Roth Conversions - McQuarrie study

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
printer86
Posts: 294
Joined: Mon Apr 25, 2016 8:45 am

Re: Roth Conversions - McQuarrie study

Post by printer86 »

McQ wrote: Mon Jun 21, 2021 2:08 pm
printer86 wrote: Mon Jun 21, 2021 8:54 am
retiredjg wrote: Mon Jun 21, 2021 8:39 am
printer86 wrote: Mon Jun 21, 2021 8:21 am I find McQ's challenge of writing a broad based study on the practicality of Roth conversions almost impossible since personal finance is really a confederation of niches.
I agree.

Th frustration about writing about Rob and Sue is that very few people will look like Rob and Sue and yet all of us are looking for guidance on this very complex issue.
Just to add, what McQ describes on page 3 of his study as a highly unlikely scenario (a retired couple with a significant TDA and hundreds of thousands of dollars just sitting in a taxable account) describes our household perfectly. For me, it's almost not worth reading the remaining 42 pages.
[with a nod to jg's response] Printer86, here is a second reply focused on your situation, which may make the paper more palatable. If you have "hundreds of thousands of dollars just sitting in a taxable account," then the proper counterfactual to a Roth conversion for you can be set up as follows:
Either:
1. deplete those funds over the next few years to fund a series of Roth conversions in the bottom brackets, each time writing a large check to the government, and locking up the converted funds under the five-year rule;
2. or, right now, take every surplus dollar and put it into the lowest cost, Total Market Index ETF you can find, there in the taxable account. Tax on dividends will subtract 30 basis points in return, otherwise you get the market return over what could be 40 - 50 years or more (heirs never have to sell the ETF, unlike the Roth).
The paper considers a similar scenario (table alpha 5, p. 25). The Roth conversion will still win if you have to liquidate the funds while alive. The reason is the embedded unrealized capital gain in the long-held ETF. But the Roth advantage over the taxable account may be much smaller than you expect; the ETF structure is incredibly tax-efficient in the case of stock index ETFs.
BTW, your situation is also considered in Table 6, pp. 14-15
McQ, I appreciate your willingness to engage in this discussion and will definitely read the rest of the paper later today (the Boston Marathon is 16 weeks from today. So, I have to get in my miles first). I will return to this thread when I complete my homework assignment.
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

Chip wrote: Mon Jun 21, 2021 2:20 pm Based on my calculations it appears that Rob & Sue can't retire in the manner envisioned in the base case and maintain anywhere close to their pre-retirement spending level. My calculations all are in constant 2020 dollars, so tax brackets don't change and neither does social security. Investment returns from the base case are reduced by the inflation rate to approximate a real return. Wage growth is assumed to be 1% real.

The base case scenario had their joint wage earnings at 400k at age 64, with 52k deferred into 401ks. Total taxes (2020 rates) are 119k (67k Federal, 29k CA and 23k FICA+Medicare. This leaves spendable income of 229k (400-52-119). The assumption in the base case is that all of this money is spent, since there is no money being saved in a taxable account.

The base case also assumes that one of the couple will retire at age 65 and that no further 401k contributions will be made from that point forward. With 200k of wage income and no deferrals, taxes will be 56k (30k Federal, 15k CA and 11k FICA+Medicare). This leaves spendable income of 144k. This calculation ignores IRMAA effects.

To achieve their pre-retirement spending level of 233k, they will have to pull enough from TDA to net 89k after tax. This works out to roughly a 129k withdrawal from their 2.9M TDAs.

This size withdrawal will actually slightly reduce their TDA balance, given that the nominal return in the base case is 5.51%; a real return of about 3%, given the assumed inflation rate of 2.5%. Note that this withdrawal rate is roughly 4.5% of their portfolio, above the oft-used 4% "guideline".

At age 70 social security begins. The 106k referenced in the base case deflates to 94k 2020 dollars, of which 85% is federally taxable and 0% taxable in CA. This income, plus the favorable tax treatment, allows much less to be pulled from the TDAs for spending; less than 10k each year.

At age 72 the second retirement begins. Total of the TDA accounts is now at about 2.8M. Replacing the lost wage income requires significant TDA withdrawals, approximately 200k/year. This corresponds to a withdrawal rate of close to 7% while facing a 23 year retirement if both live to age 95.

If the 200k withdrawals continue they will exhaust all savings before age 90, after less than 18 years of full retirement.
Chip, that's an excellent analysis with an outstanding level of attention to detail. It warms my authorial heart that you grappled with my hypothetical case to that extent. It raises a general question that must have come up in past Boglehead forums (which I'd like to peruse if anyone has links to hand): what percentage of income needs to be replaced in retirement? And does it vary by level of affluence? (of course).
In early drafts of the paper I played around with an income statement for Rob and Sue parallel to what you constructed. My goal was to show that there must have been a surplus in their early 60s; that is the only way that they could have survived on half their income when one hypothetically retired (as you show, there aren't enough tax savings for half the income at 65 to support all the expenditure from the early 60s, if everything was spent then). For critical purposes, and in the absence of the background I am supplying here, you inferred: they must have spent it all, setting the standard of living which they will attempt to maintain after retirement. And again as you show, they would then have to make premature withdrawals, and take more than the RMD, and they will run out of money all too soon, if indeed they try to match the spending level they had when making $400,000 in their early 60s.
So in my imagination, the spending they could do when one retired and their income was cut in half to $200,000 is the spending they will attempt to maintain in retirement. And social security and RMDs do that, even though they amount to far less than the 70% rule (of $400k) often seen.
My motive in devising the Rob and Sue example may be relevant here. The popular literature says: "you should consider a Roth conversion if you find yourself in a temporarily lower tax bracket, say in your 60s." So I manufactured a case where the tax bracket was "lower" (Rob and Sue go from 32% to 24%), but not low in an absolute sense, and where it was plausible that they would have millions in TDA. And I needed them to have no taxable savings, since if so, they could then wait and convert at zero percent when both retired, say at 66. In other words: I wanted to test conversion outcomes when the bracket was "lower" but not zero. (Based on this forum, I'll add the 22% case as well in a revision.)
So where did all that money go in their early 60s, if not into a taxable account, or ongoing life style? Hmm, as I recall they spent some to do a cash-in refinance to significantly lower their California house payment; another year they took their child on a graduation trip to Scotland; the next year they redid the landscaping; and the year after that they took their own trip to New Zealand. All one-time expenditures not built into their life style and not needing to be funded from TDA withdrawals. Prior to the 60s every surplus dollar was spent on college savings and staying on top of the very large mortgage on their rather fine California house ...
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

printer86 wrote: Mon Jun 21, 2021 2:26 pm
McQ wrote: Mon Jun 21, 2021 2:08 pm
printer86 wrote: Mon Jun 21, 2021 8:54 am
retiredjg wrote: Mon Jun 21, 2021 8:39 am
printer86 wrote: Mon Jun 21, 2021 8:21 am I find McQ's challenge of writing a broad based study on the practicality of Roth conversions almost impossible since personal finance is really a confederation of niches.
I agree.

Th frustration about writing about Rob and Sue is that very few people will look like Rob and Sue and yet all of us are looking for guidance on this very complex issue.
Just to add, what McQ describes on page 3 of his study as a highly unlikely scenario (a retired couple with a significant TDA and hundreds of thousands of dollars just sitting in a taxable account) describes our household perfectly. For me, it's almost not worth reading the remaining 42 pages.
[with a nod to jg's response] Printer86, here is a second reply focused on your situation, which may make the paper more palatable. If you have "hundreds of thousands of dollars just sitting in a taxable account," then the proper counterfactual to a Roth conversion for you can be set up as follows:
Either:
1. deplete those funds over the next few years to fund a series of Roth conversions in the bottom brackets, each time writing a large check to the government, and locking up the converted funds under the five-year rule;
2. or, right now, take every surplus dollar and put it into the lowest cost, Total Market Index ETF you can find, there in the taxable account. Tax on dividends will subtract 30 basis points in return, otherwise you get the market return over what could be 40 - 50 years or more (heirs never have to sell the ETF, unlike the Roth).
The paper considers a similar scenario (table alpha 5, p. 25). The Roth conversion will still win if you have to liquidate the funds while alive. The reason is the embedded unrealized capital gain in the long-held ETF. But the Roth advantage over the taxable account may be much smaller than you expect; the ETF structure is incredibly tax-efficient in the case of stock index ETFs.
BTW, your situation is also considered in Table 6, pp. 14-15
McQ, I appreciate your willingness to engage in this discussion and will definitely read the rest of the paper later today (the Boston Marathon is 16 weeks from today. So, I have to get in my miles first). I will return to this thread when I complete my homework assignment.
As they say, you can take the academic out of the classroom, but you can't stop him from giving homework :happy
User avatar
LadyGeek
Site Admin
Posts: 74118
Joined: Sat Dec 20, 2008 5:34 pm
Location: Philadelphia
Contact:

Re: Roth Conversions - McQuarrie study

Post by LadyGeek »

^^^ :)
McQ wrote: Mon Jun 21, 2021 5:55 pm It raises a general question that must have come up in past Boglehead forums (which I'd like to peruse if anyone has links to hand): what percentage of income needs to be replaced in retirement? And does it vary by level of affluence? (of course).
Consider that this is a "DIY" (Do-It-Yourself) investing and personal finance forum. We have an article series in the wiki: Retirement spending

In detail: Replacement rate models of retirement spending

We also have a DIY spreadsheet where you can input your situation to decide if Roth conversions will work for you. Retiree Portfolio Model

Wiki content is written by Bogleheads forum members - anonymously - and it represents the broad experience of the members' knowledge base.
Wiki To some, the glass is half full. To others, the glass is half empty. To an engineer, it's twice the size it needs to be.
mouth
Posts: 305
Joined: Sun Apr 19, 2015 6:40 am

Re: Roth Conversions - McQuarrie study

Post by mouth »

McQ wrote: Mon Jun 21, 2021 1:14 pm Re the "challenge of writing a broad-based study": that's all I can or want to do as a scholar. If the paper was doomed from the start, because everything is a niche, then we've all been wasting our time on this forum since Thursday night.
Please know I'm only poking fun when I say thins ... of all people, as a scholar, you should know better :p

Time is never wasted even if the conclusion of research / debate is nothing more than, "it all means nothing." Because you've still expanded the body of knowledge so that others can move on to other avenues of inquiry.

So even if that's where we land, I thank you for doing this, because it will mean I can stop fretting :beer
Last edited by mouth on Mon Jun 21, 2021 6:28 pm, edited 1 time in total.
livesoft
Posts: 76841
Joined: Thu Mar 01, 2007 8:00 pm

Re: Roth Conversions - McQuarrie study

Post by livesoft »

McQ wrote: Mon Jun 21, 2021 5:55 pm... It raises a general question that must have come up in past Boglehead forums (which I'd like to peruse if anyone has links to hand): what percentage of income needs to be replaced in retirement? And does it vary by level of affluence? (of course).
Well, it has been written often that once one stops working, then they are no longer contributing to their retirement plans nor paying FICA and medicare taxes. So if working folks are making the legal maximum contributions to 401(k)/403(b) and Roth IRAs (and maybe HSAs) as well as paying 7.65% in FICA/medicare taxes, then the income to do those things is not going to current other expenses. If the mortgage is paid off, then that expense goes away as well. Paying for children expenses and their college expenses goes away in most circumstances, too, although I will admit those are not large expenses in the first place. Plus income used to pay income taxes is not needed in retirement.

I started two threads on our situation.
Taxes on a family with $200,000 gross income
and
How to pay ZERO taxes in retirement with 6-figure expenses

While these do not show directly what percentage of income needs to be replaced in retirement, I think they show that a substantial amount of pre-retirement income is unneeded in retirement when big categories of expenses go away.
Wiki This signature message sponsored by sscritic: Learn to fish.
printer86
Posts: 294
Joined: Mon Apr 25, 2016 8:45 am

Re: Roth Conversions - McQuarrie study

Post by printer86 »

Ok, homework assignment done. Well, done close enough for this middling student. I'll admit that I skimmed over some of the additional scenarios in second half of the paper as I already understood your overall point that even affluent couples should not blindly accept the recommendation to aggressively Roth convert.

I see that the notes and tables address additional nuances. But, as other posters have mentioned, the base scenario could be better envisioned. Like having a couple retire a couple years earlier with some taxable funds enabling them to perform several years of Roth conversions. The base scenario also doesn't address the eventual surviving spouse's tax dilemma.

Us Bogleheads live for academic papers like yours. We appreciate your work and your willingness to discuss it here. I for one can't want for the sequel. I wonder if Rob and Sue stay together. :wink:

PS. I think I found a typo on the second to last paragraph on table 3, page 6. The last line says "Defined Compensation Plan". I think you meant "Deferred Compensation Plan".
User avatar
WoodSpinner
Posts: 2109
Joined: Mon Feb 27, 2017 1:15 pm

Re: Roth Conversions - McQuarrie study

Post by WoodSpinner »

McQ wrote: Fri Jun 18, 2021 1:39 pm
WoodSpinner wrote: Fri Jun 18, 2021 10:21 am
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."
Edward,

Welcome to our Forum! I really appreciate the Opportunity to discuss a paper and it’s findings with the author directly. Many thanks for the opportunity.

I am on year 4 of a Roth conversion strategy and the resulting taxes are my largest expense. Given that, I do have a vested interest in the subject and want to make sure my plans are well thought out.

My first step is a deeper reading of the paper and review of your spreadsheet. The first pass confirmed that we are likely to benefit from aggressive conversions. Will come back with additional questions.

Would you be willing to opine on our strategy if we provide the appropriate background?

Thanks

WoodSpinner
Woodspinner, here I have to be careful. I'm neither qualified nor licensed to give financial advice to individuals. Nor should you share too many financial details in a public forum. 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.
Edward,

I have finished reviewing the paper and many of the posts in this thread. As others have pointed out there are a few key differences in philosophy around Roth Conversions that aren't adequately addressed in the paper.

1. The fundamental metric should be the Present Value of the Total Expected Portfolio Value -- After Taxes are paid. I would look at this in terms of Last-to-Die for a Married Couple and then for Heirs assuming a 10 year stretch.

2. We really need some framework for planning a series of Roth Conversions that accounts for the vagaries of financial and tax situations. Some rubrics would be helpful that help plan both the current years conversions and plans for future years. For instance many advocate trying to balance AGI across your expected lifetimes which may work well for some but will provide less than optimal results for states with high taxes (e.g. CA) that do not tax Social Security.

3. You may need to expand your examples a bit to cover other Financial realities. For instance, I retired in 2018 with a very small After-Tax portfolio (<6%), a very small Roth (<1%) and a large Tax-Deferred Account (93% approx). I have encountered many others with a pretty similar situation on the BH forum and as callers on various Retirement Planning podcasts.

4. There are some very good concepts laid out in the paper about the pitfalls of analyzing Roth Conversions and some of the common mistakes. I think this list should be called out and organized in a useful Appendix -- a sort of checklist for modeling conversions.

5. Many of us struggle with just how much to leave in a TDA rather than converting. For instance leaving a portion behind for QCDs, Medical Expenses, Charitable Bequests, Bequests to Heirs in lower Marginal brackets are key to planning a conversion strategy.

6. Your Rules of Thumb for conversions in the higher brackets are a bit difficult to follow and could be improved through some clearer verbiage. The low-end is covered well (0%, 10% Marginal Rates are almost a no-brainer). The bigger problem is on the higher end Marginal Rates. I think something that includes Age, Pension, SS, Other Income, Long Term Capital Gains and TDA Balance would be helpful.

That said, there is a lot I like about your paper:
1. You lay out a set of clear and reasonable assumptions that drive your calculations.
2. There is a checklist of sorts that you lay out for how to approach modeling conversions.
3. I agree with the general point -- Roth Conversions are not for everyone. They need to be modeled based on each persons finances, plans and goals.

Many of us have either built our own Retirement Model or are using the Retirement Portfolio Model.

In my case, I have built my own Retirement Model and I was pleased that my model aligns with the best practices laid out in your paper -- this review alone was a great ROI on my time.

FWIW, I never considered myself a Super Saver. I contributed 10-15% of my salary to TDA accounts throughout my 40 year career and we never made over $200K MFJ. Time, compound growth, staying the course, and lots of luck helped grow the portfolio.

I am convinced that a series of Roth Conversions are a good strategy in our case since:
1. We retired early (Me-59, Wife-58) in 2018 so we have time for conversions
2. I have Medical Insurance through my Employers Retirement Plan -- No ACA plan needed.
3. PV of Tax Discounted Portfolio is about $300K higher with currently planned Conversions vs. Base Case.
4. PV of Total Expected Taxes Paid is $400K less with currently planned Conversions vs. Base Case.
5. PV of Total Expected IRMA Costs is $128K less with currently planned Conversions vs. Base Case.
6. Highest Marginal Bracket hit is 37.3% (Current Plan) vs 42.3% (Base Case)
7. Better Tax Diversification

Code: Select all

	                                        Base Case	 Current Plan
Taxable Accounts at 90	                          49%	           22%
Roth Accounts at 90	                          1%	           52%
IRA Accounts at 90	                          50%	           26%
8. If one of us passes early (which is probably pretty likely) the Current Plan will perform even better than the results above.

** Base Case -- no additional Roth Conversions in 2021 or beyond
** Current Plan -- Convert into IRMA Tier3 till I turn 70 in 2029 and start SS
** Assumes I live to 90, wife to 89
** Taxes and Marginal Brackets include California Taxes
** PV Discount Rate of 1.66%, 30 year Treasury as of 12/30/2020 when model last updated
** Weighted Expected Real Return 2.46% (4.82% Nominal)
** Plan to have QCDs be at least 25% of expected RMDs
** Non-spouse beneficiaries are likely to be in the 12% Marginal Bracket
curmudgeon
Posts: 2255
Joined: Thu Jun 20, 2013 11:00 pm

Re: Roth Conversions - McQuarrie study

Post by curmudgeon »

I've been rather critical of the paper, but I think the point that quite large TDAs may see limited value from Roth conversions is well worth exploring. It might be worth doing in the context of someone who is single, as that removes the question of what happens at the transition from MFJ to Single tax brackets. I really like the approach of including exploration of what happens if tax rates/brackets change; sensitivity analysis along this point is important to make sure your Roth conversion plan isn't teetering on some rare little odd spike in the tax code. It's a near certainty that tax laws will change in some direction over the next 30 years.
McQ wrote: Mon Jun 21, 2021 5:55 pm So in my imagination, the spending they could do when one retired and their income was cut in half to $200,000 is the spending they will attempt to maintain in retirement. And social security and RMDs do that, even though they amount to far less than the 70% rule (of $400k) often seen.
My motive in devising the Rob and Sue example may be relevant here. The popular literature says: "you should consider a Roth conversion if you find yourself in a temporarily lower tax bracket, say in your 60s." So I manufactured a case where the tax bracket was "lower" (Rob and Sue go from 32% to 24%), but not low in an absolute sense, and where it was plausible that they would have millions in TDA. And I needed them to have no taxable savings, since if so, they could then wait and convert at zero percent when both retired, say at 66. In other words: I wanted to test conversion outcomes when the bracket was "lower" but not zero. (Based on this forum, I'll add the 22% case as well in a revision.)
Interestingly enough, I think there is still actually a decent return to be had from making that contrived $100K conversion at age 65. Rob and Sue will be netting about $170K after tax on that $200K income, when they draw it from TDA after age 66. If, at age 70 and 71 when they are drawing $100K SS but not doing RMDs, they draw $50K from the Roth each year, they will only need to pull $25K from the TDA those years to get to $170K spendable income. They'll save about $16K because for those two years they will avoid much of the "tax torpedo" of taxation of their SS benefits. $16K is not a bad return for doing a $100K Roth conversion. It's worth noting that this does not scale much further. Once RMDs start, SS will be fully taxed unless they've done some really large conversions.
McQ wrote: Mon Jun 21, 2021 5:55 pm So where did all that money go in their early 60s, if not into a taxable account, or ongoing life style? Hmm, as I recall they spent some to do a cash-in refinance to significantly lower their California house payment; another year they took their child on a graduation trip to Scotland; the next year they redid the landscaping; and the year after that they took their own trip to New Zealand. All one-time expenditures not built into their life style and not needing to be funded from TDA withdrawals. Prior to the 60s every surplus dollar was spent on college savings and staying on top of the very large mortgage on their rather fine California house ...
I think I know some of these folks... They are financial optimists who think every dollar is there to spend. The actually tend to do cash-out refis, and maybe raid the 401K if they hit a spell of unemployment. They may end up working into their late 60's or 70's. You don't find them on Bogleheads much, though, unless they are penitents coming in for a dose of "spenders anonymous" rehab.
User avatar
dodecahedron
Posts: 5699
Joined: Tue Nov 12, 2013 12:28 pm

Re: Roth Conversions - McQuarrie study

Post by dodecahedron »

curmudgeon wrote: Tue Jun 22, 2021 12:41 am I've been rather critical of the paper, but I think the point that quite large TDAs may see limited value from Roth conversions is well worth exploring. It might be worth doing in the context of someone who is single, as that removes the question of what happens at the transition from MFJ to Single tax brackets. I really like the approach of including exploration of what happens if tax rates/brackets change; sensitivity analysis along this point is important to make sure your Roth conversion plan isn't teetering on some rare little odd spike in the tax code. It's a near certainty that tax laws will change in some direction over the next 30 years.
+1 to the need for sensitivity analysis.

Not only is it a near certainty that tax laws will change in some direction over the next 30 years, even in the absence of tax law change, there is a huge amount of uncertainty about future path of one's AGI since the future stream of returns on investment is also highly unpredictable.

My late husband did quite a bit of Roth conversion in the decade before his unexpected death in his late 50s eight years ago. As a finance professor with a highly volatile but often very substantial side consulting income, he did so opportunistically in years when AGI dipped. He was also relatively very aggressive (and very lucky!) in his investments in the Roth, so ex post, it has turned out quite well for me as surviving spouse, despite the fact that tax bracket rate schedules were higher back when he was converting than they are now.

My portfolio now is roughly 30% Roth, 20% tax-deferred, 50% taxable. My tax-deferred is entirely in fixed income (TIAA Trad SRA yielding 3%, which is their guaranteed minimum) so very predictable stream of future RMDs for me [unless interest rates ever mean-revert!].

But most folks with large proportion of tax-deferred typically would have some equities in their tax-deferred and have far more unpredictable future RMDs than I do. Also, I am reasonably certain about my tax filing status staying the same (Single) over the remainder of my life.

At this point in my life (age 68, with RMDs four years away), I have crunched the numbers relevant to my situation going forward and my personal conclusions about whether and to what extent additional conversions make sense for me generally comport with most of the broad conclusions of the paper as stated in the abstract.
paper's abstract wrote: • Future tax rates need not be higher for a conversion to pay off;
• Nor is it all that helpful to pay the tax on conversion from outside funds;
• Nor are Roth conversions especially beneficial for top bracket taxpayers as compared to middle class taxpayers;
• Rather, the greatest benefit accrues to taxpayers who can make the conversion partly in the zero percent tax bracket, i.e., during a year with no other taxable income.
Edited to add: at least in my case, it is a no-brainer to use outside funds to pay the tax on any conversions I do, since I have a large taxable account and there is no reason not to do so. Barring major unanticipated future disasters (e.g., need for many years of expensive LTC), my Roth will probably go as a bequest to my daughters. I live in a state (NY) with an estate tax that kicks in (with a cliff!) well before the federal estate tax does. Kudos to forum member bsteiner for frequently reminding me of this. As he has pointed out, Roth conversions make particular sense if there is any possibility of state estate taxes kicking in.

That said, any additional Roth conversions that I do will likely be "tinkering around the margins." For 2021, I have a pretty good handle on my effective marginal tax rate (taking into account many things besides my official "bracket") and believe that I can convert about $12K at an effective marginal tax rate of 18.5% (Fed) which would also be state tax free (NY allows up to $20K per year in tax free retirement distributions). If I go much beyond that, I quickly approach a range where the effective federal marginal tax rate would be 49.5% (hot spot on the heat map) and also start to approach IRMAA issues, issues with qualifying income for senior property tax breaks, etc.

$12K is a tiny "drop in the bucket" relative to the size of my tax-deferred account, but whatever. I will do it in the spirit of stewardship as a tribute to my beloved late husband, since it seems a no-brainer.

I do have the ability to manage my future taxable RMDs to an unusual extent, since a significant portion of my current budget is charitable giving, which will take the form of QCDs once I am over 70 1/2. (Currently funding my charitable giving by donating appreciated securities and itemizing deductions.) I also still work (very part-time) and may continue to do so after 72, which will give me further options on delaying RMDs (since I have rolled all the tax-deferred into a 403b at my employer.) Another unusual factor for me is that it made no sense to wait until 70 to file for SS, since widow's benefits maximize at FRA and do not get delayed retirement credits, so I am already collecting SS. Also, my AGI includes a large amount of qualified dividend income due to my large taxable account. So I have many "moving parts" to my analysis.

Bottom line: the amount of money saved by doing additional Roth conversions at this stage of my life is probably not worth the time I have spent thinking about it (except that it keeps my brain engaged!) Kind of like doing a crossword puzzle.
cas
Posts: 1383
Joined: Wed Apr 26, 2017 8:41 am

Re: Roth Conversions - McQuarrie study

Post by cas »

curmudgeon wrote: Tue Jun 22, 2021 12:41 am Interestingly enough, I think there is still actually a decent return to be had from making that contrived $100K conversion at age 65. Rob and Sue will be netting about $170K after tax on that $200K income, when they draw it from TDA after age 66. If, at age 70 and 71 when they are drawing $100K SS but not doing RMDs, they draw $50K from the Roth each year, they will only need to pull $25K from the TDA those years to get to $170K spendable income. They'll save about $16K because for those two years they will avoid much of the "tax torpedo" of taxation of their SS benefits. $16K is not a bad return for doing a $100K Roth conversion. It's worth noting that this does not scale much further. Once RMDs start, SS will be fully taxed unless they've done some really large conversions.
It is true that the age 70/71 "avoid tax torpedo" tactic can't be expected to extend beyond those two years.

However, the example points out the potential usefulness of Rob and Sue having the availability of at least some assets in Roth.

In another very similar example, in 2008 and 2020, Congress waived RMDs due to economic circumstances. While those one-off waivers can't be assumed to ever happen again, *if* an RMD waiver ever happened again, Rob and Sue would find having Roth assets very useful. Roth assets would give them the flexibility to manage their taxable income and, once again, avoid their SS tax torpedo. (They would have to have converted more than $100K, though, in order to still have Roth assets *after* they used Roth assets to avoid the SS tax torpedo at age 70/71.)

In another example of where Rob and Sue might find Roth assets useful in managing a potential "tax torpedo," it has been niggling at me that Rob and Sue look like the perfect candidates to be flirting with the Alternative Minimum Tax "tax hump" under the pre-2018/post-2025 tax law. Like the SS tax hump, the AMT tax hump is very dependent on individual circumstances, but its 32.5% and 35% marginal rate "bump" tends to cover quite a big span in what is now the 22% and 24% brackets under the conventional tax system.***

Dr. McQuarrie already discusses an alternative scenario where the conventional tax brackets revert to their pre-2018 form, as currently scheduled. He also discusses that Rob and Sue might be living in California in their "rather fine house" for all or part of their retirement. Income in Rob and Sue's range + high state taxes + high property taxes is a stereotypical scenario where the AMT tax system would kick in pre-2018. But someone who knows more than I about California income and property tax law would need to judge whether Rob and Sue, as retirees and having apparently owned their house for quite a while, would have high enough state taxes and property taxes to be at risk of switching from the conventional tax system to the AMT tax system.

*If* Rob and Sue are determined to be seriously flirting with AMT in their retirement years, then Roth conversions come into play in two ways:
1. Could some reasonably attainable amount of Roth conversions pull their RMDs down just enough for them to avoid triggering AMT for some number of years? Or at least avoid the 32.5% or 35% AMT marginal rates on the extra amount of RMD that is there without Roth conversion and isn't there with Roth conversion? (Similar to the way that Roth conversions helped them avoid an additional IRMAA tier for some years in the "base case".)
2. Would the availability of Roth assets help them avoid AMT in years where they might need a bit of extra income beyond what is coming from SS + RMDs? Or at least avoid the 32.5% or 35% AMT marginal rates on the extra withdrawn amount? (For example, if they need to replace the roof or siding on their "rather fine" house or if they wanted to take a special vacation with their young adult children.)


***For more info on the general concept of the AMT "tax hump" see the "AMT bump zone" graph near the bottom of Michael Kitces 2014 article "Evaluating Exposure To The Alternative Minimum Tax And Strategies To Reduce The AMT Bite" . And, yes, that is a 2014 article, so all the numbers are dated and calculations would need to be adjusted for inflation. But the general concept holds for the post-2025 tax law as currently scheduled to return.
Last edited by cas on Tue Jun 22, 2021 8:41 am, edited 2 times in total.
Dottie57
Posts: 10112
Joined: Thu May 19, 2016 5:43 pm
Location: Earth Northern Hemisphere

Re: Roth Conversions - McQuarrie stud

Post by Dottie57 »

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.
+1. And tax brackets are a challenge in retirement too, and in medicare rates.
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

printer86 wrote: Mon Jun 21, 2021 8:41 pm Ok, homework assignment done. Well, done close enough for this middling student. I'll admit that I skimmed over some of the additional scenarios in second half of the paper as I already understood your overall point that even affluent couples should not blindly accept the recommendation to aggressively Roth convert.

I see that the notes and tables address additional nuances. But, as other posters have mentioned, the base scenario could be better envisioned. Like having a couple retire a couple years earlier with some taxable funds enabling them to perform several years of Roth conversions. The base scenario also doesn't address the eventual surviving spouse's tax dilemma.

Us Bogleheads live for academic papers like yours. We appreciate your work and your willingness to discuss it here. I for one can't want for the sequel. I wonder if Rob and Sue stay together. :wink:

PS. I think I found a typo on the second to last paragraph on table 3, page 6. The last line says "Defined Compensation Plan". I think you meant "Deferred Compensation Plan".
Spot on, Printer86: should have been Deferred Compensation Plan (for the cognoscenti, a 409A plan, a type of non-qualified plan). And BTW, spotting a typo in footnote 18 on p. 28 rather proves that you completed the assignment, don’t you think? Best wishes!
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

WoodSpinner wrote: Mon Jun 21, 2021 11:13 pm
McQ wrote: Fri Jun 18, 2021 1:39 pm
WoodSpinner wrote: Fri Jun 18, 2021 10:21 am
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."
Edward,

Welcome to our Forum! I really appreciate the Opportunity to discuss a paper and it’s findings with the author directly. Many thanks for the opportunity.

I am on year 4 of a Roth conversion strategy and the resulting taxes are my largest expense. Given that, I do have a vested interest in the subject and want to make sure my plans are well thought out.

My first step is a deeper reading of the paper and review of your spreadsheet. The first pass confirmed that we are likely to benefit from aggressive conversions. Will come back with additional questions.

Would you be willing to opine on our strategy if we provide the appropriate background?

Thanks

WoodSpinner
Woodspinner, here I have to be careful. I'm neither qualified nor licensed to give financial advice to individuals. Nor should you share too many financial details in a public forum. 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.
Edward,

I have finished reviewing the paper and many of the posts in this thread. As others have pointed out there are a few key differences in philosophy around Roth Conversions that aren't adequately addressed in the paper.

1. The fundamental metric should be the Present Value of the Total Expected Portfolio Value -- After Taxes are paid. I would look at this in terms of Last-to-Die for a Married Couple and then for Heirs assuming a 10 year stretch.

2. We really need some framework for planning a series of Roth Conversions that accounts for the vagaries of financial and tax situations. Some rubrics would be helpful that help plan both the current years conversions and plans for future years. For instance many advocate trying to balance AGI across your expected lifetimes which may work well for some but will provide less than optimal results for states with high taxes (e.g. CA) that do not tax Social Security.

3. You may need to expand your examples a bit to cover other Financial realities. For instance, I retired in 2018 with a very small After-Tax portfolio (<6%), a very small Roth (<1%) and a large Tax-Deferred Account (93% approx). I have encountered many others with a pretty similar situation on the BH forum and as callers on various Retirement Planning podcasts.

4. There are some very good concepts laid out in the paper about the pitfalls of analyzing Roth Conversions and some of the common mistakes. I think this list should be called out and organized in a useful Appendix -- a sort of checklist for modeling conversions.

5. Many of us struggle with just how much to leave in a TDA rather than converting. For instance leaving a portion behind for QCDs, Medical Expenses, Charitable Bequests, Bequests to Heirs in lower Marginal brackets are key to planning a conversion strategy.

6. Your Rules of Thumb for conversions in the higher brackets are a bit difficult to follow and could be improved through some clearer verbiage. The low-end is covered well (0%, 10% Marginal Rates are almost a no-brainer). The bigger problem is on the higher end Marginal Rates. I think something that includes Age, Pension, SS, Other Income, Long Term Capital Gains and TDA Balance would be helpful.

That said, there is a lot I like about your paper:
1. You lay out a set of clear and reasonable assumptions that drive your calculations.
2. There is a checklist of sorts that you lay out for how to approach modeling conversions.
3. I agree with the general point -- Roth Conversions are not for everyone. They need to be modeled based on each persons finances, plans and goals.

Many of us have either built our own Retirement Model or are using the Retirement Portfolio Model.

In my case, I have built my own Retirement Model and I was pleased that my model aligns with the best practices laid out in your paper -- this review alone was a great ROI on my time.

FWIW, I never considered myself a Super Saver. I contributed 10-15% of my salary to TDA accounts throughout my 40 year career and we never made over $200K MFJ. Time, compound growth, staying the course, and lots of luck helped grow the portfolio.

I am convinced that a series of Roth Conversions are a good strategy in our case since:
1. We retired early (Me-59, Wife-58) in 2018 so we have time for conversions
2. I have Medical Insurance through my Employers Retirement Plan -- No ACA plan needed.
3. PV of Tax Discounted Portfolio is about $300K higher with currently planned Conversions vs. Base Case.
4. PV of Total Expected Taxes Paid is $400K less with currently planned Conversions vs. Base Case.
5. PV of Total Expected IRMA Costs is $128K less with currently planned Conversions vs. Base Case.
6. Highest Marginal Bracket hit is 37.3% (Current Plan) vs 42.3% (Base Case)
7. Better Tax Diversification

Code: Select all

	                                        Base Case	 Current Plan
Taxable Accounts at 90	                          49%	           22%
Roth Accounts at 90	                          1%	           52%
IRA Accounts at 90	                          50%	           26%
8. If one of us passes early (which is probably pretty likely) the Current Plan will perform even better than the results above.

** Base Case -- no additional Roth Conversions in 2021 or beyond
** Current Plan -- Convert into IRMA Tier3 till I turn 70 in 2029 and start SS
** Assumes I live to 90, wife to 89
** Taxes and Marginal Brackets include California Taxes
** PV Discount Rate of 1.66%, 30 year Treasury as of 12/30/2020 when model last updated
** Weighted Expected Real Return 2.46% (4.82% Nominal)
** Plan to have QCDs be at least 25% of expected RMDs
** Non-spouse beneficiaries are likely to be in the 12% Marginal Bracket
Woodspinner—thank you for that detailed reply. Over the decades, I’ve endured many peer reviews from scholarly journals that were far less informative (or constructive).
Your comment that you were never a super saver but still ended up with (millions) in your TDA is well worth restating. It truly is amazing what monthly contributions, invested in the Boglehead way, can mount up to over a 40 year working career.
Now, since I put you off in my first reply (can’t give individual finance advice, yada yada), please let me make amends with a few specific comments on what you have done.
1. You built a spreadsheet to capture your likely situation. ‘Nuff said—those who build their own spreadsheets, they shall be saved. Those who can’t or won’t, will have to find a friend, or pay an advisor—the equivalent of extra tax drag. However, building your own custom spreadsheet is, I believe you will agree, not that easy to do.
2. I discourage you from assuming a specific age at death, the more so jointly. And your chosen ages—90 for you, 89 for wife—are far too young for the affluent, conscientious demographic in which you reside. Playing things out until 95 for you (if you are male) and 105 for your wife (if female) would be a worthwhile exercise.
a. Sorry, I live among the woke, and dare not infer gender from terms like ‘wife’
b. It probably won’t change anything, but … will the heirs still be in the 12% bracket if your spouse does not pass until 105?
c. Aside: I have no training in financial economics and never developed a loyalty to net present value calculations. They seem to enforce selection of an ending date—i.e., your death at 90. I prefer future value constructions, where the rate of appreciation can be varied at will, and the scenario played out to an arbitrarily late date, and then roots computed to get at ROI. FYI: not a few life insurance companies run their illustrations out to 120 these days.
d. NPV analyses also demand selection of a discount rate. Your choice of rate is blue chip respectable; but again, I’d rather let the future play out, under annualized assumptions or pasting in actual historical data, and see how things change, when the rates change.
3. Very smart that you model IRMAA. When I applied the learning from the paper back to my personal situation, it became apparent that we were overly vulnerable to entering the second (third) IRMAA level, and now I am noodling about how to stay in the first IRMAA level—which as a previous poster remarked, is almost impossible to avoid if you are going to live large in retirement. Which I fully intend to do.
4. Per your statement, QCD are the all purpose buffer. Find yourself $5268 into an IRMAA bracket, marginal tax rate on the $5268 = 60%, no problem—make a QCD of $5500 that year. Typically less than 2.5% of income, not heroically charitable, doable if you have any charitable recipients in mind.
5. Last, play around with the decision to convert amounts through IRMAA #3. Still working on that new paper (“How to Cope with IRMAA”), but very preliminary analyses cause me to say, hey, better give that a second look. Yes, certainly, always go to the top of whatever IRMAA bracket applies—but to willfully push beyond, into the next bracket, or the bracket beyond, is not a decision I can as yet support.
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

iceport wrote: Mon Jun 21, 2021 12:21 pm
retiredjg wrote: Mon Jun 21, 2021 8:39 am
printer86 wrote: Mon Jun 21, 2021 8:21 am I find McQ's challenge of writing a broad based study on the practicality of Roth conversions almost impossible since personal finance is really a confederation of niches.
I agree.

Th frustration about writing about Rob and Sue is that very few people will look like Rob and Sue and yet all of us are looking for guidance on this very complex issue.
And another frustration (which gets the better of me, at times) is the frequency with which changes in Rob and Sue's particular lives or changes in tax laws will render all the tediously constructed deterministic tax planning scenarios moot.

From the 2005 Vanguard paper (now outdated), Tax Diversification and the Roth 401(k):
The history of continuous change in the tax code, along with these research findings, underscores the inherent uncertainty of future tax rates. Participants cannot be sure of their tax rate in retirement, and so cannot be sure of whether pre-tax or Roth savings are inherently superior. The tax system is dynamic and seems subject to almost continuous change. Current reform proposals call for everything from higher tax rates (favoring Roth savings) to a scrapping of the income tax entirely (possibly favoring pre-tax savings).

In this environment of uncertainty, how should participants manage the risk that taxes could be higher or lower in retirement? In the face of such risk, our recommendation is to diversify. In an uncertain tax world, participants should hold both pre-tax and Roth savings—the former to benefit in the event of lower tax rates in retirement, the latter to benefit in the event of higher tax rates. This is the notion of tax diversification—hedging the risk of uncertain future tax rates by holding both types of contributions.

There is a direct analogy between tax risk and investment risk. Investors may believe that common stocks are likely to generate a substantial equity risk premium. Yet they also recognize that higher equity returns are not guaranteed, and so diversify against that risk by holding other assets such as fixed income securities. In the case of taxes, participants may expect taxes to be lower (suggesting pre-tax savings) or higher (Roth savings) in retirement. But in pursuing a strategy of tax diversification, they will acknowledge the inherent uncertainty of forecasting any future tax rates, including their own, and so hold both types of savings.
"...acknowledge the inherent uncertainty of forecasting any future tax rates, including their own..."

The author does address this tax uncertainty, but only marginally, and only assuming minor, easily envisioned changes, which result in only minor differences in outcomes, apparently. And that's ironic, given that a major inspiration for the detailed analysis was the changes in tax laws in only the last 10 years — highlighted in the very first paragraph — the kind so bombastically dismissed later as irrelevant (because we only have one government :confused ).
Definitely gonna need to revise that section. Let me explain why the section is there and why I chose a provocative title. The section provides a vehicle to make the following claim: “It is only a matter of time before “tax-free” Roth distributions join “tax free” municipal bond interest in the calculation of MAGI used to determine whether Medicare IRMAA surcharges will be imposed.”
All Bogleheads who approach Roth conversions as a means of providing future tax-favored income need to accept this risk: that Roth distributions will cease to be tax-favored with respect to IRMAA.
That tax law change won’t matter if the converted funds are truly surplus, intended for beneficiaries and not for distribution while alive. Here the risk is that Congress imposes a stacking rule on heirs of a Roth account, as is now done for international income. Under stacking, a $100,000 Roth distribution will be tax-free, as before; but it will be placed at the bottom of the income stack, such that the next dollar of ordinary income will be taxed as if it were the $100,001st dollar of income, in the bracket that pertains to that income. The Roth will still be tax free to heirs, but the tax savings will be drastically reduced, since it averts tax at 0%, 10%, and 12% (=the bottom $100,000 in the current tax structure MFJ).
Iceport, can you help by suggesting a better title / positioning for the section, which would still allow these points to be made? I find far too much blithe optimism concerning the tax risk inherent in Roth conversions, and that's what I wish to combat.
User avatar
iceport
Posts: 4995
Joined: Sat Apr 07, 2007 4:29 pm

Re: Roth Conversions - McQuarrie study

Post by iceport »

McQ wrote: Tue Jun 22, 2021 1:17 pm
Definitely gonna need to revise that section. Let me explain why the section is there and why I chose a provocative title. The section provides a vehicle to make the following claim: “It is only a matter of time before “tax-free” Roth distributions join “tax free” municipal bond interest in the calculation of MAGI used to determine whether Medicare IRMAA surcharges will be imposed.”
All Bogleheads who approach Roth conversions as a means of providing future tax-favored income need to accept this risk: that Roth distributions will cease to be tax-favored with respect to IRMAA.
That tax law change won’t matter if the converted funds are truly surplus, intended for beneficiaries and not for distribution while alive. Here the risk is that Congress imposes a stacking rule on heirs of a Roth account, as is now done for international income. Under stacking, a $100,000 Roth distribution will be tax-free, as before; but it will be placed at the bottom of the income stack, such that the next dollar of ordinary income will be taxed as if it were the $100,001st dollar of income, in the bracket that pertains to that income. The Roth will still be tax free to heirs, but the tax savings will be drastically reduced, since it averts tax at 0%, 10%, and 12% (=the bottom $100,000 in the current tax structure MFJ).
Iceport, can you help by suggesting a better title / positioning for the section, which would still allow these points to be made? I find far too much blithe optimism concerning the tax risk inherent in Roth conversions, and that's what I wish to combat.
Professor McQuarrie,

Thank you for the explanation. And allow me to apologize for the intemperate tone of some of my comments. (An uncle once described how he thought he had outgrown a lousy personality; unfortunately, I never outgrew mine.)

The aspect of your analysis that triggered my ire was the lack of any real qualification of the results, which, as I envision it, would acknowledge the uncertainties inherent in any kind of tax planning, really. I think such a qualification is an essential component of your analysis, which is amazingly comprehensive and detailed, and has enormous value. Nevertheless, the topic needs to be addressed with a degree of humility, because the future is uncertain.

(As an aside, the same kind of thing used to make me fume when a certain poster who had mastered the art of back-testing and graphing the results would adamantly urge everyone who’d listen to adopt massive small value and REIT tilts. Like the pied piper, folks followed his lead, clamoring for the superior performance, which, it was strongly implied, was a practical certainty. Well, that argument has become more muted, lately.)

Personally, I agree with and fully support your intentions. I don’t think Roth conversions should be considered an obvious slam-dunk. By the same token, I don’t think even the best, most complete and flawless deterministic analysis of the tax benefits or detriments of Roth conversions should be treated as the last word on the subject, but rather as a valuable and powerful tool to assist in making informed choices in the face of tax laws that are confusing today *and* that have an uncertain future.

From your further explanation, it seems that you are not so much dismissive of what I’ve come to know as tax diversification (a term which some folks think is semantically incorrect, under a narrower definition that you apply in the paper), as you are critical of folks focusing on one type of tax risk and overlooking others. Really, your point could possibly be better made by embracing the notion that the future is unknowable, the concept of tax risk. That’s where your claim fits in: as a counterpoint to the thinking that tax risk only cuts one way with respect to Roth accounts. As you so correctly point out, new laws could just as easily undermine the tax benefit of Roth accounts. The level of detail with which you cover the possible treatment of Roth withdrawals provides a vivid illustration of a scenario in which tax risk shows up in an unexpected form.

However, it’s misguided, in my opinion, to cross the line from identifying what’s possible into making a prediction. As tempting as it is, and as well-founded as it might be, and as interesting as it might be to formulate, a prediction like that should not form the sole basis of action. Similarly, I think it’s misguided to position your portfolio to benefit only in the event of massive tax hikes, based on such a prediction. The whole point is that nobody really knows what the tax laws will look like in 10, 15 or 20 years from now. So I don’t agree with stating your claim so confidently as a prediction.

(I’ve predicted for the better part of 20 years now that federal income taxes only have one direction to go from here. Obviously, I’ve been wrong for a long, long time. Thankfully, I had the sense not to act on my prediction.)

To satisfy my own perspective, I’d love to see you incorporate the idea of not putting all one’s eggs in one tax treatment basket, or tax diversification, if you will. However, while directly related, I think your real concern is that folks aren’t taking the full scope of tax risk into account, but are instead only picking and choosing which forms of risk they want to address — based on imperfect and biased predictions. Your topic is highlighting tax uncertainty/tax risk, not attacking tax diversification.

If one truly accepts that they cannot know the future, it’s less likely they’ll fully commit to a grossly lopsided investing approach.

I hope this makes a little sense, that I'm explaining myself well enough.
"Discipline matters more than allocation.” ─William Bernstein
DSBH
Posts: 249
Joined: Tue Jul 21, 2020 3:31 pm
Location: Texas

Re: Roth Conversions - McQuarrie study

Post by DSBH »

I downloaded the spreadsheet and am still studying the paper, and then got confused on this one:
-------------------------
"...
2. Nor is it all that helpful to pay the tax on conversion from outside funds;
...
The analyst simply assumes that Rob and Sue have some hundreds of thousands of dollars set aside. This assumption goes unexamined for good reason: the logic of arranging enough conversions, at a low enough tax rate, to make a material difference virtually requires the existence of that large sum of taxable dollars sitting around doing nothing.
...
For these reasons, the base case in the analysis of intertemporal tax arbitrage must be that tax on the conversion is paid from the converted funds.
Payment from some other source is one of many interesting alternatives to be explored but cannot be the base case."
-------------------------
As the spreadsheet shows, Rob and Sue made a 100K conversion in 2020, paid 24K in taxes and ended up with 76K in the Roth account. Table 3 shows that the floor of the 24% bracket is 198K, so it looks like Rob and Sue got quite a bit of money in their Taxable account to live without having to touch either tIRA or Roth account.

Does it appear from the example that it is not helpful to pay the tax on conversion from the converted funds (instead of outside funds)?
John C. Bogle: "Never confuse genius with luck and a bull market".
VanGar+Goyle
Posts: 67
Joined: Sat May 29, 2021 1:31 pm

Re: Roth Conversions - McQuarrie study

Post by VanGar+Goyle »

I wonder where the median Boglehead fits in the larger universe of individuals with substantial IRA or 401(k) balances? Reports by Fidelity indicate that even a $1 million dollar TDA balance puts a saver in the 98th to 99th percentile among the 16 million or so accounts they administer.
Vanguard may be a more appropriate place to check for Bogleheads.
Their how-america-invests-2020.pdf report is 58 pages long, breaks data down by average and quartiles, but not as detailed.
It is subtitled 'A look at the personal investing behavior of Vanguard’s 5 million retail households'.
Average figures: Household size is two persons. Portfolio holdings, on average, are two accounts and about five investments,
with average account balance of $352,200. The median account balance was $60,900, so not close to $1 million.

Is there a better Vanguard report, perhaps more detailed and institutional oriented?
The How America Saves 2021 report does cover institutional accounts, with 29% of balances more than $100,000.
At age 55–64, the average balance was $232,379, the median balance was $84,714, so still not detailed enough and not that close to $1 million.
Well, you pay a little bit, we're a little bit tough. | You pay very much,very much tough. | You pay a too much, we're too much a tough. | How much you pay? ... Well, then we're plenty tough. - Marx
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

iceport wrote: Tue Jun 22, 2021 5:05 pm
McQ wrote: Tue Jun 22, 2021 1:17 pm
Definitely gonna need to revise that section. Let me explain why the section is there and why I chose a provocative title. The section provides a vehicle to make the following claim: “It is only a matter of time before “tax-free” Roth distributions join “tax free” municipal bond interest in the calculation of MAGI used to determine whether Medicare IRMAA surcharges will be imposed.”
All Bogleheads who approach Roth conversions as a means of providing future tax-favored income need to accept this risk: that Roth distributions will cease to be tax-favored with respect to IRMAA.
That tax law change won’t matter if the converted funds are truly surplus, intended for beneficiaries and not for distribution while alive. Here the risk is that Congress imposes a stacking rule on heirs of a Roth account, as is now done for international income. Under stacking, a $100,000 Roth distribution will be tax-free, as before; but it will be placed at the bottom of the income stack, such that the next dollar of ordinary income will be taxed as if it were the $100,001st dollar of income, in the bracket that pertains to that income. The Roth will still be tax free to heirs, but the tax savings will be drastically reduced, since it averts tax at 0%, 10%, and 12% (=the bottom $100,000 in the current tax structure MFJ).
Iceport, can you help by suggesting a better title / positioning for the section, which would still allow these points to be made? I find far too much blithe optimism concerning the tax risk inherent in Roth conversions, and that's what I wish to combat.
Professor McQuarrie,

Thank you for the explanation. And allow me to apologize for the intemperate tone of some of my comments. (An uncle once described how he thought he had outgrown a lousy personality; unfortunately, I never outgrew mine.)

The aspect of your analysis that triggered my ire was the lack of any real qualification of the results, which, as I envision it, would acknowledge the uncertainties inherent in any kind of tax planning, really. I think such a qualification is an essential component of your analysis, which is amazingly comprehensive and detailed, and has enormous value. Nevertheless, the topic needs to be addressed with a degree of humility, because the future is uncertain.

(As an aside, the same kind of thing used to make me fume when a certain poster who had mastered the art of back-testing and graphing the results would adamantly urge everyone who’d listen to adopt massive small value and REIT tilts. Like the pied piper, folks followed his lead, clamoring for the superior performance, which, it was strongly implied, was a practical certainty. Well, that argument has become more muted, lately.)

Personally, I agree with and fully support your intentions. I don’t think Roth conversions should be considered an obvious slam-dunk. By the same token, I don’t think even the best, most complete and flawless deterministic analysis of the tax benefits or detriments of Roth conversions should be treated as the last word on the subject, but rather as a valuable and powerful tool to assist in making informed choices in the face of tax laws that are confusing today *and* that have an uncertain future.

From your further explanation, it seems that you are not so much dismissive of what I’ve come to know as tax diversification (a term which some folks think is semantically incorrect, under a narrower definition that you apply in the paper), as you are critical of folks focusing on one type of tax risk and overlooking others. Really, your point could possibly be better made by embracing the notion that the future is unknowable, the concept of tax risk. That’s where your claim fits in: as a counterpoint to the thinking that tax risk only cuts one way with respect to Roth accounts. As you so correctly point out, new laws could just as easily undermine the tax benefit of Roth accounts. The level of detail with which you cover the possible treatment of Roth withdrawals provides a vivid illustration of a scenario in which tax risk shows up in an unexpected form.

However, it’s misguided, in my opinion, to cross the line from identifying what’s possible into making a prediction. As tempting as it is, and as well-founded as it might be, and as interesting as it might be to formulate, a prediction like that should not form the sole basis of action. Similarly, I think it’s misguided to position your portfolio to benefit only in the event of massive tax hikes, based on such a prediction. The whole point is that nobody really knows what the tax laws will look like in 10, 15 or 20 years from now. So I don’t agree with stating your claim so confidently as a prediction.

(I’ve predicted for the better part of 20 years now that federal income taxes only have one direction to go from here. Obviously, I’ve been wrong for a long, long time. Thankfully, I had the sense not to act on my prediction.)

To satisfy my own perspective, I’d love to see you incorporate the idea of not putting all one’s eggs in one tax treatment basket, or tax diversification, if you will. However, while directly related, I think your real concern is that folks aren’t taking the full scope of tax risk into account, but are instead only picking and choosing which forms of risk they want to address — based on imperfect and biased predictions. Your topic is highlighting tax uncertainty/tax risk, not attacking tax diversification.

If one truly accepts that they cannot know the future, it’s less likely they’ll fully commit to a grossly lopsided investing approach.

I hope this makes a little sense, that I'm explaining myself well enough.
Iceport: you explain yourself very well, and I want to thank you for exactly the feedback I requested. I see now that I can't confidently predict "in the near future Roth distributions will count toward MAGI," and still be true to my general point, which is that the future tax treatment of Roth accounts, like all other accounts, must always be uncertain. Very helpful, and I will be re-reading your posts when I revise
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

DSBH wrote: Tue Jun 22, 2021 6:20 pm I downloaded the spreadsheet and am still studying the paper, and then got confused on this one:
-------------------------
"...
2. Nor is it all that helpful to pay the tax on conversion from outside funds;
...
The analyst simply assumes that Rob and Sue have some hundreds of thousands of dollars set aside. This assumption goes unexamined for good reason: the logic of arranging enough conversions, at a low enough tax rate, to make a material difference virtually requires the existence of that large sum of taxable dollars sitting around doing nothing.
...
For these reasons, the base case in the analysis of intertemporal tax arbitrage must be that tax on the conversion is paid from the converted funds.
Payment from some other source is one of many interesting alternatives to be explored but cannot be the base case."
-------------------------
As the spreadsheet shows, Rob and Sue made a 100K conversion in 2020, paid 24K in taxes and ended up with 76K in the Roth account. Table 3 shows that the floor of the 24% bracket is 198K, so it looks like Rob and Sue got quite a bit of money in their Taxable account to live without having to touch either tIRA or Roth account.

Does it appear from the example that it is not helpful to pay the tax on conversion from the converted funds (instead of outside funds)?
Hi: not sure I quite follow your concern, but I think I can address the thrust of your comment. Paying the conversion tax from outside funds doesn't hurt, and helps a bit: now you have $100,000 in Roth, not $76,000. But in calculating ROI, $100,000 is the base, not $76,000. See Table 4, middle column versus left column, bottom panel. So the ROI doesn't change much.
PS: and yes, I did assume that Rob and Sue could scare up $24,000 from taxable funds, even though they don't have that much available in taxable funds (an assumption relaxed in later analyses).
PPS: Sue is still employed in 2020, so they had $200,000 of wage income
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

VanGar+Goyle wrote: Tue Jun 22, 2021 7:53 pm
I wonder where the median Boglehead fits in the larger universe of individuals with substantial IRA or 401(k) balances? Reports by Fidelity indicate that even a $1 million dollar TDA balance puts a saver in the 98th to 99th percentile among the 16 million or so accounts they administer.
Vanguard may be a more appropriate place to check for Bogleheads.
Their how-america-invests-2020.pdf report is 58 pages long, breaks data down by average and quartiles, but not as detailed.
It is subtitled 'A look at the personal investing behavior of Vanguard’s 5 million retail households'.
Average figures: Household size is two persons. Portfolio holdings, on average, are two accounts and about five investments,
with average account balance of $352,200. The median account balance was $60,900, so not close to $1 million.

Is there a better Vanguard report, perhaps more detailed and institutional oriented?
The How America Saves 2021 report does cover institutional accounts, with 29% of balances more than $100,000.
At age 55–64, the average balance was $232,379, the median balance was $84,714, so still not detailed enough and not that close to $1 million.
Search for the equivalent Fidelity reports, issued quarterly. They do call out the count of 401(k) or IRA millionaires in their system. But as an earlier poster noted, it is a count of accounts, not of persons, so an underestimate.
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

cas wrote: Tue Jun 22, 2021 7:12 am
curmudgeon wrote: Tue Jun 22, 2021 12:41 am Interestingly enough, I think there is still actually a decent return to be had from making that contrived $100K conversion at age 65. Rob and Sue will be netting about $170K after tax on that $200K income, when they draw it from TDA after age 66. If, at age 70 and 71 when they are drawing $100K SS but not doing RMDs, they draw $50K from the Roth each year, they will only need to pull $25K from the TDA those years to get to $170K spendable income. They'll save about $16K because for those two years they will avoid much of the "tax torpedo" of taxation of their SS benefits. $16K is not a bad return for doing a $100K Roth conversion. It's worth noting that this does not scale much further. Once RMDs start, SS will be fully taxed unless they've done some really large conversions.
It is true that the age 70/71 "avoid tax torpedo" tactic can't be expected to extend beyond those two years.

However, the example points out the potential usefulness of Rob and Sue having the availability of at least some assets in Roth.

In another very similar example, in 2008 and 2020, Congress waived RMDs due to economic circumstances. While those one-off waivers can't be assumed to ever happen again, *if* an RMD waiver ever happened again, Rob and Sue would find having Roth assets very useful. Roth assets would give them the flexibility to manage their taxable income and, once again, avoid their SS tax torpedo. (They would have to have converted more than $100K, though, in order to still have Roth assets *after* they used Roth assets to avoid the SS tax torpedo at age 70/71.)

In another example of where Rob and Sue might find Roth assets useful in managing a potential "tax torpedo," it has been niggling at me that Rob and Sue look like the perfect candidates to be flirting with the Alternative Minimum Tax "tax hump" under the pre-2018/post-2025 tax law. Like the SS tax hump, the AMT tax hump is very dependent on individual circumstances, but its 32.5% and 35% marginal rate "bump" tends to cover quite a big span in what is now the 22% and 24% brackets under the conventional tax system.***

Dr. McQuarrie already discusses an alternative scenario where the conventional tax brackets revert to their pre-2018 form, as currently scheduled. He also discusses that Rob and Sue might be living in California in their "rather fine house" for all or part of their retirement. Income in Rob and Sue's range + high state taxes + high property taxes is a stereotypical scenario where the AMT tax system would kick in pre-2018. But someone who knows more than I about California income and property tax law would need to judge whether Rob and Sue, as retirees and having apparently owned their house for quite a while, would have high enough state taxes and property taxes to be at risk of switching from the conventional tax system to the AMT tax system.

*If* Rob and Sue are determined to be seriously flirting with AMT in their retirement years, then Roth conversions come into play in two ways:
1. Could some reasonably attainable amount of Roth conversions pull their RMDs down just enough for them to avoid triggering AMT for some number of years? Or at least avoid the 32.5% or 35% AMT marginal rates on the extra amount of RMD that is there without Roth conversion and isn't there with Roth conversion? (Similar to the way that Roth conversions helped them avoid an additional IRMAA tier for some years in the "base case".)
2. Would the availability of Roth assets help them avoid AMT in years where they might need a bit of extra income beyond what is coming from SS + RMDs? Or at least avoid the 32.5% or 35% AMT marginal rates on the extra withdrawn amount? (For example, if they need to replace the roof or siding on their "rather fine" house or if they wanted to take a special vacation with their young adult children.)


***For more info on the general concept of the AMT "tax hump" see the "AMT bump zone" graph near the bottom of Michael Kitces 2014 article "Evaluating Exposure To The Alternative Minimum Tax And Strategies To Reduce The AMT Bite" . And, yes, that is a 2014 article, so all the numbers are dated and calculations would need to be adjusted for inflation. But the general concept holds for the post-2025 tax law as currently scheduled to return.
Ahh, the AMT in California, from the Bush tax cuts forward: long were we penned in that dungeon, where so much tax was painfully extracted from our flesh of wages.
And indeed, if the AMT reverts to the inflation-adjusted equivalent of what it was pre-TCJA, couples like Rob and Sue, living in a state like California, may well fall into it again in retirement. And if they do, then a Roth conversion will pay off earlier and to a greater extent.
But remember, it was going to pay off anyway, just slowly, and not in a large amount until after age 100 (Table alpha.3, figure alpha.1, pp. 22-23].
And last, nightmare scenario: Congress puts a line into form 6251 that adds back Roth distributions in determining tentative AMTI. Then Rob and Sue, to use a technical term, are screwed, conversion or no.
User avatar
WoodSpinner
Posts: 2109
Joined: Mon Feb 27, 2017 1:15 pm

Re: Roth Conversions - McQuarrie study

Post by WoodSpinner »

McQ wrote: Tue Jun 22, 2021 11:25 am Woodspinner—thank you for that detailed reply. Over the decades, I’ve endured many peer reviews from scholarly journals that were far less informative (or constructive).
Your comment that you were never a super saver but still ended up with (millions) in your TDA is well worth restating. It truly is amazing what monthly contributions, invested in the Boglehead way, can mount up to over a 40 year working career.
Now, since I put you off in my first reply (can’t give individual finance advice, yada yada), please let me make amends with a few specific comments on what you have done.
Thanks for the kind words. Frankly without the help of this forum and especially #Celia, I would have been a lost babe in the Retirement woods. I had been on autopilot for many years and hadn’t a real clue if or when I could retire.
1. You built a spreadsheet to capture your likely situation. ‘Nuff said—those who build their own spreadsheets, they shall be saved. Those who can’t or won’t, will have to find a friend, or pay an advisor—the equivalent of extra tax drag. However, building your own custom spreadsheet is, I believe you will agree, not that easy to do.
Agreed!

Building out our retirement model helped us really understand how our Retirement will financially work. It gave us the confidence to actually retire — We simply couldn’t trust other peoples tools and models since we didn’t fully understand all of the assumptions and trade offs.

That said, #Bigfoot48 has put together a pretty impressive Retirement Portfolio Model that is free and available through our WIKI.

Planning out Roth Conversions is a very challenging task and we need all of the help, insight and tools that we can get our hands on. I don’t know much about the academic world but this is a real world problem that can use the intellect and rigor of the academic community.
2. I discourage you from assuming a specific age at death, the more so jointly. And your chosen ages—90 for you, 89 for wife—are far too young for the affluent, conscientious demographic in which you reside. Playing things out until 95 for you (if you are male) and 105 for your wife (if female) would be a worthwhile exercise.
FWIW, I am male and my wife female.

This puzzles me a bit given the actuarial probabilities (under 15% for me, under 8% for my wife). Source.

I will need to update my model to complete the analysis. I would expect the IRA to be drained and After-Tax and Roth to grow significantly.
a. Sorry, I live among the woke, and dare not infer gender from terms like ‘wife’
b. It probably won’t change anything, but … will the heirs still be in the 12% bracket if your spouse does not pass until 105?
I suspect it won’t make a difference, but it’s just a reasonable guess. At 105 the RMDs would have almost eliminated the IRA and they would be inheriting After-Tax and Roth accounts.
c. Aside: I have no training in financial economics and never developed a loyalty to net present value calculations. They seem to enforce selection of an ending date—i.e., your death at 90. I prefer future value constructions, where the rate of appreciation can be varied at will, and the scenario played out to an arbitrarily late date, and then roots computed to get at ROI. FYI: not a few life insurance companies run their illustrations out to 120 these days.
That’s easy enough but it really doesn’t change the key metrics. Significant Roth conversions still pay off, even more so if one of us passes early.

My model allows yearly changes in AA, Expected Returns, Inflation etc. As other’s have mentioned we really could use some more insights into the sensitivity analysis around tax rates and expected returns. This isn’t something that I have tackled in my modeling and would love to benefit from others work in this area.
d. NPV analyses also demand selection of a discount rate. Your choice of rate is blue chip respectable; but again, I’d rather let the future play out, under annualized assumptions or pasting in actual historical data, and see how things change, when the rates change.


3. Very smart that you model IRMAA. When I applied the learning from the paper back to my personal situation, it became apparent that we were overly vulnerable to entering the second (third) IRMAA level, and now I am noodling about how to stay in the first IRMAA level—which as a previous poster remarked, is almost impossible to avoid if you are going to live large in retirement. Which I fully intend to do.
In my analysis, IRMAA (even at the third tier) is a fairly small number (under 5%) of the overall Taxes owed. Are you seeing different ratios in your analysis? This might be an interesting rubric to help people asses the impact of IRMAA on their Roth conversion plans.
4. Per your statement, QCD are the all purpose buffer. Find yourself $5268 into an IRMAA bracket, marginal tax rate on the $5268 = 60%, no problem—make a QCD of $5500 that year. Typically less than 2.5% of income, not heroically charitable, doable if you have any charitable recipients in mind.
Since we are charitably inclined our current plan is to target QCDs at 25% of RMDs. We will ramp this up if future returns are significantly higher than expected. This becomes a great control valve for tax and estate planning.
5. Last, play around with the decision to convert amounts through IRMAA #3. Still working on that new paper (“How to Cope with IRMAA”), but very preliminary analyses cause me to say, hey, better give that a second look. Yes, certainly, always go to the top of whatever IRMAA bracket applies—but to willfully push beyond, into the next bracket, or the bracket beyond, is not a decision I can as yet support.
This goes back to my earlier comment, IRMAA is expected to be a relatively minor portion of our overall taxes.

I suspect that this will be true for others with SS, Pension and larger IRAs — looking forward to your analysis.

Thanks

WoodSpinner
User avatar
FiveK
Posts: 11763
Joined: Sun Mar 16, 2014 2:43 pm

Re: Roth Conversions - McQuarrie study

Post by FiveK »

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. ....
Quoting in case this was overlooked.
sc9182
Posts: 587
Joined: Wed Aug 17, 2016 7:43 pm

Re: Roth Conversions - McQuarrie study

Post by sc9182 »

FiveK wrote: Tue Jun 22, 2021 11:37 pm
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 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. ....
Quoting in case this was overlooked.
This situation applies to highly successful BH crowd who lives on single/double pensions, double max delayed-SS, some deferred comp, part time job, some inheritance, and/or quite large brokerage accounts ; and live modestly, without ever needing/wanting the TDA amounts/withdrawals (except for pesky RMDs). You couldn’t convert sufficient into Roth due to most of the above income sources pushing to large enough tax/IRMAA brackets, Now, all TDA doing is sitting around and keep growing/compounding over 20-30-40-50 or more years — and good to have growing balances - but TDA withdrawals/conversions (or RMDs) could push into higher marginal tax brackets!

Consider if someone made Roth conversions prior to TCJA, they paid possibly 25% instead of 22% ! You already paid 1% extra-tax/year even if you were pushed into the now-24% bracket due to market caused growth of your TDA! Or, a COVID Crash or two, your withdrawals or RMDs could be lot lower possibly pushing you back into 22% or even lower brackets ..

If 34% or 50% market crashes were to occur — I would rather have 50% of $1000 in TDA rather than 50% of $780 in Roth., and worse yet, needing to withdraw some/fixed amount from either of these portfolios at such time !! You would rather prefer to be in TDA (and larger balance that is). Larger portfolios (TDA), is an insurance against worse-market conditions— allowing portfolio longevity and successful retirement withdrawals.

Then again - some highly successful BH’s don’t ever need TDA/Roth monies ever to live on (nor much interested in QCDs, or not ever needing monies for large health-care costs tax-efficiently coming out of TDA ..) — Roth May perform better - because RMDs on TDA are forcing withdrawals from even during market crashes — thus chipping away harshly at ‘then’ beaten down portfolio. Roths, not needing RMDs (in your joint lifetime), and you strictly not-needing withdrawals from it., Roth accounts won’t get beaten down further during market crashes. — this could apply to select highly successful BH’s who have multiple sources stable incomes (and inflation is not rampant, and health is top-notch forever)

Or worse yet - considering 70s style inflation — I would rather have $1000 in TDA than $780 in Roth ..

Hard to predict future — like someone says..
Last edited by sc9182 on Wed Jun 23, 2021 7:51 am, edited 3 times in total.
Chip
Posts: 3439
Joined: Wed Feb 21, 2007 4:57 am

Re: Roth Conversions - McQuarrie study

Post by Chip »

McQ wrote: Mon Jun 21, 2021 5:55 pm It raises a general question that must have come up in past Boglehead forums (which I'd like to peruse if anyone has links to hand): what percentage of income needs to be replaced in retirement? And does it vary by level of affluence? (of course).
From my POV there is quite a bit of disdain here for any attempt to quantify a retirement spending level as a fixed percentage (or percentage range) of pre-retirement gross income. It's an ill-advised shortcut often seen on brokerage websites and in retirement planning articles meant to be read in five minutes or less.

We usually tell people here to figure out their current spending, then make adjustments similar to those mentioned by livesoft, plus anticipated changes in spending patterns (e.g. no commuting/work wardrobe costs, more travel, higher medical expenses, home remodeling).

I am one of many here that track spending (perhaps with a degree of OCD :D). If I compare pre-retirement, inflation-adjusted, after-tax spending to 20 years of post-retirement, we have spent 20% more post-retirement. This doesn't include 2020 covid-restricted spending, which I suspect will be more than made up in 2021 and 2022. My explanation for this is that we have more time in retirement to do things that cost money.

Of course income taxes are much lower post-retirement, even with the Roth conversions.

Here's a recent thread where several members weighed in on their pre and post retirement spending.
User avatar
tadamsmar
Posts: 9368
Joined: Mon May 07, 2007 12:33 pm

Re: Roth Conversions - McQuarrie study

Post by tadamsmar »

The OP link says: "You should therefore consider a Roth conversion for a one-time lump sum withdrawal later in retirement (such as a down payment on entering a nursing home) or what you anticipate leaving as a bequest."

But, in some cases, I think even the down payment would be tax deductible. This is an argument against Roth conversions.

It depends on whether you plan to stick it out in your own home till you qualify for a tax deduction on the down payment.
FlamePoint
Posts: 64
Joined: Wed Nov 11, 2020 10:45 pm

Re: Roth Conversions - McQuarrie study

Post by FlamePoint »

I’ve been reading this thread with a great deal of interest given our current situation. Even with all the good information out there I still feel like there’s a large hole/gap in available tools to model Roth conversions.

Through a series of what I’ll call fortunate events we ended up with rather large TDA upon retirement last year at age 58; started investing in our 20’s, full employment over a 30 year career with same employer, generous compensation plan and retirement benefits, pension that was taken as a lump sum. We are currently sitting on a combined $4.5 M retirement bucket of funds; $3.6M tax deferred, $450K Roth, and the remainder taxable. We will both be receiving close to max SS, and plan to take it at age 70. Current expenses are around $96K per year. SS will cover the vast majority of our expenses from age 70 on.

All the modeling I’ve done to date suggests we will benefit from Roth conversions between now and age 72. Unless we really start to ramp up our spending, it looks like our Roth funds will never be touched and will be our legacy funds for the kids.

At this point, I’m mainly concerned about the tax impact once one of us passes away. Without conversions taxes as MFJ start to really jump once RMD’s kick in. The single taxes brackets are even worse, so there appears to be a significant benefit to converting at least 50% of the TDA over the next 10 years or so.

Certainly a good problem to have, but challenging to determine best course of action given all the future unknowns (ie. inflation rate, tax law changes, our health, market fluctuations, etc).

I continue to watch and learn.
DSBH
Posts: 249
Joined: Tue Jul 21, 2020 3:31 pm
Location: Texas

Re: Roth Conversions - McQuarrie study

Post by DSBH »

McQ wrote: Tue Jun 22, 2021 10:52 pm
DSBH wrote: Tue Jun 22, 2021 6:20 pm I downloaded the spreadsheet and am still studying the paper, and then got confused on this one:
-------------------------
"...
2. Nor is it all that helpful to pay the tax on conversion from outside funds;
...
The analyst simply assumes that Rob and Sue have some hundreds of thousands of dollars set aside. This assumption goes unexamined for good reason: the logic of arranging enough conversions, at a low enough tax rate, to make a material difference virtually requires the existence of that large sum of taxable dollars sitting around doing nothing.
...
For these reasons, the base case in the analysis of intertemporal tax arbitrage must be that tax on the conversion is paid from the converted funds.
Payment from some other source is one of many interesting alternatives to be explored but cannot be the base case."
-------------------------
As the spreadsheet shows, Rob and Sue made a 100K conversion in 2020, paid 24K in taxes and ended up with 76K in the Roth account. Table 3 shows that the floor of the 24% bracket is 198K, so it looks like Rob and Sue got quite a bit of money in their Taxable account to live without having to touch either tIRA or Roth account.

Does it appear from the example that it is not helpful to pay the tax on conversion from the converted funds (instead of outside funds)?
Hi: not sure I quite follow your concern, but I think I can address the thrust of your comment. Paying the conversion tax from outside funds doesn't hurt, and helps a bit: now you have $100,000 in Roth, not $76,000. But in calculating ROI, $100,000 is the base, not $76,000. See Table 4, middle column versus left column, bottom panel. So the ROI doesn't change much.
PS: and yes, I did assume that Rob and Sue could scare up $24,000 from taxable funds, even though they don't have that much available in taxable funds (an assumption relaxed in later analyses).
PPS: Sue is still employed in 2020, so they had $200,000 of wage income
It's not necessarily a concern, I'd just like to understand the study case because we are somewhat similar to Rob and Sue - except that we're both retired as of Jan 2020 - trying to decide how much to convert each year at least till 72.

So it appears that Rob and Sue have 200K from her job and outside taxable funds (enough to exceed the 24% bracket AGI floor) to pay for the tax on 100K of Roth conversion in 2020, but decide to pay for it using the money from the converted funds.

I haven't yet run across any ROI work on that decision in the paper, but I am still working on my review. As noted in the paper, the conventional wisdom is that you get more benefit from a Roth IRA conversion if you can pay for the tax using outside funds, not to mention that the conversion could be more beneficial if you have no other significant income when you convert.

Thank you for your work and your time.
John C. Bogle: "Never confuse genius with luck and a bull market".
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

dodecahedron wrote: Tue Jun 22, 2021 5:25 am
curmudgeon wrote: Tue Jun 22, 2021 12:41 am I've been rather critical of the paper, but I think the point that quite large TDAs may see limited value from Roth conversions is well worth exploring. It might be worth doing in the context of someone who is single, as that removes the question of what happens at the transition from MFJ to Single tax brackets. I really like the approach of including exploration of what happens if tax rates/brackets change; sensitivity analysis along this point is important to make sure your Roth conversion plan isn't teetering on some rare little odd spike in the tax code. It's a near certainty that tax laws will change in some direction over the next 30 years.
+1 to the need for sensitivity analysis.

Not only is it a near certainty that tax laws will change in some direction over the next 30 years, even in the absence of tax law change, there is a huge amount of uncertainty about future path of one's AGI since the future stream of returns on investment is also highly unpredictable.

My late husband did quite a bit of Roth conversion in the decade before his unexpected death in his late 50s eight years ago. As a finance professor with a highly volatile but often very substantial side consulting income, he did so opportunistically in years when AGI dipped. He was also relatively very aggressive (and very lucky!) in his investments in the Roth, so ex post, it has turned out quite well for me as surviving spouse, despite the fact that tax bracket rate schedules were higher back when he was converting than they are now.

My portfolio now is roughly 30% Roth, 20% tax-deferred, 50% taxable. My tax-deferred is entirely in fixed income (TIAA Trad SRA yielding 3%, which is their guaranteed minimum) so very predictable stream of future RMDs for me [unless interest rates ever mean-revert!].

But most folks with large proportion of tax-deferred typically would have some equities in their tax-deferred and have far more unpredictable future RMDs than I do. Also, I am reasonably certain about my tax filing status staying the same (Single) over the remainder of my life.

At this point in my life (age 68, with RMDs four years away), I have crunched the numbers relevant to my situation going forward and my personal conclusions about whether and to what extent additional conversions make sense for me generally comport with most of the broad conclusions of the paper as stated in the abstract.
paper's abstract wrote: • Future tax rates need not be higher for a conversion to pay off;
• Nor is it all that helpful to pay the tax on conversion from outside funds;
• Nor are Roth conversions especially beneficial for top bracket taxpayers as compared to middle class taxpayers;
• Rather, the greatest benefit accrues to taxpayers who can make the conversion partly in the zero percent tax bracket, i.e., during a year with no other taxable income.
Edited to add: at least in my case, it is a no-brainer to use outside funds to pay the tax on any conversions I do, since I have a large taxable account and there is no reason not to do so. Barring major unanticipated future disasters (e.g., need for many years of expensive LTC), my Roth will probably go as a bequest to my daughters. I live in a state (NY) with an estate tax that kicks in (with a cliff!) well before the federal estate tax does. Kudos to forum member bsteiner for frequently reminding me of this. As he has pointed out, Roth conversions make particular sense if there is any possibility of state estate taxes kicking in.

That said, any additional Roth conversions that I do will likely be "tinkering around the margins." For 2021, I have a pretty good handle on my effective marginal tax rate (taking into account many things besides my official "bracket") and believe that I can convert about $12K at an effective marginal tax rate of 18.5% (Fed) which would also be state tax free (NY allows up to $20K per year in tax free retirement distributions). If I go much beyond that, I quickly approach a range where the effective federal marginal tax rate would be 49.5% (hot spot on the heat map) and also start to approach IRMAA issues, issues with qualifying income for senior property tax breaks, etc.

$12K is a tiny "drop in the bucket" relative to the size of my tax-deferred account, but whatever. I will do it in the spirit of stewardship as a tribute to my beloved late husband, since it seems a no-brainer.

I do have the ability to manage my future taxable RMDs to an unusual extent, since a significant portion of my current budget is charitable giving, which will take the form of QCDs once I am over 70 1/2. (Currently funding my charitable giving by donating appreciated securities and itemizing deductions.) I also still work (very part-time) and may continue to do so after 72, which will give me further options on delaying RMDs (since I have rolled all the tax-deferred into a 403b at my employer.) Another unusual factor for me is that it made no sense to wait until 70 to file for SS, since widow's benefits maximize at FRA and do not get delayed retirement credits, so I am already collecting SS. Also, my AGI includes a large amount of qualified dividend income due to my large taxable account. So I have many "moving parts" to my analysis.

Bottom line: the amount of money saved by doing additional Roth conversions at this stage of my life is probably not worth the time I have spent thinking about it (except that it keeps my brain engaged!) Kind of like doing a crossword puzzle.
I applaud your good fortune. A word of advice from an academic who also has access to a TIAA traditional account (not available to most Bogleheads; akin to, but far superior to the stable value funds from profit-seeking insurance companies that are widely available).
You have to plan on living a very long time. Therefore, the graded TIAA annuity is the way to go, as an ersatz inflation-adjusted annuity. Maximum guarantee period will be set by the applicable life expectancy (not a CPA, can’t tell you what will apply in your situation). I would take the maximum guarantee, for your daughters’ sake. As age 72 approaches, get TIAA on the phone and put them through the paces on your options.
The annuitized TIAA income may well cover all your ordinary day to day living expenses, leaving you free to deploy the remainder of your income / wealth to charity, and to grandchildren (do bone up on the five-year 529 gift rules).
Let it, along with social security, be the totally safe foundation of your post 72 income (annuitize in the calendar year you turn 71, and it won’t have an RMD, I believe).
And then, enjoy your good fortune. You can’t take it with you, and how much do your heirs really need?
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

FlamePoint wrote: Wed Jun 23, 2021 11:12 am I’ve been reading this thread with a great deal of interest given our current situation. Even with all the good information out there I still feel like there’s a large hole/gap in available tools to model Roth conversions.

Through a series of what I’ll call fortunate events we ended up with rather large TDA upon retirement last year at age 58; started investing in our 20’s, full employment over a 30 year career with same employer, generous compensation plan and retirement benefits, pension that was taken as a lump sum. We are currently sitting on a combined $4.5 M retirement bucket of funds; $3.6M tax deferred, $450K Roth, and the remainder taxable. We will both be receiving close to max SS, and plan to take it at age 70. Current expenses are around $96K per year. SS will cover the vast majority of our expenses from age 70 on.

All the modeling I’ve done to date suggests we will benefit from Roth conversions between now and age 72. Unless we really start to ramp up our spending, it looks like our Roth funds will never be touched and will be our legacy funds for the kids.

At this point, I’m mainly concerned about the tax impact once one of us passes away. Without conversions taxes as MFJ start to really jump once RMD’s kick in. The single taxes brackets are even worse, so there appears to be a significant benefit to converting at least 50% of the TDA over the next 10 years or so.

Certainly a good problem to have, but challenging to determine best course of action given all the future unknowns (ie. inflation rate, tax law changes, our health, market fluctuations, etc).

I continue to watch and learn.
Dear Flamepoint: quite a few people in this forum and elsewhere have mentioned your fear: What happens to the survivor, when most income continues, but it is now taxed as a single, where the bracket floors kick in at half the MFJ dollars? I heard it so many times that I started to build a spreadsheet to quantify what might be called "the widow tax hit." Thus far the analysis is about as well-developed and coordinated as a 3-week old puppy, so please don't put too much weight on the next paragraph.
Long story short: it is not to be feared. The key is to evaluate the disposable income available to the widow (after tax, after Medicare), and *NOT* the dollars paid in tax or the average tax rate. Next, if the deceased spent about 25% of couple disposable income while alive (as in, like, eating), then the widow only has to cobble together 75% of the disposable income the couple had to maintain the same lifestyle. And analyses thus far show that, for couple income in the 22% through the 24% brackets ($100K - $350K), she does, despite increases in dollars paid in tax and tax rate, consequent to being a single.
I'll probably write a paper about it if the analyses hold up; and if I do, I'll post a link here on a BH forum. Just a caution: I move at an academic pace, which compares to glacial as my running speed compares to that of Usain Bolt.
User avatar
FiveK
Posts: 11763
Joined: Sun Mar 16, 2014 2:43 pm

Re: Roth Conversions - McQuarrie study

Post by FiveK »

McQ wrote: Wed Jun 23, 2021 2:26 pmThe key is to evaluate the disposable income available to the widow (after tax, after Medicare), and *NOT* the dollars paid in tax or the average tax rate.
Yes, spendable after-tax amount is a good metric.

One gets to different results when following various paths down the assumption rabbit hole. E.g.,
- will the spending follow some Safe Withdrawal Ratio, or a fixed amount?
- does the spendable after-tax amount calculation extend to heirs, and if so
- are the heirs charities (and thus the inheritance is not subject to tax), or people subject to the 10 year distribution rule and thus paying the marginal rate for those distributions on top of other income?
Northern Flicker
Posts: 8037
Joined: Fri Apr 10, 2015 12:29 am

Re: Roth Conversions - McQuarrie study

Post by Northern Flicker »

FiveK wrote: The paper continues to compare current marginal vs. future effective rates instead of the correct comparison between current and future marginal rates.
Both are incorrect. There is no single future marginal rate in retirement. It varies significantly from year to year based on how much you spend in a given year and based on what assets are used to generate income. Based on personal experience, I believe that the McQuarrie metric will result in better trad vs Roth decisions made while still working than the metric in the BH wiki.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
User avatar
FiveK
Posts: 11763
Joined: Sun Mar 16, 2014 2:43 pm

Re: Roth Conversions - McQuarrie study

Post by FiveK »

Northern Flicker wrote: Wed Jun 23, 2021 3:10 pm
FiveK wrote: The paper continues to compare current marginal vs. future effective rates instead of the correct comparison between current and future marginal rates.
Both are incorrect. There is no single future marginal rate in retirement. It varies significantly from year to year based on how much you spend in a given year and based on what assets are used to generate income. Based on personal experience, I believe that the McQuarrie metric will result in better trad vs Roth decisions made while still working than the metric in the BH wiki.
Do you have a numerical example in mind?
bsteiner
Posts: 5954
Joined: Sat Oct 20, 2012 9:39 pm
Location: NYC/NJ/FL

Re: Roth Conversions - McQuarrie study

Post by bsteiner »

McQ wrote: Wed Jun 23, 2021 2:26 pm ... quite a few people in this forum and elsewhere have mentioned your fear: What happens to the survivor, when most income continues, but it is now taxed as a single, where the bracket floors kick in at half the MFJ dollars? I heard it so many times that I started to build a spreadsheet to quantify what might be called "the widow tax hit." Thus far the analysis is about as well-developed and coordinated as a 3-week old puppy, so please don't put too much weight on the next paragraph.

Long story short: it is not to be feared. The key is to evaluate the disposable income available to the widow (after tax, after Medicare), and *NOT* the dollars paid in tax or the average tax rate. Next, if the deceased spent about 25% of couple disposable income while alive (as in, like, eating), then the widow only has to cobble together 75% of the disposable income the couple had to maintain the same lifestyle. And analyses thus far show that, for couple income in the 22% through the 24% brackets ($100K - $350K), she does, despite increases in dollars paid in tax and tax rate, consequent to being a single.
...
While the married brackets up to 32% are twice the width of the single brackets for 2018 through 2025, before 2018 that was only the case through the 15% bracket. The pre-2018 law is scheduled to return in 2026,

2017 brackets: https://taxfoundation.org/2017-tax-brackets/.

2021 brackets: https://taxfoundation.org/publications/ ... -brackets/.
Northern Flicker
Posts: 8037
Joined: Fri Apr 10, 2015 12:29 am

Re: Roth Conversions - McQuarrie study

Post by Northern Flicker »

FiveK wrote: Wed Jun 23, 2021 3:24 pm
Northern Flicker wrote: Wed Jun 23, 2021 3:10 pm
FiveK wrote: The paper continues to compare current marginal vs. future effective rates instead of the correct comparison between current and future marginal rates.
Both are incorrect. There is no single future marginal rate in retirement. It varies significantly from year to year based on how much you spend in a given year and based on what assets are used to generate income. Based on personal experience, I believe that the McQuarrie metric will result in better trad vs Roth decisions made while still working than the metric in the BH wiki.
Do you have a numerical example in mind?
Yes. My own marginal rate has not been consistent from year to year. Moreover, I avoid trad withdrawals/conversions in years when my marginal rate is higher, so the variable time series of marginal rates on trad income realization doesn't even track the variable time series of annual marginal rates.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
User avatar
FiveK
Posts: 11763
Joined: Sun Mar 16, 2014 2:43 pm

Re: Roth Conversions - McQuarrie study

Post by FiveK »

Northern Flicker wrote: Wed Jun 23, 2021 4:42 pm Yes. My own marginal rate has not been consistent from year to year. Moreover, I avoid trad withdrawals/conversions in years when my marginal rate is higher, so the variable time series of marginal rates on trad income realization doesn't even track the variable time series of annual marginal rates.
That does tend to support the idea that It’s Difficult to Make Predictions, Especially About the Future.
Northern Flicker
Posts: 8037
Joined: Fri Apr 10, 2015 12:29 am

Re: Roth Conversions - McQuarrie study

Post by Northern Flicker »

The difficulty of comparing to some hypothetical marginal rate in retirement is more than just the difficulty in predicting the future. Even if you have a good estimate of your average marginal tax rate averaged over your retirement years, because you have substantial control over the timing of trad IRA withdrawals, the average marginal tax rate of trad IRA withdrawals will typically be less than the average marginal tax rate.

Another wrinkle is IRMAA (adjustments to medicare premiums based on income), which need to be included in the marginal tax rate for this purpose for those with sufficiently high income to be affected by it.
Last edited by Northern Flicker on Thu Jun 24, 2021 7:02 pm, edited 1 time in total.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
Lee_WSP
Posts: 5338
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: Roth Conversions - McQuarrie study

Post by Lee_WSP »

bsteiner wrote: Sat Jun 19, 2021 10:14 am
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).
And even less if one is stuck with a SIMPLE IRA or even no employer sponsored plan. Or an SEP IRA as a non-owner employee.
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

Lee_WSP wrote: Thu Jun 24, 2021 1:09 pm
bsteiner wrote: Sat Jun 19, 2021 10:14 am
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).
And even less if one is stuck with a SIMPLE IRA or even no employer sponsored plan. Or an SEP IRA as a non-owner employee.
Good point. My wife and I have always lived under an institutional umbrella, with always a 401(k) or 403(b) plan, and employer matches to boot. I forget that's not every retirement saver.
But, if your TDA balances are smallish and taxable account balances large, and you intend to convert and pay the conversion tax from the taxable account funds, the relevant counterfactual would be to leave the money in the taxable account, in a low cost tax-managed index fund. That won't do as well over the long term as a Roth account, but there's no initial tax debit, no five-year rule, and the potential for a step up in basis at death, in which case the Roth advantage may be minuscule. Thinkers.
McQ
Posts: 109
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Roth Conversions - McQuarrie study

Post by McQ »

BTW, woodspinner and some others mentioned a checklist as a value add. I had prepared such a checklist but cut it from the paper for length reasons. Not sure if this is what folks had in mind, but here goes.

Here is how a Boglehead might approach the Roth conversion decision, assuming their personal situation aligns with the examples profiled in the paper (generally, anticipated retirement income between $150,000 and $500,000 in today’s dollars).
1. Make sure the funds to be converted and their future value are truly surplus. Remember that if these converted funds begin to be distributed annually in the 80s, the conversion may not meet its goals.
2. Stay focused: converted funds are designed to handle late in life financial emergencies and /or leave a larger after-tax bequest to heirs. Make sure you are comfortable with a payoff horizon that may fall between age 90 and 100.
a. If you or your spouse snorts at the idea that they might live so long, consider life insurance products as an alternative means to pass on a tax-free bequest.
3. Can you arrange your affairs to have no taxable income for one or more years apart from the conversion itself? Design the maximum tax-efficient conversion that can occur during that period. In most cases that maximum will be the lesser of your expected surplus TDA balance or the AGI at the ceiling of the 24% bracket, unless you have begun to take Medicare, in which case it will be the threshold where IRMAA surcharges begin.
a. For a mild boost to the payoff, pay the conversion tax from outside funds.
4. If you cannot arrange that most favorable circumstance, but can only “top off” a middling tax bracket with conversions, be aware that the payoff associated with a conversion under constant or similar future tax rates is going to be smaller and come slower.
5. Explore the alternatives to a Roth conversion, including back door Roth IRAs, insurance products of various types, and, most important, the flexibility and tax-managed potential of funds in a brokerage account maintained outside of any tax-deferred structure.
6. Do not rely on simple heuristics like the expected level of future tax rates. Do not express distant future results in nominal dollars. Run the numbers for yourself using a spreadsheet like the one at http://www.edwardfmcquarrie.com/?p=573.
Lee_WSP
Posts: 5338
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: Roth Conversions - McQuarrie study

Post by Lee_WSP »

McQ wrote: Thu Jun 24, 2021 3:45 pm
Lee_WSP wrote: Thu Jun 24, 2021 1:09 pm
bsteiner wrote: Sat Jun 19, 2021 10:14 am
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).
And even less if one is stuck with a SIMPLE IRA or even no employer sponsored plan. Or an SEP IRA as a non-owner employee.
Good point. My wife and I have always lived under an institutional umbrella, with always a 401(k) or 403(b) plan, and employer matches to boot. I forget that's not every retirement saver.
But, if your TDA balances are smallish and taxable account balances large, and you intend to convert and pay the conversion tax from the taxable account funds, the relevant counterfactual would be to leave the money in the taxable account, in a low cost tax-managed index fund. That won't do as well over the long term as a Roth account, but there's no initial tax debit, no five-year rule, and the potential for a step up in basis at death, in which case the Roth advantage may be minuscule. Thinkers.
Not exactly true IMO. The dividend tax drag (even a qualified dividend) is noticeable even if you are 100% equities (I used to think otherwise, but I did the research and ran the numbers and the difference is a noticeable expense ratio (I forget what it came out to, but it was greater than the expense ratio of the funds looked at)). I'd say it's quite noticeable over ten years, and certainly 30. But that is open for interpretation.

It's a ten year rule for designated beneficiaries, btw. If it goes to your estate, it's either a five year or if you started, the life expectancy table of the decedent is used.

As for the TDA balances being small; your paper uses a 30 year contribution period, but simply extending it by an additional five to ten years and maximal SIMPLE IRA contributions will easily reach/exceed the amounts of maximized 401k contributions over 30 years.

Obviously this is only a problem for super savers, but nevertheless some food for thought.
User avatar
jeffyscott
Posts: 9753
Joined: Tue Feb 27, 2007 9:12 am
Location: Wisconsin

Re: Roth Conversions - McQuarrie study

Post by jeffyscott »

Lee_WSP wrote: Thu Jun 24, 2021 1:09 pm
bsteiner wrote: Sat Jun 19, 2021 10:14 am
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).
And even less if one is stuck with a SIMPLE IRA or even no employer sponsored plan. Or an SEP IRA as a non-owner employee.
OTOH, those of us that had more modest incomes can have a lifetime average savings rate of 30% and yet have retired with a very small taxable account. At retirement we had ~67% tax deferred, 26% Roth, and 7% taxable (mostly I and EE bonds).

Our ~30% average was very uneven, nearly all of it was saved during the 20 year period that we had 2 incomes. We had 3 children while in our 20s and spouse did not work until she was about 30. All the money that we contributed to savings comes, not coincidentally, to 110% of her lifetime earnings.

Those, like us, who live on << $100K per year can also do conversions at 12% year after year, without having a large taxable account.
The two greatest enemies of the equity fund investor are expenses and emotions. ― John C. Bogle
N.Y.Cab
Posts: 121
Joined: Wed Feb 07, 2018 8:46 pm

Re: Roth Conversions - McQuarrie study

Post by N.Y.Cab »

I’ve been using i-ORP to compare Roth conversions at different tax bracket ceilings and was surprised to see no change in disposable income. There is some increase in total plan value so I see no problem with Roth conversions to the top of 22% while it is available.
User avatar
WoodSpinner
Posts: 2109
Joined: Mon Feb 27, 2017 1:15 pm

Re: Roth Conversions - McQuarrie study

Post by WoodSpinner »

Northern Flicker wrote: Thu Jun 24, 2021 12:22 pm The difficulty of comparing to some hypothetical marginal rate in retirement is more than just the difficulty in predicting the future. Even if you have a good estimate of your average marginal tax rate averaged over your retirement years, because you have substantial control over the timing of trad IRA withdrawals, the average marginal tax rate of trad IRA withdrawals will typically be less than the average marginal tax rate

Another wrinkle is that IRMAA (adjustments to medicare premiums based on income), which need to be included in the marginal tax rate for this purpose for those with sufficiently high income to be affected by it.
Huh? Can you help me understand your observation?

WoodSpinner
Lee_WSP
Posts: 5338
Joined: Fri Apr 19, 2019 5:15 pm
Location: Arizona

Re: Roth Conversions - McQuarrie study

Post by Lee_WSP »

jeffyscott wrote: Thu Jun 24, 2021 4:16 pm
Lee_WSP wrote: Thu Jun 24, 2021 1:09 pm
bsteiner wrote: Sat Jun 19, 2021 10:14 am
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).
And even less if one is stuck with a SIMPLE IRA or even no employer sponsored plan. Or an SEP IRA as a non-owner employee.
OTOH, those of us that had more modest incomes can have a lifetime average savings rate of 30% and yet have retired with a very small taxable account. At retirement we had ~67% tax deferred, 26% Roth, and 7% taxable (mostly I and EE bonds).

Our ~30% average was very uneven, nearly all of it was saved during the 20 year period that we had 2 incomes. We had 3 children while in our 20s and spouse did not work until she was about 30. All the money that we contributed to savings comes, not coincidentally, to 110% of her lifetime earnings.

Those, like us, who live on << $100K per year can also do conversions at 12% year after year, without having a large taxable account.
Agreed. I don't think a savings rate of 30% is destined to *need* to invest a lot in a taxable account. Start nearing 50% and with SIMPLE IRA contribution limits and your taxable account is going to outstrip the SIMPLE account very quickly at incomes above 80k.
bsteiner
Posts: 5954
Joined: Sat Oct 20, 2012 9:39 pm
Location: NYC/NJ/FL

Re: Roth Conversions - McQuarrie study

Post by bsteiner »

McQ wrote: Thu Jun 24, 2021 3:52 pm ...
If you or your spouse snorts at the idea that they might live so long, consider life insurance products as an alternative means to pass on a tax-free bequest.
...
Explore the alternatives to a Roth conversion, including back door Roth IRAs, insurance products of various types, and, most important, the flexibility and tax-managed potential of funds in a brokerage account maintained outside of any tax-deferred structure.
...
To be better off with life insurance you have to die substantially before life expectancy.
User avatar
jeffyscott
Posts: 9753
Joined: Tue Feb 27, 2007 9:12 am
Location: Wisconsin

Re: Roth Conversions - McQuarrie study

Post by jeffyscott »

Lee_WSP wrote: Thu Jun 24, 2021 4:59 pm
jeffyscott wrote: Thu Jun 24, 2021 4:16 pm
Lee_WSP wrote: Thu Jun 24, 2021 1:09 pm
bsteiner wrote: Sat Jun 19, 2021 10:14 am
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).
And even less if one is stuck with a SIMPLE IRA or even no employer sponsored plan. Or an SEP IRA as a non-owner employee.
OTOH, those of us that had more modest incomes can have a lifetime average savings rate of 30% and yet have retired with a very small taxable account. At retirement we had ~67% tax deferred, 26% Roth, and 7% taxable (mostly I and EE bonds).

Our ~30% average was very uneven, nearly all of it was saved during the 20 year period that we had 2 incomes. We had 3 children while in our 20s and spouse did not work until she was about 30. All the money that we contributed to savings comes, not coincidentally, to 110% of her lifetime earnings.

Those, like us, who live on << $100K per year can also do conversions at 12% year after year, without having a large taxable account.
Agreed. I don't think a savings rate of 30% is destined to *need* to invest a lot in a taxable account. Start nearing 50% and with SIMPLE IRA contribution limits and your taxable account is going to outstrip the SIMPLE account very quickly at incomes above 80k.
Sure and some do not even have that or a 401K at all. Or they have a bad one that makes the effective limit 6% of pay or something like that (when it is only worth contributing to the level of the match).

For my spouse and me, it was two 401K equivalents and 2 roth IRA accounts, at 50 or over today that would be $66,000, which is 30% of $220K.
The two greatest enemies of the equity fund investor are expenses and emotions. ― John C. Bogle
Northern Flicker
Posts: 8037
Joined: Fri Apr 10, 2015 12:29 am

Re: Roth Conversions - McQuarrie study

Post by Northern Flicker »

WoodSpinner wrote: Thu Jun 24, 2021 4:59 pm
Northern Flicker wrote: Thu Jun 24, 2021 12:22 pm The difficulty of comparing to some hypothetical marginal rate in retirement is more than just the difficulty in predicting the future. Even if you have a good estimate of your average marginal tax rate averaged over your retirement years, because you have substantial control over the timing of trad IRA withdrawals, the average marginal tax rate of trad IRA withdrawals will typically be less than the average marginal tax rate

Another wrinkle is that IRMAA (adjustments to medicare premiums based on income), which need to be included in the marginal tax rate for this purpose for those with sufficiently high income to be affected by it.
Huh? Can you help me understand your observation?

WoodSpinner
Sure. In a year where our marginal tax rate is higher, it is because of income realized from other sources, and we do not do trad IRA/401K withdrawals or conversions in that year.

And if you consider the degenerate case where the IRA is the total source of income, and model a LIFO withdrawal pattern, the last dollar contributed is the first withdrawn, it is not hard to construct counterexamples at the level of rigor of a mathematical proof to show that both marginal tax rate in retirement and average tax rate in retirement fail as the target comparison with the marginal rate deferred for a contribution when making a trad vs Roth decision or evaluating a potential conversion.

And RMDs are another complication requiring an estimate of taxes over one's lifecourse from that point forward.
My postings are my opinion, and never should be construed as a recommendation to buy, sell, or hold any particular investment.
Tdubs
Posts: 1220
Joined: Tue Apr 24, 2018 7:50 pm

Re: Roth Conversions - McQuarrie study

Post by Tdubs »

McQ wrote: Wed Jun 23, 2021 2:26 pm
FlamePoint wrote: Wed Jun 23, 2021 11:12 am I’ve been reading this thread with a great deal of interest given our current situation. Even with all the good information out there I still feel like there’s a large hole/gap in available tools to model Roth conversions.

Through a series of what I’ll call fortunate events we ended up with rather large TDA upon retirement last year at age 58; started investing in our 20’s, full employment over a 30 year career with same employer, generous compensation plan and retirement benefits, pension that was taken as a lump sum. We are currently sitting on a combined $4.5 M retirement bucket of funds; $3.6M tax deferred, $450K Roth, and the remainder taxable. We will both be receiving close to max SS, and plan to take it at age 70. Current expenses are around $96K per year. SS will cover the vast majority of our expenses from age 70 on.

All the modeling I’ve done to date suggests we will benefit from Roth conversions between now and age 72. Unless we really start to ramp up our spending, it looks like our Roth funds will never be touched and will be our legacy funds for the kids.

At this point, I’m mainly concerned about the tax impact once one of us passes away. Without conversions taxes as MFJ start to really jump once RMD’s kick in. The single taxes brackets are even worse, so there appears to be a significant benefit to converting at least 50% of the TDA over the next 10 years or so.

Certainly a good problem to have, but challenging to determine best course of action given all the future unknowns (ie. inflation rate, tax law changes, our health, market fluctuations, etc).

I continue to watch and learn.
Dear Flamepoint: quite a few people in this forum and elsewhere have mentioned your fear: What happens to the survivor, when most income continues, but it is now taxed as a single, where the bracket floors kick in at half the MFJ dollars? I heard it so many times that I started to build a spreadsheet to quantify what might be called "the widow tax hit." Thus far the analysis is about as well-developed and coordinated as a 3-week old puppy, so please don't put too much weight on the next paragraph.
Long story short: it is not to be feared. The key is to evaluate the disposable income available to the widow (after tax, after Medicare), and *NOT* the dollars paid in tax or the average tax rate. Next, if the deceased spent about 25% of couple disposable income while alive (as in, like, eating), then the widow only has to cobble together 75% of the disposable income the couple had to maintain the same lifestyle. And analyses thus far show that, for couple income in the 22% through the 24% brackets ($100K - $350K), she does, despite increases in dollars paid in tax and tax rate, consequent to being a single.
I'll probably write a paper about it if the analyses hold up; and if I do, I'll post a link here on a BH forum. Just a caution: I move at an academic pace, which compares to glacial as my running speed compares to that of Usain Bolt.
While you are at it, I saw this nugget for your spreadsheet: "The probability that the wife will outlive her husband is 0.63 and, if she does, her survivor life expectancy is 12.5 years." https://www.nber.org/papers/w25009
Post Reply