Jun 18, 2021 7:43 pm
lazynovice wrote: Thu Jun 17, 2021 4:00 pm I’d really encourage everyone to read the study itself. The author addresses many of the objections being raised here.

One finding is that if you withdraw the money from the Roth before a certain break even period, you didn’t get any benefit. The break even point is 13 to 15 years best case.
Didn't read it (yet)
CAn you explain that break even point?

I thougght if you do a conversion at a tax bracket 12% and you spend the money later when your tax bracket is 22%, then you win. Am I wrong?
See upthread where McQ has answered your second question. McQ is the study author. He is going to do a better job than I ever could of answering. Yes, you are correct.
Exchme
Posts: 1335
Joined: Sun Sep 06, 2020 3:00 pm

Re: Roth Conversions - McQuarrie study

Post by Exchme »

Prof McQ:

I love that you took the time to produce the paper, it's great having some academic interest in a subject of practical importance to millions of people. Seems like most writing on such topics these days comes from people selling something.

I like the precision of referring to inter temporal tax arbitrage part of a Roth specifically. But not sure I understand the dismissal of loading the Roth with stocks while slowing tax deferred with bonds. Are you saying that it is incorrect to do that or that Roth advocates do careless/opportunistic math (bottom of p10/top of p11)?

It does seem like bonds in TDA and stocks in Roth is good advice and doesn't holding two different types of accounts give us the chance to get a boost from this? If there is a boost, is it really separable from the inter temporal tax arbitrage benefit?

Thanks!
User avatar
beyou
Posts: 6915
Joined: Sat Feb 27, 2010 2:57 pm
Location: If you can make it there

Re: Roth Conversions - McQuarrie study

Post by beyou »

To me the driver is to max the benefit of a low married tax bracket while my spouse and I are alive. Frankly whether i would spend to 401k t-ira withdrawals or convert to roth is purely a function of having sufficient cash to pay bills. Why NOT put in Roth IRA if I don’t need the cash ?


As long as I have taxable account with some income and assets without huge gains (bonds), and later SS, why do I need the 401k or t-ira ? I just want to minimize taxes on those accounts given they have to be paid sometime, pay when likely a low rate. Year by year decision.
ChrisC
Posts: 1475
Joined: Tue Jun 19, 2012 9:10 am
Location: North Carolina

Re: Roth Conversions - McQuarrie study

Post by ChrisC »

McQ wrote: Fri Jun 18, 2021 6:10 pm Higher pension, a 457 as well as a 401k balance per an earlier comment (civil servant case), etc. make it more likely that you will be high up in the 22% bracket, which may mean IRMAA, which will make a conversion today at 22% relatively more attractive. BTW, when I set a $1.8 million TDA balance as the threshold for entering the 22% bracket seven years hence, I assumed $70,000 social security + pension for MFJ. If that is only $50,000, then you will need $2.3 million to nose into the 22% bracket.
Do you have TDA balance threshold for entering the 24%? A number of civil servant/military annuitants/pensioners easily enter the 22% bracket in retirement at an early retirement age, and some can, or have a spouse that can, draw SS retirement benefits that place them as a couple squarely within the 24% bracket. We've been converting into the higher range of the 24% the past several years, with the expectation that (1) we will never be below that bracket when SS retirement benefits kick-in for my spouse (which will occur later this year); (2) we will never touch Roth or TDA balances for living expenses (and can pay Federal and state income taxes from pension draws and taxable accounts); (3) all of our retirements accounts in TSP, 401Ks or Roth IRAs will be inherited by our children who are in tax brackets higher than their parents.

My gut tells me that we should still be converting into the 24% bracket and perhaps even higher pre-and post RMDS. If one of us passes, the survivor will wind up in the 32% bracket even if some of the pension and SS retirement benefits decrease. Also, some former civil servants have the ability to avoid IRMAA altogether by not enrolling in Medicare Part B or Part D and by relying on Government employer sponsored retiree health coverage (FEHB).

Edited: So, after posting the above, I read this note to your Table 2 regarding TDA threshold balance for entering into the 24% bracket:

"Pension income. If the client is a public sector employee or other individual who expects a pension as well as social security,
TDA balances would be lower than in the table. Each $10,000 in pension payments would reduce the required TDA balance to
enter a bracket by $273,000. Returning to Rob and Sue, if one of them had a pension of $50,000, then instead of needing
$4.175 million to enter the 24% tax bracket, only about $2.8 million would be required."

Perhaps, I'm misuderstanding the above note but if my pension/social security income is $174,000, then the $4.175 million in the threshold balance you postulated for TDA would be reduced by 17.4 x $273,000, which results in -$575,200.
curmudgeon
Posts: 2630
Joined: Thu Jun 20, 2013 11:00 pm

Re: Roth Conversions - McQuarrie study

Post by curmudgeon »

curmudgeon wrote: Thu Jun 17, 2021 5:38 pm I took the time to read the paper and it's appendices. I actually agree with the general point that for *many* people, Roth conversions will have very limited effect. But I wasn't too impressed with the paper otherwise. It started with a rather off-the-wall "base case", and continued on to other more or less realistic extensions. A huge, gaping hole was that it made zero acknowledgment of the fact that there will almost certainly be a period of years where a surviving spouse will be being taxed with much smaller single tax brackets.

The author does acknowledge the possibility of people with the ability to live off of taxable savings while doing Roth conversions (someone who sold a business), but writes it off as an edge case. He neglects the reality that his "high income" couple is likely to have saved beyond just their 401Ks, may downsize, have an inheritance, etc. He also seriously neglects the "early retiree" case; age 65 is not early retiring unless you are a college professor.
So I went back and read the paper again. I can't say it improved upon re-reading. A few sections of it are argumentative without solid foundation (tax diversification, savings rates, other results). The base case is highly contrived (more on that later).

The single biggest flaw is that it mentions the "tax torpedo" of taxation of SS benefits, but then proceeds to ignore it (buried in note 3) by skewing the cases examined to be outside the impacted boundaries. Using the example of a joint SS benefit of $100K, retirees with IRA balances between $250K and $2.5 million will be in that tax torpedo zone at initial RMD and subject to considerably higher marginal rates than used in this paper.

I find the case of a high income couple having no savings outside their 401K to be a stretch, though that could certainly apply to some.

It appears that the "end wealth", used to calculate his "breakeven age", values tIRA accounts and Roth as equivalent. I can say without question that your heirs, unless they are a charity, will take a very different view of the matter.

While it's good to investigate the effects of tax rate changes, a starting point for that would logically be the ones that are currently being actively pushed (loss of step-up basis, higher LTCG rates) in addition to bracket moves scheduled in current law. Estate tax may also become more relevant.

There's a fair bit of hand-waving around the gap between age 65 and 72. The author asserts that his couple has no other savings, but he has them delay SS until age 70, and IRA withdrawal until age 72. What are they living on? Fuzzily implied in there is that one of them can't retire until 72!

All in all, the premise that "Roth conversions aren't needed/valuable for everyone" (which is quite true) gets badly mangled in this paper. If you are going to stake out one corner of the SS taxaction space, you should be more forthcoming about the rest of the zone.
cas
Posts: 2258
Joined: Wed Apr 26, 2017 8:41 am

Re: Roth Conversions - McQuarrie study

Post by cas »

curmudgeon wrote: Fri Jun 18, 2021 11:42 pm
The single biggest flaw is that it mentions the "tax torpedo" of taxation of SS benefits, but then proceeds to ignore it (buried in note 3) by skewing the cases examined to be outside the impacted boundaries. Using the example of a joint SS benefit of $100K, retirees with IRA balances between $250K and $2.5 million will be in that tax torpedo zone at initial RMD and subject to considerably higher marginal rates than used in this paper.
As demonstrated rather vividly by the much less affluent, real life random poster who showed up earlier in the thread, asked a specific "Should I convert in my specific situation?" question, and was mildly rebuked by the moderators to not hijack the thread. But FiveK produced this graph for him showing what his marginal rate landscape would be for his RMDs (plus any additional withdrawals beyond RMDs).

Note that, due to the SS "tax hump", plus some small amount of pension and other income, this much-less-affluent-than-paper's-example real life person has a marginal rate on his RMDs that never drops below 22% once he reaches only ~2K in RMDs. And it has a stretch of 40.7% marginal rate on his RMD, starting at about only 17K in RMDs. (The marginal rate spikes further out are IRMAA.)

FiveK wrote: Fri Jun 18, 2021 4:32 pm
hulburt1 wrote: Fri Jun 18, 2021 3:38 pm I'm 68 single with 1.1m Ira and 220000 Roth. I can easily live at 12%. But once SS kicks in and RMD at 72 I will be at top 22%. I live very nice at 25000 a year. Running all my no. (10000 pension, RMD 40000, SS 26000 and a part time job 10000)
Once the $26K SS benefit kicks in, you'll pay at least 24% on the RMD (assuming a minimum $40K RMD), and more than that on higher withdrawals.
Image
This shows why statements like the following from p. 4 of the paper are problematic:
Now turn that around: advisors must recognize, and clients need to be told, that if you will not have peak wage earnings in excess of $200,000 as a couple, and/or circumstances have not allowed you to be a tenacious and prodigious saver, then most likely you are going to be in the 12% tax bracket when you begin to draw RMDs.
12% *nominal* tax bracket, perhaps. But not 12% *marginal rate*, which is the relevant rate for considering Roth conversions.

(Rob and Sue in the article's example actually have the same SS "tax hump" problem. And the same quite-early-on 22+% marginal rates on their RMD. But it is so far in the rear view mirror that, under any Roth conversion/RMD scenario, it is always there, so it can be ignored in the analysis.)
User avatar
jeffyscott
Posts: 13487
Joined: Tue Feb 27, 2007 8:12 am

Re: Roth Conversions - McQuarrie study

Post by jeffyscott »

curmudgeon wrote: Fri Jun 18, 2021 11:42 pm There's a fair bit of hand-waving around the gap between age 65 and 72. The author asserts that his couple has no other savings, but he has them delay SS until age 70, and IRA withdrawal until age 72. What are they living on? Fuzzily implied in there is that one of them can't retire until 72!
Or just doesn't want to retire, near the beginning he seems to make it clear that these are workaholics that he is describing :mrgreen: .
In their peak earning year they had enjoyed wage and salary income totaling about $400,000 in 2019 dollars...one of them has retired "early", at age 65 they find themselves in a lower tax bracket than while working...

I've not read all of the paper as it really is not pertinent to us (aside from confirming that conversions at 12% are highly likely to be beneficial), but it seems to be about doing a single roth conversion?
Should they make a Roth conversion now to reduce their future RMDs

The Hulbert article looks at that sort of scenario, "...in order for such a retiree to take advantage of a Roth conversion at the lowest tax rate, only a small fraction of the retiree’s IRA or 401(k) can be converted".

More often here people talk about doing a series of Roth conversions, such as we are doing (in our case filling up the 12% bracket each year). We're not that affluent couple, but each annual conversion is still only a small percentage of the balance. However, when you repeat this for 10-15 years it adds up. If instead, I were to look at doing one giant Roth conversion in a single year, it would probably not be worth doing as we'd be paying 22-32% tax on a lot of it.
User avatar
RickBoglehead
Posts: 7877
Joined: Wed Feb 14, 2018 8:10 am
Location: In a house

Re: Roth Conversions - McQuarrie study

Post by RickBoglehead »

Haven't read any of it. Sometimes I think people over think things. In a tax bracket of 22/24% now? Convert. Very likely to equal, or be lower, than retirement. We had some years where income was low and we converted. Also Roth 401ks. Now we do pre-tax, then evaluate converting near year-end. Retiring this year with low 80% of retirement Roth, will likely convert 100% before RMDs and have zero RMDs. Also unlikely to tap any retirement accounts for a long time, if at all.
Avid user of forums on variety of interests-financial, home brewing, F-150, EV, home repair, etc. Enjoy learning & passing on knowledge. It's PRINCIPAL, not PRINCIPLE. I ADVISE you to seek ADVICE.
privateID
Posts: 801
Joined: Sat Oct 18, 2014 4:59 pm

Re: Roth Conversions - McQuarrie study

Post by privateID »

Can this conversation also apply to retirement contributions? For someone currently in the 22%, the decision to contribute to a t-ira vs Roth would seem to have similar criteria. If I shouldn't be converting while working to fill up the 22%, then I would think I shouldn't be contributing to a Roth vs a t-ira.
The Stone Wall
Posts: 159
Joined: Thu Mar 22, 2018 1:18 pm

Re: Roth Conversions - McQuarrie study

Post by The Stone Wall »

jeffyscott wrote: Sat Jun 19, 2021 7:01 am
curmudgeon wrote: Fri Jun 18, 2021 11:42 pm There's a fair bit of hand-waving around the gap between age 65 and 72. The author asserts that his couple has no other savings, but he has them delay SS until age 70, and IRA withdrawal until age 72. What are they living on? Fuzzily implied in there is that one of them can't retire until 72!
Or just doesn't want to retire, near the beginning he seems to make it clear that these are workaholics that he is describing :mrgreen: .
In their peak earning year they had enjoyed wage and salary income totaling about $400,000 in 2019 dollars...one of them has retired "early", at age 65 they find themselves in a lower tax bracket than while working...

I've not read all of the paper as it really is not pertinent to us (aside from confirming that conversions at 12% are highly likely to be beneficial), but it seems to be about doing a single roth conversion?
Should they make a Roth conversion now to reduce their future RMDs

The Hulbert article looks at that sort of scenario, "...in order for such a retiree to take advantage of a Roth conversion at the lowest tax rate, only a small fraction of the retiree’s IRA or 401(k) can be converted".

More often here people talk about doing a series of Roth conversions, such as we are doing (in our case filling up the 12% bracket each year). We're not that affluent couple, but each annual conversion is still only a small percentage of the balance. However, when you repeat this for 10-15 years it adds up. If instead, I were to look at doing one giant Roth conversion in a single year, it would probably not be worth doing as we'd be paying 22-32% tax on a lot of it.
I kept expecting the paper to look at multiple Roth conversions, bit it did not. Why would that couple only do one conversion? He admits it is a paltry sum considering the size of the account. However, that one paltry conversion (for them) did make a difference (albeit small). If so, wouldn't multiple conversions between 65 and 70 have a greater effect? Of course, the surviving spouse is going to have all kinds of tax problems because they didn't retire earlier and do more conversions! I am really glad I found the Boglehead site and knew to start thinking ahead.
curmudgeon
Posts: 2630
Joined: Thu Jun 20, 2013 11:00 pm

Re: Roth Conversions - McQuarrie study

Post by curmudgeon »

curmudgeon wrote: Fri Jun 18, 2021 11:42 pm
The single biggest flaw is that it mentions the "tax torpedo" of taxation of SS benefits, but then proceeds to ignore it (buried in note 3) by skewing the cases examined to be outside the impacted boundaries. Using the example of a joint SS benefit of $100K, retirees with IRA balances between $250K and $2.5 million will be in that tax torpedo zone at initial RMD and subject to considerably higher marginal rates than used in this paper.

....

All in all, the premise that "Roth conversions aren't needed/valuable for everyone" (which is quite true) gets badly mangled in this paper. If you are going to stake out one corner of the SS taxaction space, you should be more forthcoming about the rest of the zone.
To be fair, I should add that I think examining the value of Roth conversions for those who don't hit the SS tax hump is a useful zone of consideration. Because the SS benefit tax limits are not indexed for inflation, the hump is effectively moving "down and to the left" over time.
bsteiner
Posts: 9211
Joined: Sat Oct 20, 2012 9:39 pm
Location: NYC/NJ/FL

Re: Roth Conversions - McQuarrie study

Post by bsteiner »

lazynovice wrote: Thu Jun 17, 2021 4:38 pm ... he ... misses the legacy impact of heirs in high tax brackets.

Unless you have pension income, I don’t see a lot of reason to convert beyond the current 12% bracket- future 15%. I’ve thought that for awhile. ...
The beneficiaries may be in higher brackets than before the SECURE Act since the distributions will generally be bunched into 10 or 11 years. That's especially the case for trusts.

Many people only convert to the extent they can do so within the 12% (15% before 2018 and scheduled to resume in 2026).
smitcat
Posts: 13308
Joined: Mon Nov 07, 2016 9:51 am

Re: Roth Conversions - McQuarrie study

Post by smitcat »

bsteiner wrote: Sat Jun 19, 2021 9:05 am
lazynovice wrote: Thu Jun 17, 2021 4:38 pm ... he ... misses the legacy impact of heirs in high tax brackets.

Unless you have pension income, I don’t see a lot of reason to convert beyond the current 12% bracket- future 15%. I’ve thought that for awhile. ...
The beneficiaries may be in higher brackets than before the SECURE Act since the distributions will generally be bunched into 10 or 11 years. That's especially the case for trusts.

Many people only convert to the extent they can do so within the 12% (15% before 2018 and scheduled to resume in 2026).
"The beneficiaries may be in higher brackets than before the SECURE Act since the distributions will generally be bunched into 10 or 11 years. That's especially the case for trusts."
Exactly - one big issue for many of us.
Tdubs
Posts: 1832
Joined: Tue Apr 24, 2018 7:50 pm

Re: Roth Conversions - McQuarrie study

Post by Tdubs »

The Stone Wall wrote: Sat Jun 19, 2021 8:15 am
jeffyscott wrote: Sat Jun 19, 2021 7:01 am
curmudgeon wrote: Fri Jun 18, 2021 11:42 pm There's a fair bit of hand-waving around the gap between age 65 and 72. The author asserts that his couple has no other savings, but he has them delay SS until age 70, and IRA withdrawal until age 72. What are they living on? Fuzzily implied in there is that one of them can't retire until 72!
Or just doesn't want to retire, near the beginning he seems to make it clear that these are workaholics that he is describing :mrgreen: .
In their peak earning year they had enjoyed wage and salary income totaling about $400,000 in 2019 dollars...one of them has retired "early", at age 65 they find themselves in a lower tax bracket than while working...

I've not read all of the paper as it really is not pertinent to us (aside from confirming that conversions at 12% are highly likely to be beneficial), but it seems to be about doing a single roth conversion?
Should they make a Roth conversion now to reduce their future RMDs

The Hulbert article looks at that sort of scenario, "...in order for such a retiree to take advantage of a Roth conversion at the lowest tax rate, only a small fraction of the retiree’s IRA or 401(k) can be converted".

More often here people talk about doing a series of Roth conversions, such as we are doing (in our case filling up the 12% bracket each year). We're not that affluent couple, but each annual conversion is still only a small percentage of the balance. However, when you repeat this for 10-15 years it adds up. If instead, I were to look at doing one giant Roth conversion in a single year, it would probably not be worth doing as we'd be paying 22-32% tax on a lot of it.
I kept expecting the paper to look at multiple Roth conversions, bit it did not. Why would that couple only do one conversion? He admits it is a paltry sum considering the size of the account. However, that one paltry conversion (for them) did make a difference (albeit small). If so, wouldn't multiple conversions between 65 and 70 have a greater effect? Of course, the surviving spouse is going to have all kinds of tax problems because they didn't retire earlier and do more conversions! I am really glad I found the Boglehead site and knew to start thinking ahead.
To support the basic point of the paper--that Roth conversions pay off, though slowly and unspectacularly--tracking just one conversion is fine. The best bang for the buck comes with the first one. Rob and Sue will be even older when later conversions finally hit the breakeven point.
bsteiner
Posts: 9211
Joined: Sat Oct 20, 2012 9:39 pm
Location: NYC/NJ/FL

Re: Roth Conversions - McQuarrie study

Post by bsteiner »

cbeck wrote: Thu Jun 17, 2021 7:11 pm
lazynovice wrote: Thu Jun 17, 2021 4:00 pm ...

One finding is that if you withdraw the money from the Roth before a certain break even period, you didn’t get any benefit. The break even point is 13 to 15 years best case.
Which points out a problem that I have noticed in discussions of Roth. The value of the Roth depends on the expected lifespan of the Roth, not the owner. I expect my Roth to be feeding my younger wife forty years from now.

Another point that I don't see addressed in most of these discussions is the benefit from paying for the Roth conversion with after-tax funds, thereby enabling us to replace the portion of the TIRA which belongs to the govt with our own money. Also, the estate planning benefit of assets exempt from RMDs for the surviving spouse is usually overlooked.

Having done my conversions between the ages of 61 and 70 when my taxable income was close to zero and having a (much) younger wife, I don't see how having most of my assets in the Roth could fail to pay off.
cbeck is correct. He mentions breakeven point many times. But that's not a relevant concept. The proper comparison is between (i) how much money there would be after all taxes on the day the beneficiaries have to withdraw the last dollar from the IRA if you convert, and (ii) how much money there would be after all taxes on the day the beneficiaries have to withdraw the last dollar from the IRA if you don't convert (or if you convert some specified amounts).

cbeck is also correct that whether you pay the tax on the conversion out of other money is a significant factor. That analysis is the same for contributions as for conversions.
Last edited by bsteiner on Sat Jun 19, 2021 7:37 pm, edited 1 time in total.
bsteiner
Posts: 9211
Joined: Sat Oct 20, 2012 9:39 pm
Location: NYC/NJ/FL

Re: Roth Conversions - McQuarrie study

Post by bsteiner »

McQ wrote: Fri Jun 18, 2021 1:04 am Bogleheads: Edward McQuarrie here, happy to answer questions directly. An example of the spreadsheet I used to compare outcomes can be downloaded here: http://www.edwardfmcquarrie.com/?p=573. If you don't think my conclusions apply to your particular case, the spreadsheet allows you to confirm that mismatch. But if you haven't set up a spreadsheet to evaluate your Roth conversion, then you may have fallen prey to what my behavioral finance colleagues call "misleading heuristics." It took me about six attempts before I had a spreadsheet that passed muster Early attempts failed to maintain an exact counterfactual to the conversion. Try it; but you'll need the paper appendix open on a large screen monitor also.
Next, I want to reiterate a key conclusion: "A Roth conversion, within the parameters examined, always pays off, but always slowly, and always in small amounts." The exception is conversions in the 0% bracket and neighboring brackets, which pay off early and sizably.
On reviewing the discussion thus far, let me acknowledge the following limitations:
1. Indeed, I did not examine the situation of the widow/er subject to future single tax rates. On tap for the revision. FYI: 90% of the taxpayers who were in my target income range are married filing joint, per the IRS SOI.
2. Bequests: in another revision, I will emphasize that conversions are particularly suitable when you have no need for some of the money in traditional accounts. I mean no need at all, true surplus. Rationale: a bequest adds ten years to your life expectancy (SECURE act), and the longer the compounding period, the more successful the Roth conversion. But, start to tap that conversion annually in your 80s to supplement income, and the payoff blows up (=it wasn't really surplus).
3. It is a long paper, it has footnotes, and footnotes to the tables, and appendices, and appendix tables, and footnotes to the appendix tables. Sorry; the topic is complicated and I am a life-long academic. I ask you not to dismiss what you have not read.
PS: thank you, lazynovice
It is late here in California; replies will begin by about noon Pacific time Friday.
Thanks for participating here.

There are so many variables that it's difficult to create a spreadsheet.

Many people convert up to the top of the 12% (15% before 2018 and scheduled to return in 2026) bracket. They view the risk (of withdrawals or conversions in later years at lower rates) as being modest. The amount of tax they have to pay on the conversion is likewise modest.

Some people who don't expect that they or their beneficiaries are likely to be in a bracket less than 22% convert up to the top of the 24% bracket. That takes into account the large gap between the 24% bracket and the 32% bracket. Given the widening of more of the joint return brackets to double the width of the single brackets for 2018-2025 (the 24% bracket goes up to $329,850 of taxable income in 2021), that allows for substantial conversions at 24% or less.
User avatar
David Jay
Posts: 14587
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: Roth Conversions - McQuarrie study

Post by David Jay »

Welcome to the forum, this is a “hot” topic right now, as we are seeing the first big tranche of individuals with defined contribution plans reaching retirement age.
McQ wrote: Fri Jun 18, 2021 1:04 am1. Indeed, I did not examine the situation of the widow/er subject to future single tax rates. On tap for the revision. FYI: 90% of the taxpayers who were in my target income range are married filing joint, per the IRS SOI.
I look forward to the revision. As you say, 90% of the demographic may currently be married but virtually 100% of married couples will face some period where the survivor is “filing single”.
Last edited by David Jay on Sat Jun 19, 2021 10:05 am, edited 1 time in total.
It's not an engineering problem - Hersh Shefrin | To get the "risk premium", you really do have to take the risk - nisiprius
ChrisC
Posts: 1475
Joined: Tue Jun 19, 2012 9:10 am
Location: North Carolina

Re: Roth Conversions - McQuarrie study

Post by ChrisC »

Exchme wrote: Fri Jun 18, 2021 8:53 pm Prof McQ:

I love that you took the time to produce the paper, it's great having some academic interest in a subject of practical importance to millions of people. Seems like most writing on such topics these days comes from people selling something.

I like the precision of referring to inter temporal tax arbitrage part of a Roth specifically. But not sure I understand the dismissal of loading the Roth with stocks while slowing tax deferred with bonds. Are you saying that it is incorrect to do that or that Roth advocates do careless/opportunistic math (bottom of p10/top of p11)?

It does seem like bonds in TDA and stocks in Roth is good advice and doesn't holding two different types of accounts give us the chance to get a boost from this? If there is a boost, is it really separable from the inter temporal tax arbitrage benefit?

Thanks!
I've never understood the advice of loading up the TDA with bonds to slow down the growth of that account to lessen the impact of RMDs; seemed like this advice was partly driven by asset allocation goals and that it was purportedly optimal not to grow your TDA so that you would decrease RMDs -- sounded nonsensical to me, especially if you already dealt with asset allocation constructs in your entire investment portolio, including taxable accounts. For instance, I maintain 80/20 equity to fixed income allocation in our Roth accounts and 84/16 equity to fixed income allocations in our TDA. But I'm 100% in cash in my taxable account, which is entirely in savings accounts, CDs and money market funds, and this cash represents 10% of my investment portfolio, excluding illiquid real estate. (When I do convert from my TDA, I'm taking primarily from the equity portion of the balance in this account so there's that for slowing down the growth -- I admit that but it's more like locking in gains to me.)

So why should I slow down the growth of my TDA to avoid greater RMD and tax exposure. Admittedly, I'm fighting a losing battle in that my conversions don't keep pace with growth in my TDA, but this is a battle I really don't have to win or that I need to make battlefield adjustments to go heavily into bonds in my TDA.
bsteiner
Posts: 9211
Joined: Sat Oct 20, 2012 9:39 pm
Location: NYC/NJ/FL

Re: Roth Conversions - McQuarrie study

Post by bsteiner »

sc9182 wrote: Fri Jun 18, 2021 11:07 am ... if one is single/widowed, retired living in CA/NY as oft mentioned ...
There's another important difference between the two. New York has a state estate tax with an exclusion amount lower than the Federal (it's equal to what the Federal exclusion amount would have been but for the Tax Cuts and Jobs Act of 2017, presently $5,930,000). California has no state estate tax.

Section 691(c) allows an income tax deduction for the Federal (but not the state) estate tax, at the marginal rate. That fully eliminates the double tax in states like California that don't have a state estate tax, but not for people in states like New York that have a state estate or inheritance tax.
User avatar
David Jay
Posts: 14587
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: Roth Conversions - McQuarrie study

Post by David Jay »

ChrisC wrote: Sat Jun 19, 2021 10:03 amI've never understood the advice of loading up the TDA with bonds to slow down the growth of that account to lessen the impact of RMDs; seemed like this advice was partly driven by asset allocation goals and that it was purportedly optimal not to grow your TDA so that you would decrease RMDs -- sounded nonsensical to me, especially if you already dealt with asset allocation constructs in your entire investment portolio, including taxable accounts. For instance, I maintain 80/20 equity to fixed income allocation in our Roth accounts and 84/16 equity to fixed income allocations in our TDA. But I'm 100% in cash in my taxable account, which is entirely in savings accounts, CDs and money market funds, and this cash represents 10% of my investment portfolio, excluding illiquid real estate. (When I do convert from my TDA, I'm taking primarily from the equity portion of the balance in this account so there's that for slowing down the growth -- I admit that but it's more like locking in gains to me.)

So why should I slow down the growth of my TDA to avoid greater RMD and tax exposure. Admittedly, I'm fighting a losing battle in that my conversions don't keep pace with growth in my TDA, but this is a battle I really don't have to win or that I need to make battlefield adjustments to go heavily into bonds in my TDA.
I am not comfortable with your “loading up” language. I have not seen any recommendation to change your overall asset allocation to reduce taxes.

Since one has a certain percentage of their portfolio in stocks and a certain percentage in fixed income to meet their personal growth-vs-volatility needs, why would one put any of the low-yielding fixed income in Roth? The math makes no sense to me.
It's not an engineering problem - Hersh Shefrin | To get the "risk premium", you really do have to take the risk - nisiprius
bsteiner
Posts: 9211
Joined: Sat Oct 20, 2012 9:39 pm
Location: NYC/NJ/FL

Re: Roth Conversions - McQuarrie study

Post by bsteiner »

McQ wrote: Fri Jun 18, 2021 1:24 pm ... it does not seem reasonable, in the general case, that anyone other than a Boglehead would have both maxed out their 401(k) and piled up hundreds of thousands of dollars in taxable funds + saved kid's college + lived in a nice house (e.g., 30% savings rate, where I assumed 15% was the max reasonable). Nonetheless, if you do have years of living expenses in your taxable account, then you can convert at 0%/10%/12%, year after year. That was absolutely the best Roth conversion outcome I could find among all examined. But don't convert so much that age 72 income falls out of the 22% bracket.
You don't have to be a super saver to have a 6-figure taxable account (where the first digit isn't a 1). There's a limit as to how much you can put into a 401(k) plan (presently $19,500, or $26,000 after age 50).

I agree that you generally shouldn't convert at 22% if you' expect to be in the 12% bracket later on.
bsteiner
Posts: 9211
Joined: Sat Oct 20, 2012 9:39 pm
Location: NYC/NJ/FL

Re: Roth Conversions - McQuarrie study

Post by bsteiner »

sc9182 wrote: Fri Jun 18, 2021 2:39 pm ...
Someone is one life event away, or one/two divorces away, big medical condition, or disability, or GFC type 50% or worse crash or two away — from falling significantly in tax bracket(s) or amount of tax paid. We all hope/plan, but markets/life deliver. Even if we imagine 22% possible future tax bracket — it may not happen thus so. So, pre-paying at 22% and locking the potential savings may not materialize.
...
On the other hand, things may go better than expected and you may unexpectedly end up in a bracket higher than 22%.

Note that now only are the rates scheduled to revert to pre-2018 levels in 2026, but the doubling of the width of the present 22%, 24% and 32% brackets on a joint return to twice the width of the single brackets is scheduled to expire at the end of 2025.
ChrisC
Posts: 1475
Joined: Tue Jun 19, 2012 9:10 am
Location: North Carolina

Re: Roth Conversions - McQuarrie study

Post by ChrisC »

David Jay wrote: Sat Jun 19, 2021 10:11 am
ChrisC wrote: Sat Jun 19, 2021 10:03 amI've never understood the advice of loading up the TDA with bonds to slow down the growth of that account to lessen the impact of RMDs; seemed like this advice was partly driven by asset allocation goals and that it was purportedly optimal not to grow your TDA so that you would decrease RMDs -- sounded nonsensical to me, especially if you already dealt with asset allocation constructs in your entire investment portolio, including taxable accounts. For instance, I maintain 80/20 equity to fixed income allocation in our Roth accounts and 84/16 equity to fixed income allocations in our TDA. But I'm 100% in cash in my taxable account, which is entirely in savings accounts, CDs and money market funds, and this cash represents 10% of my investment portfolio, excluding illiquid real estate. (When I do convert from my TDA, I'm taking primarily from the equity portion of the balance in this account so there's that for slowing down the growth -- I admit that but it's more like locking in gains to me.)

So why should I slow down the growth of my TDA to avoid greater RMD and tax exposure. Admittedly, I'm fighting a losing battle in that my conversions don't keep pace with growth in my TDA, but this is a battle I really don't have to win or that I need to make battlefield adjustments to go heavily into bonds in my TDA.
I am not comfortable with your “loading up” language. I have not seen any recommendation to change your overall asset allocation to reduce taxes.

Since one has a certain percentage of their portfolio in stocks and a certain percentage in fixed income to meet their personal growth-vs-volatility needs, why would one put any of the low-yielding fixed income in Roth? The math makes no sense to me.
Well, there's one recommendation in this thread:viewtopic.php?f=2&t=350940&p=6059316#p6059316 The point I was making was that there are recommendations to load up your TDA with bonds and that is ostensibly done to slow the growth of that account to lessen the RMD hit, which would in turn, lessen tax exposure or permit greater Roth conversions from the TDA. Have I misrepresented that advice?

The math might make no sense to you, but some of us are afflicted by behavorial issues that don't permit us to go 100% equities in Roth. Given my high pension income stream with a COLA adjusted Federal pension that takes care of all living expenses, I should not be in any low-yielding fixed income in Roth, TDA or taxable accounts -- I should be letting it all ride in equities, right?
User avatar
retiredjg
Posts: 54082
Joined: Thu Jan 10, 2008 11:56 am

Re: Roth Conversions - McQuarrie study

Post by retiredjg »

ChrisC wrote: Sat Jun 19, 2021 10:42 am
David Jay wrote: Sat Jun 19, 2021 10:11 am
ChrisC wrote: Sat Jun 19, 2021 10:03 amI've never understood the advice of loading up the TDA with bonds to slow down the growth of that account to lessen the impact of RMDs; seemed like this advice was partly driven by asset allocation goals and that it was purportedly optimal not to grow your TDA so that you would decrease RMDs -- sounded nonsensical to me, especially if you already dealt with asset allocation constructs in your entire investment portolio, including taxable accounts. For instance, I maintain 80/20 equity to fixed income allocation in our Roth accounts and 84/16 equity to fixed income allocations in our TDA. But I'm 100% in cash in my taxable account, which is entirely in savings accounts, CDs and money market funds, and this cash represents 10% of my investment portfolio, excluding illiquid real estate. (When I do convert from my TDA, I'm taking primarily from the equity portion of the balance in this account so there's that for slowing down the growth -- I admit that but it's more like locking in gains to me.)

So why should I slow down the growth of my TDA to avoid greater RMD and tax exposure. Admittedly, I'm fighting a losing battle in that my conversions don't keep pace with growth in my TDA, but this is a battle I really don't have to win or that I need to make battlefield adjustments to go heavily into bonds in my TDA.
I am not comfortable with your “loading up” language. I have not seen any recommendation to change your overall asset allocation to reduce taxes.

Since one has a certain percentage of their portfolio in stocks and a certain percentage in fixed income to meet their personal growth-vs-volatility needs, why would one put any of the low-yielding fixed income in Roth? The math makes no sense to me.
Well, there's one recommendation in this thread:viewtopic.php?f=2&t=350940&p=6059316#p6059316 The point I was making was that there are recommendations to load up your TDA with bonds and that is ostensibly done to slow the growth of that account to lessen the RMD hit, which would in turn, lessen tax exposure or permit greater Roth conversions from the TDA. Have I misrepresented that advice?
You have not misrepresented that advice. I did say that in that one thread specifically for that one situation. It is not something I suggest on any regular basis.

In other words, David Jay is correct that usually the advice is to put what bonds you have decided to own in the tax-deferred accounts as a measure to slow down growth of that account without slowing down the growth of the portfolio as a whole.
Angst
Posts: 2968
Joined: Sat Jun 09, 2007 11:31 am

Re: Roth Conversions - McQuarrie study

Post by Angst »

RickBoglehead wrote: Sat Jun 19, 2021 7:22 am Haven't read any of it. Sometimes I think people over think things. In a tax bracket of 22/24% now? Convert. Very likely to equal, or be lower, than retirement. We had some years where income was low and we converted. Also Roth 401ks. Now we do pre-tax, then evaluate converting near year-end. Retiring this year with low 80% of retirement Roth, will likely convert 100% before RMDs and have zero RMDs. Also unlikely to tap any retirement accounts for a long time, if at all.
For those who might live in a municipality that levies (e.g.) a 1% "earnings tax", contributing to one's trad 401k and converting later to a Roth can allow one to entirely avoid ever paying the 1% local tax on that income. Our city does not levy the tax on payroll income that goes straight into one's trad 401k, nor is it charged on 401k or IRA redemptions or conversions.
User avatar
David Jay
Posts: 14587
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: Roth Conversions - McQuarrie study

Post by David Jay »

ChrisC wrote: Sat Jun 19, 2021 10:42 amGiven my high pension income stream with a COLA adjusted Federal pension that takes care of all living expenses, I should not be in any low-yielding fixed income in Roth, TDA or taxable accounts -- I should be letting it all ride in equities, right?
Absolutely not. AA comes first. Then portfolio. Then account distribution.

AA is intensely personal, it is imperative that one not overestimate their ability to handle risk, it is so important that one not sell after a large downturn. I always use Bernstein’s quote: “mistiming the market is probably the most frequent and severe form of permanent capital loss”.

That said, I think it is a very minor behavioral issue to adjust AA by individual account and not by total portfolio.
It's not an engineering problem - Hersh Shefrin | To get the "risk premium", you really do have to take the risk - nisiprius
McQ
Posts: 1425
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

ChrisC wrote: Fri Jun 18, 2021 9:53 pm
McQ wrote: Fri Jun 18, 2021 6:10 pm Higher pension, a 457 as well as a 401k balance per an earlier comment (civil servant case), etc. make it more likely that you will be high up in the 22% bracket, which may mean IRMAA, which will make a conversion today at 22% relatively more attractive. BTW, when I set a $1.8 million TDA balance as the threshold for entering the 22% bracket seven years hence, I assumed $70,000 social security + pension for MFJ. If that is only $50,000, then you will need $2.3 million to nose into the 22% bracket.
Do you have TDA balance threshold for entering the 24%? A number of civil servant/military annuitants/pensioners easily enter the 22% bracket in retirement at an early retirement age, and some can, or have a spouse that can, draw SS retirement benefits that place them as a couple squarely within the 24% bracket. We've been converting into the higher range of the 24% the past several years, with the expectation that (1) we will never be below that bracket when SS retirement benefits kick-in for my spouse (which will occur later this year); (2) we will never touch Roth or TDA balances for living expenses (and can pay Federal and state income taxes from pension draws and taxable accounts); (3) all of our retirements accounts in TSP, 401Ks or Roth IRAs will be inherited by our children who are in tax brackets higher than their parents.

My gut tells me that we should still be converting into the 24% bracket and perhaps even higher pre-and post RMDS. If one of us passes, the survivor will wind up in the 32% bracket even if some of the pension and SS retirement benefits decrease. Also, some former civil servants have the ability to avoid IRMAA altogether by not enrolling in Medicare Part B or Part D and by relying on Government employer sponsored retiree health coverage (FEHB).

Edited: So, after posting the above, I read this note to your Table 2 regarding TDA threshold balance for entering into the 24% bracket:

"Pension income. If the client is a public sector employee or other individual who expects a pension as well as social security,
TDA balances would be lower than in the table. Each $10,000 in pension payments would reduce the required TDA balance to
enter a bracket by $273,000. Returning to Rob and Sue, if one of them had a pension of $50,000, then instead of needing
$4.175 million to enter the 24% tax bracket, only about $2.8 million would be required."

Perhaps, I'm misunderstanding the above note but if my pension/social security income is $174,000, then the $4.175 million in the threshold balance you postulated for TDA would be reduced by 17.4 x $273,000, which results in -$575,200.
I see now the table note would be more clear if I had written: "each $10,000 *in excess of* the value in the table" which would be $90,000 here. So in your case, (174,000 - 90,000) = $84,000, or 8.4 * 273,000, about $2.3 million to be subtracted from $4.175 million (warning: table values for 2027; increase by 1.025 for each additional year before you turn 72). So a pensioner like yourself hits the 24% bracket with a TDA balance of $1.8 million. If you convert so much that you have less than $1.8 million, then you convert now at 24% to save 22%, and breakeven will be further into your 90s. But you will breakeven, especially adding 10 years to the last to die (=your bequest).
I debated adding a section on those with significant pensions and removed for length reasons. Might have to revisit that.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
McQ
Posts: 1425
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

privateID wrote: Sat Jun 19, 2021 7:55 am Can this conversation also apply to retirement contributions? For someone currently in the 22%, the decision to contribute to a t-ira vs Roth would seem to have similar criteria. If I shouldn't be converting while working to fill up the 22%, then I would think I shouldn't be contributing to a Roth vs a t-ira.
Appendix alpha in the paper considers the contributions case, but contributions were not the focus, since they have to be modeled over 60 - 70 years. :shock: Generally, the same rule applies: if you wouldn't convert at 22%, you probably should not give up the 22% government subsidy for traditional 401(k)s. But contributions may have many decades before tapped, and since Roth accounts always prevail when the compounding period is long enough, the decision calculus is somewhat different for contributions.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Stuckinmn
Posts: 35
Joined: Mon May 13, 2019 5:11 pm

Re: Roth Conversions - McQuarrie study

Post by Stuckinmn »

As someone a couple of years from early retirement in my 50s I've been considering this subject. To me, my gut says up to 12% bracket is a no brainer but 22% is probably not worth it in my situation.

However, taxable accounts throw a new wrinkle into it. Specifically, if I keep my qualified dividends and cap gains in the 12% bracket I pay zero on those and save 15%. So to me it makes sense to convert the amount that will take you to top of the 12% bracket after including cap gains and divs for the year. Yes, the 15% you'd pay in cap gains if you go over the 12% bracket is probably lower than the marginal RMD rate you'll eventually pay, but that's a 15% gain on taxes (going from 15 long term cap gains to zero) versus a possible 10% gain (going from 22% to 12%).

Not sure if that makes sense but it's basically take the bird in the hand and pay zero percent on long term caps while you liquidate for living expenses, then fill up the rest of the 12% bracket with conversions.
Exchme
Posts: 1335
Joined: Sun Sep 06, 2020 3:00 pm

Re: Roth Conversions - McQuarrie study

Post by Exchme »

curmudgeon wrote: Fri Jun 18, 2021 11:42 pm It appears that the "end wealth", used to calculate his "breakeven age", values tIRA accounts and Roth as equivalent. I can say without question that your heirs, unless they are a charity, will take a very different view of the matter.
The author acknowledges the shortcoming of net worth in Appendix alpha-4 and makes an approximate attempt at correction to the actual utility of the different accounts by applying an assumed tax rate to the TDA, so the author is clearly aware of the math. I agree with "curmudgeon" that using net worth sets a confusing frame for the reader. Graphs with negative numbers regarding money are powerful persuasion tools and may convince people not to do Roths if they feel their lifespan is short when I don't think that is the author's intended message.

To hammer the point further, if you did a Roth conversion and then immediately died, the value of your estate to your heirs is not lower than if you did no conversion, at least in the case where your heirs have the same average tax burden on inherited IRA withdrawals as you had on the Roth conversion. So when measured against actual utility of the money, I don't see that there was ever a reduction of spendable money, so no "investment" in the traditional sense, so breakeven ages and ROIs are odd measures.

My suggestion to compare cases is to get everything equalized. I did a pair of runs in Retiree Portfolio Model which keeps track of each type of account. To be the simplest base case I could think of, I simply extended my assumed lifetime by 10 years and evenly withdrew the remaining funds from the TDA during that final 10 years to simulate the action of heirs. Then I simply compared the ending portfolio value of the Roth vs. no Roth case with all accounts converted to taxable. If I would rather have the reference point be day of death instead of 10 years on, I could use a present value formula, using the assumed rate of return as the discount rate. Obviously there are lots of permutations with regard spousal lifespans, heirs with different incomes, etc. but for the simplest case, that seems to work and provides a way to compare everything on an equal footing.
mouth
Posts: 654
Joined: Sun Apr 19, 2015 6:40 am

Re: Roth Conversions - McQuarrie study

Post by mouth »

McQ wrote: Sat Jun 19, 2021 12:34 pm I debated adding a section on those with significant pensions and removed for length reasons. Might have to revisit that.
Yes please.

Single working (32% fed bracket right now) 48yo Military retiree with $46k COLA'd pension looking to retire at 55 if not 50. Just as an one example of your target audience :)
Big Dog
Posts: 4608
Joined: Mon Sep 07, 2015 4:12 pm

Re: Roth Conversions - McQuarrie study

Post by Big Dog »

smitcat wrote: Sat Jun 19, 2021 9:10 am
bsteiner wrote: Sat Jun 19, 2021 9:05 am
lazynovice wrote: Thu Jun 17, 2021 4:38 pm ... he ... misses the legacy impact of heirs in high tax brackets.

Unless you have pension income, I don’t see a lot of reason to convert beyond the current 12% bracket- future 15%. I’ve thought that for awhile. ...
The beneficiaries may be in higher brackets than before the SECURE Act since the distributions will generally be bunched into 10 or 11 years. That's especially the case for trusts.

Many people only convert to the extent they can do so within the 12% (15% before 2018 and scheduled to resume in 2026).
"The beneficiaries may be in higher brackets than before the SECURE Act since the distributions will generally be bunched into 10 or 11 years. That's especially the case for trusts."
Exactly - one big issue for many of us.
Not a trust, but my kids are currently in a bracket higher than 22%, and I expect them to remain, so my conversions are really for passing the most after-tax wealth to them. (As I started Roths much later than I should have, I shouldn't need those dollars to live on.)
TheDogFather
Posts: 200
Joined: Sun Jul 14, 2019 11:02 am
Location: Virginia

Re: Roth Conversions - McQuarrie study

Post by TheDogFather »

With dividends from taxable, pensions, SS (taken at 70) and RMDs (even with minimal growth in tax-advantaged accounts) I will be in the 24% bracket and so it seems to make sense to do all the Roth conversions I can at 22% from the time I retire until I am required to take RMDs.
ChrisC
Posts: 1475
Joined: Tue Jun 19, 2012 9:10 am
Location: North Carolina

Re: Roth Conversions - McQuarrie study

Post by ChrisC »

McQ wrote: Sat Jun 19, 2021 12:34 pm
I see now the table note would be more clear if I had written: "each $10,000 *in excess of* the value in the table" which would be $90,000 here. So in your case, (174,000 - 90,000) = $84,000, or 8.4 * 273,000, about $2.3 million to be subtracted from $4.175 million (warning: table values for 2027; increase by 1.025 for each additional year before you turn 72). So a pensioner like yourself hits the 24% bracket with a TDA balance of $1.8 million. If you convert so much that you have less than $1.8 million, then you convert now at 24% to save 22%, and breakeven will be further into your 90s. But you will breakeven, especially adding 10 years to the last to die (=your bequest).
I debated adding a section on those with significant pensions and removed for length reasons. Might have to revisit that.
Thanks for reaching back to us on this issue. I do hope you add a section there, though perhaps this a diminishing population. I hate to sidetrack the general tenor of the discussions here, but I'm wondering about the $90,000 assumed social security retirement benefits, when coupled with high pensions. In my situation, my pension alone brings us above the top half of the 22% bracket, which would also increase by 10% if my spouse predeceased me. I get a modest social security retirement benefit because of WEP, and no spousal or survivor benefits from my wife's modest social security benefits because of GPO. The $90,000 figure, in my situation, is almost 4 times the actual social security benefits we would receive as a couple. How do you account for assumed social security benefits that are grossly over-stated in individual cases?

One other point I'd like to make is about the IRMAA references in Table 2 for those in upper IRMAA tiers: do you reduce the TDA balance that corresponds to the IRMAA tier level by the same high pension factors? In other words, if we're in IRMAA tier 3 with a TDA threshold balance of $6,561,000 do we reduce that by $2.3 million in my case? And what about the situation where only one of us is subject to IRMAA (as I didn't enroll in Medicare Part B or D but my spouse did) or where we can easily adjust IRMAA tiers by doing lower conversion amounts or dis-enrolling from Medicare Part B?
milktoast
Posts: 626
Joined: Wed Jul 10, 2019 8:17 pm

Re: Roth Conversions - McQuarrie study

Post by milktoast »

One thing I don’t understand is converting deep into 24% bracket.

Don’t you slam into several interesting thresholds? For MFJ the 24% bracket starts at roughly $171k + 24k = $195k MAGI

But for those on Medicare you hit IRMAA thresholds at $176k, $222k, and $276k MAGI.

And if you have taxable accounts (which I think you must to convert that much) the 3.8% NIIT hits at $250k.

So you end up having an effective 27.8% marginal rate in there along with short bursts in marginal rates due to IRMAA.

Are people considering this when thinking about going beyond $176k in MAGI after age 63?

The other question I have, and maybe this is only due to my planning for high spending, is why convert rather than just spending it? Is it better to convert and sell in taxable to cover spending or just to spend it outright.

Guess I’ll have to make a spreadsheet. The pre-made ones overwhelm me.
smitcat
Posts: 13308
Joined: Mon Nov 07, 2016 9:51 am

Re: Roth Conversions - McQuarrie study

Post by smitcat »

Big Dog wrote: Sat Jun 19, 2021 1:17 pm
smitcat wrote: Sat Jun 19, 2021 9:10 am
bsteiner wrote: Sat Jun 19, 2021 9:05 am
lazynovice wrote: Thu Jun 17, 2021 4:38 pm ... he ... misses the legacy impact of heirs in high tax brackets.

Unless you have pension income, I don’t see a lot of reason to convert beyond the current 12% bracket- future 15%. I’ve thought that for awhile. ...
The beneficiaries may be in higher brackets than before the SECURE Act since the distributions will generally be bunched into 10 or 11 years. That's especially the case for trusts.

Many people only convert to the extent they can do so within the 12% (15% before 2018 and scheduled to resume in 2026).
"The beneficiaries may be in higher brackets than before the SECURE Act since the distributions will generally be bunched into 10 or 11 years. That's especially the case for trusts."
Exactly - one big issue for many of us.
Not a trust, but my kids are currently in a bracket higher than 22%, and I expect them to remain, so my conversions are really for passing the most after-tax wealth to them. (As I started Roths much later than I should have, I shouldn't need those dollars to live on.)
Exactly our case - good luck to you.
User avatar
jeffyscott
Posts: 13487
Joined: Tue Feb 27, 2007 8:12 am

Re: Roth Conversions - McQuarrie study

Post by jeffyscott »

lazynovice wrote: Fri Jun 18, 2021 3:19 pm
FiveK wrote: Fri Jun 18, 2021 2:57 pm
lazynovice wrote: Fri Jun 18, 2021 2:50 pm
FiveK wrote: Fri Jun 18, 2021 2:23 pm
lazynovice wrote: Fri Jun 18, 2021 2:03 pmFor whom does it make sense to convert at 22%?
For starters, anyone who expects future distributions would be taxed at marginal rates higher than 22%.
Not exactly per the study? Look at Table 7. Converting at 22% at a 22% rate during RMDs the money has to be untouched for 17 years- case #2. So the advice should be amended to “if you can leave it in the Roth for quite awhile post conversion.”
If I convert $1000 to Roth and pay 22% tax from the conversion, that leaves $780. If that increases in value by 10% in one year, I have $858 to spend.
If I wait one year and that $1000 grows to $1100, and I convert or withdraw at 24%, I have $836 to spend.

No need to wait 17 years.
I’d like to see McQ’s answer to that. I think he is telling us something more nuanced than that.Not that I disagree with your math, but the rate of growth versus the tax rate change are both factors he looked at. And the conclusion was that the outcome had different sensitivities to both.
I'd like to understand why this, tax rate in vs. tax rate out, is not all that matters. I or someone is going to pay tax on the money in a TDA. Why isn't doing so at the lowest possible rate the correct answer?

Of course, there is uncertainty that may make converting at 12%, when expecting a future rate of 22%, a better bet than doing so at 22%, when expecting a future rate of 24%. But I don't think that is what the paper is about :?: .
McQ
Posts: 1425
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

ChrisC wrote: Sat Jun 19, 2021 2:03 pm
McQ wrote: Sat Jun 19, 2021 12:34 pm
I see now the table note would be more clear if I had written: "each $10,000 *in excess of* the value in the table" which would be $90,000 here. So in your case, (174,000 - 90,000) = $84,000, or 8.4 * 273,000, about $2.3 million to be subtracted from $4.175 million (warning: table values for 2027; increase by 1.025 for each additional year before you turn 72). So a pensioner like yourself hits the 24% bracket with a TDA balance of $1.8 million. If you convert so much that you have less than $1.8 million, then you convert now at 24% to save 22%, and breakeven will be further into your 90s. But you will breakeven, especially adding 10 years to the last to die (=your bequest).
I debated adding a section on those with significant pensions and removed for length reasons. Might have to revisit that.
Thanks for reaching back to us on this issue. I do hope you add a section there, though perhaps this a diminishing population. I hate to sidetrack the general tenor of the discussions here, but I'm wondering about the $90,000 assumed social security retirement benefits, when coupled with high pensions. In my situation, my pension alone brings us above the top half of the 22% bracket, which would also increase by 10% if my spouse predeceased me. I get a modest social security retirement benefit because of WEP, and no spousal or survivor benefits from my wife's modest social security benefits because of GPO. The $90,000 figure, in my situation, is almost 4 times the actual social security benefits we would receive as a couple. How do you account for assumed social security benefits that are grossly over-stated in individual cases?

One other point I'd like to make is about the IRMAA references in Table 2 for those in upper IRMAA tiers: do you reduce the TDA balance that corresponds to the IRMAA tier level by the same high pension factors? In other words, if we're in IRMAA tier 3 with a TDA threshold balance of $6,561,000 do we reduce that by $2.3 million in my case? And what about the situation where only one of us is subject to IRMAA (as I didn't enroll in Medicare Part B or D but my spouse did) or where we can easily adjust IRMAA tiers by doing lower conversion amounts or dis-enrolling from Medicare Part B?
Yes, you would hit the stated IRMAA thresholds at $2.3 million less than the stated values. If only one of you is subject to IRMAA, the dollar bite is less and the benefit of conversion is less (but maybe not by much); but the threshold doesn't change.
A word about the $90,000 assumed social security + pension benefit: the adjustment factor works in either direction. You expect more than $90,000 so your TDA threshold went down; someone who expects less would have a threshold that moves up. I'll be the first to admit that $90,000 taxable, from social security alone, even deferred to age 70, and seven years hence, is high. But I had a specific rhetorical purpose in constructing the table. It drives me nuts when journalists write "if you'll be in a higher tax bracket when retired, you should consider converting." What does it take to be even in the 24% bracket? Millions. I had to spot the skeptical reader the max possible social security to make the table entries as persuasive as possible. Thresholds would be about $2.5 million higher if for some reason there was no SS and no pension and no other income.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
McQ
Posts: 1425
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

Exchme wrote: Sat Jun 19, 2021 12:54 pm
curmudgeon wrote: Fri Jun 18, 2021 11:42 pm It appears that the "end wealth", used to calculate his "breakeven age", values tIRA accounts and Roth as equivalent. I can say without question that your heirs, unless they are a charity, will take a very different view of the matter.
The author acknowledges the shortcoming of net worth in Appendix alpha-4 and makes an approximate attempt at correction to the actual utility of the different accounts by applying an assumed tax rate to the TDA, so the author is clearly aware of the math. I agree with "curmudgeon" that using net worth sets a confusing frame for the reader. Graphs with negative numbers regarding money are powerful persuasion tools and may convince people not to do Roths if they feel their lifespan is short when I don't think that is the author's intended message.

To hammer the point further, if you did a Roth conversion and then immediately died, the value of your estate to your heirs is not lower than if you did no conversion, at least in the case where your heirs have the same average tax burden on inherited IRA withdrawals as you had on the Roth conversion. So when measured against actual utility of the money, I don't see that there was ever a reduction of spendable money, so no "investment" in the traditional sense, so breakeven ages and ROIs are odd measures.

My suggestion to compare cases is to get everything equalized. I did a pair of runs in Retiree Portfolio Model which keeps track of each type of account. To be the simplest base case I could think of, I simply extended my assumed lifetime by 10 years and evenly withdrew the remaining funds from the TDA during that final 10 years to simulate the action of heirs. Then I simply compared the ending portfolio value of the Roth vs. no Roth case with all accounts converted to taxable. If I would rather have the reference point be day of death instead of 10 years on, I could use a present value formula, using the assumed rate of return as the discount rate. Obviously there are lots of permutations with regard spousal lifespans, heirs with different incomes, etc. but for the simplest case, that seems to work and provides a way to compare everything on an equal footing.
Please keep in mind that the focus in the article is a onetime conversion followed by an ongoing stream of RMDs which have been reduced by the conversion. The breakeven point is a quick summary, relatively meaningless in the abstract for all the reasons stated in various posts, but a useful point of comparison across scenarios. The article, and the spreadsheet, also examine a second set of metrics, in which the converted funds and the matching counterfactual are completely liquidated at one point in time, per your post. By these metrics the conversion always wins from the get-go, as soon as an RMD has to be taken.
That result is completely uncontroversial--you've seen complete liquidation analyses a hundred times. The contribution of the paper in this regard, to my mind, is to quantify the benefit using a real ROI framework. Sure the conversion comes out ahead, and the more so if future tax rates are higher; but as seen in the spreadsheet, the real ROI in the constant tax rate case (like the 22% - 22% scenario that crops up on this forum), may be less than 1.0% annualized in the 80s, and still not 2.0% even in the late 90s.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
McQ
Posts: 1425
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

milktoast wrote: Sat Jun 19, 2021 2:13 pm One thing I don’t understand is converting deep into 24% bracket.

Don’t you slam into several interesting thresholds? For MFJ the 24% bracket starts at roughly $171k + 24k = $195k MAGI

But for those on Medicare you hit IRMAA thresholds at $176k, $222k, and $276k MAGI.

And if you have taxable accounts (which I think you must to convert that much) the 3.8% NIIT hits at $250k.

So you end up having an effective 27.8% marginal rate in there along with short bursts in marginal rates due to IRMAA.

Are people considering this when thinking about going beyond $176k in MAGI after age 63?

The other question I have, and maybe this is only due to my planning for high spending, is why convert rather than just spending it? Is it better to convert and sell in taxable to cover spending or just to spend it outright.

Guess I’ll have to make a spreadsheet. The pre-made ones overwhelm me.
Indeed, for best results conversions have to either respect the IRMAA threshold if made after 65, or be made before Medicare. To your point about spending: way in the back of the paper, pp. 25-28, tables alpha 5 & 6, you will find a comparison of just withdrawing the money instead of converting, and of withdrawing to fund a year of spending so that the most advantageous conversion can be made.

Re: the spreadsheet: yeah, it was only after multiple attempts to construct a spreadsheet from scratch that I really got a grasp of all the moving parts. Best wishes, and even if you don't find my spreadsheet useful as a starting point, the Appendix explaining how and why it was constructed the way it is may be useful to you.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
User avatar
David Jay
Posts: 14587
Joined: Mon Mar 30, 2015 5:54 am
Location: Michigan

Re: Roth Conversions - McQuarrie study

Post by David Jay »

Dr. McQuarrie:

For background on the demographics here, the Boglehead average and median age of retirement is 58. Fully 70% retire before age 62 so it would be commonplace for folks to have a few years of 100% self funded retirement.
It's not an engineering problem - Hersh Shefrin | To get the "risk premium", you really do have to take the risk - nisiprius
Post Reply