Why Roth conversions always pay off—if you can hold on long enough

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
smitcat
Posts: 13227
Joined: Mon Nov 07, 2016 9:51 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by smitcat »

marcopolo wrote: Fri Oct 01, 2021 6:46 pm
smitcat wrote: Fri Oct 01, 2021 6:37 pm
marcopolo wrote: Fri Oct 01, 2021 5:07 pm
FiveK wrote: Fri Oct 01, 2021 5:01 pm
McQ wrote: Fri Oct 01, 2021 4:37 pm One clarification re the past 24 hours of posts, and the definition of "wealthy":
-$313,000, and $4.5 million, are inflated future 2036 dollars, today's thresholds inflated by a factor of about 1.55X (fifteen years of inflation at 3%)
-the values are those required for smallish SS, plus $100K pension, plus RMDs, to add up to enough to push a couple to the very top of what will, in 2036, be the projected ceiling of the 22% bracket.
-As a matter of principle, I refuse to call anyone located in the 22% income tax bracket, based on income from those three sources, wealthy.
Returning to today's dollars, and assuming for MFJ age 72:
top of the 22% bracket = $200K AGI
smallish SS = $20K/yr
pension = $100K/yr/1.55 = $64.5K/yr
RMD = $200K - ($20K + $64.5K) = $115.5K
Traditional balance = $115.5K * 27.4 = $3,165K

Based on Net Worth Percentile Calculator – United States, that puts them at the 96% percentile. Of course, "wealthy" is a subjective term....
Did you include the NPV of that generous pension?
What fraction of people getting ready to retire soon have pensions like that?
Most of us are replacing that pension income from savings.

Take away the pension, and you are close to $5M investable assets, putting you closer to 97+% percentile.
I agree "wealthy" is subjective.
"What fraction of people getting ready to retire soon have pensions like that?
Most of us are replacing that pension income from savings."

I would think the number of folks recieving pensions is less than 30% as a whole.
If you remove the large % of the population that has little or no savings it looks quite different.
But then number of folks recieving pensions that would be applicable to these Roth conversion discusions will be much higher.
Last I checked the number of folks in pensions was about 30 million people.
The ones we know that have have pensions include teachers, law enforcement, water district, firefighters, nurses, janitors, small goverment officials, dept of public works, SLP, therapists , school principals and superintendents, etc.

The 30% number is likely people receiving pensions now.
As you may be aware, many institutions have been steadily eliminating pensions. I suspect that number will be lower in the future.

We will probably get included in the number of people receiving pensions, we will be getting a grand total of $420/mo, not inflation adjusted, between the two us starting at age 65, then years from now. I would guess those receiving in the neighborhood of $65k in COLA'd pensions is significantly smaller.
Last data I could find participating in pensions was from 2019 - larger than I thought.
http://www.pensionrights.org/publicatio ... ment-plans [link fixed by admin LadyGeek]

The pension amounts we are used to seeing from folks we know are often near or exceeding 6 figures.
User avatar
FiveK
Posts: 15692
Joined: Sun Mar 16, 2014 2:43 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by FiveK »

McQ wrote: Fri Oct 01, 2021 5:53 pm The goal is greater after-tax wealth available for spending. Period.
Yes.
A Roth conversion, then, is like the buyout of a contract.
...
Paying tax upfront for an uncertain future outcome is rather like making an investment.
Between these two, the "buyout of a contract" analogy seems more apt.

With an investment, one hopes to get back gains but also the invested amount. With a Roth conversion the tax paid is gone to the government, the same as if one had been in business with a partner and bought out the partner to become the sole owner.
marcopolo
Posts: 8411
Joined: Sat Dec 03, 2016 9:22 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by marcopolo »

smitcat wrote: Fri Oct 01, 2021 6:54 pm
marcopolo wrote: Fri Oct 01, 2021 6:46 pm
smitcat wrote: Fri Oct 01, 2021 6:37 pm
marcopolo wrote: Fri Oct 01, 2021 5:07 pm
FiveK wrote: Fri Oct 01, 2021 5:01 pm Returning to today's dollars, and assuming for MFJ age 72:
top of the 22% bracket = $200K AGI
smallish SS = $20K/yr
pension = $100K/yr/1.55 = $64.5K/yr
RMD = $200K - ($20K + $64.5K) = $115.5K
Traditional balance = $115.5K * 27.4 = $3,165K

Based on Net Worth Percentile Calculator – United States, that puts them at the 96% percentile. Of course, "wealthy" is a subjective term....
Did you include the NPV of that generous pension?
What fraction of people getting ready to retire soon have pensions like that?
Most of us are replacing that pension income from savings.

Take away the pension, and you are close to $5M investable assets, putting you closer to 97+% percentile.
I agree "wealthy" is subjective.
"What fraction of people getting ready to retire soon have pensions like that?
Most of us are replacing that pension income from savings."

I would think the number of folks recieving pensions is less than 30% as a whole.
If you remove the large % of the population that has little or no savings it looks quite different.
But then number of folks recieving pensions that would be applicable to these Roth conversion discusions will be much higher.
Last I checked the number of folks in pensions was about 30 million people.
The ones we know that have have pensions include teachers, law enforcement, water district, firefighters, nurses, janitors, small goverment officials, dept of public works, SLP, therapists , school principals and superintendents, etc.

The 30% number is likely people receiving pensions now.
As you may be aware, many institutions have been steadily eliminating pensions. I suspect that number will be lower in the future.

We will probably get included in the number of people receiving pensions, we will be getting a grand total of $420/mo, not inflation adjusted, between the two us starting at age 65, then years from now. I would guess those receiving in the neighborhood of $65k in COLA'd pensions is significantly smaller.
Last data I could find participating in pensions was from 2019 - larger than I thought.
http://www.pensionrights.org/publicatio ... ment-plans [link fixed by admin LadyGeek]

The pension amounts we are used to seeing from folks we know are often near or exceeding 6 figures.
That is because you live in a lake Wobegon world.
Median pension sizes:

http://www.pensionrights.org/publicatio ... e-pensions
Once in a while you get shown the light, in the strangest of places if you look at it right.
curmudgeon
Posts: 2630
Joined: Thu Jun 20, 2013 11:00 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by curmudgeon »

McQ wrote: Fri Oct 01, 2021 5:53 pm What should be the metric for Roth conversion outcomes?

For conversion under a constant tax rate of 22%, the Roth surplus at age 85 is $11,540. That’s the difference between Roth wealth of $296,205 and the after-tax wealth of (no-convert TDA + taxable) of $284,665. The ratio of those two is 1.0405. The 14th root is 1.00284, making the ROI under the new metric equal to an annualized 0.28%--as of age 85, after fourteen years.

That’s not a very large number. Rather puts the $11,540 “gain from conversion” into a different perspective, don’t you think?
Those are both reasonable numbers as a way of looking at the outcome. in comparison to a typical AUM load of 1%, or 100 basis points, the .28% seems moderate. On the other hand, Bogleheads are prone to counting every basis point when examining the merits of various index funds or classes of shares, in which case 28 basis points is a lot.

The absolute dollar amount is also useful, in the sense that you can consider it as "I spent a few hours planning a Roth conversion now, and I got back a nice vacation (or vacation upgrade) 15 years later.

For myself, if the tax rates are expected to be the same and the only gain from doing Roth conversions is via the tax drag on RMDs, I wouldn't find that very interesting. Because I have low income early retirement years, no pensions, and the potential to avoid some taxability of SS benefits, I'm more drawn to conversions.
sc9182
Posts: 2165
Joined: Wed Aug 17, 2016 7:43 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by sc9182 »

marcopolo wrote: Fri Oct 01, 2021 7:28 pm
smitcat wrote: Fri Oct 01, 2021 6:54 pm ..

Last data I could find participating in pensions was from 2019 - larger than I thought.
http://www.pensionrights.org/publicatio ... ment-plans [link fixed by admin LadyGeek]

The pension amounts we are used to seeing from folks we know are often near or exceeding 6 figures.
That is because you live in a lake Wobegon world.
Median pension sizes:

http://www.pensionrights.org/publicatio ... e-pensions
Don’t know why I can’t click-through to one of those links. What percentage are covered doesn’t really matter much - but how much of Pension benefits is rendered matters.

Also - Railroad, Union, Private sector pensions are steadily becoming a Dodo. Where as Military/Veteran pensions/benefits are slowly eroding. We don’t necessarily believe Private Pensions., and/or annuities (especially Private/Unions ones) offer much long-term hope/success for younger folks.

If one had/covered with top-dollar pension(s), with spousal benefit, especially with COLA — do enjoy, and plan accordingly— but let it be known that it’s a steadily becoming more selective (rare). We would not count on any type of pensions (except for $48/month equivalent pension at retirement when briefly covered during 90s working years at Mega corp)
User avatar
LadyGeek
Site Admin
Posts: 95466
Joined: Sat Dec 20, 2008 4:34 pm
Location: Philadelphia
Contact:

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by LadyGeek »

sc9182 wrote: Fri Oct 01, 2021 8:08 pm Don’t know why I can’t click-through to one of those links.
The link got cut short. It's fixed: http://www.pensionrights.org/publicatio ... ment-plans

In human readable format: How many American workers participate in workplace retirement plans?
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.
Running Bum
Posts: 356
Joined: Sun Feb 10, 2013 1:00 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by Running Bum »

McQ wrote: Fri Oct 01, 2021 5:53 pm
Paying tax upfront for an uncertain future outcome is rather like making an investment. Given the uncertainty of future tax structures and portfolio returns, it could even be called a wager. I will call the conversion an investment; and as such, the metric must be return on investment, or ROI, with the investment equal to the tax debit that had to be paid upfront to buy out of the TDA “contract.”
I don't understand how that mindset applies only when doing a Roth conversion. If you choose to continue to defer the tax liability (no conversion), you are dealing with those same uncertainties. In fact, doing a Roth conversion is taking a bird in the hand by locking in the current tax rate with the converted amount. That seems like less of a risk to me.

I've only skimmed this thread. The topic comes up a lot on early-retirement.org, though not to this depth. I think this was brought up before, but it seems like all of McQ's spreadsheets start at age 71. I'm trying to do as much Roth conversion as makes sense before age 70. By 70, I'll be collecting SS and a small pension, so I'm converting before those increase my income. Once RMDs start I'm not sure it makes any sense to convert more. Maybe I'm missing something by not studying these posts in depth.

I've seen other points made by McQ that I don't agree with, but they've probably been rebutted by others, so I won't rehash them. I'm well into my conversion strategy anyway, and I'm not seeing anything here that changes my plan.
marcopolo
Posts: 8411
Joined: Sat Dec 03, 2016 9:22 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by marcopolo »

smitcat wrote: Fri Oct 01, 2021 6:28 pm
marcopolo wrote: Fri Oct 01, 2021 1:36 pm
FiveK wrote: Fri Oct 01, 2021 1:06 pm
marcopolo wrote: Fri Oct 01, 2021 2:21 am
FiveK wrote: Fri Oct 01, 2021 1:31 am Other than "invest the RMD and don't withdraw converted funds from the Roth account" what criteria apply?
Well, quoting directly from Prof McQ post title Part II (describing the scenario he will be exploring):
McQ wrote: Wed Sep 29, 2021 6:00 pm For discussion purposes, I’ll take the middle time frame of 15 years and the 22%/24% boundary. For you to have a lock on the juicy Roth conversion outcomes discussed above, you would need an AGI, in 2036, RMDs included, of $313,000.
McQ wrote: Wed Sep 29, 2021 6:00 pm That indicates a TDA balance at age 72 of just over $7.2 million dollars.
McQ wrote: Wed Sep 29, 2021 6:00 pm They had pension income twice their social security income, an extra $100,000 in 2036. Which means they only have to project $4.5 million in their TDAs
We don't anticipate having $313k of taxable income in 15 years
We don't anticipate having $7.2M in a TDA
We don't have a $100k pension (more like 5k), nor likely to have a TDA of $4.5M

We have a little over $1.5M in our TDA, we keep only fixed income assets in the TDA, which will likely not do much more than keep up with inflation.
I suspect that is not that unusual for people in our ball park of net worth.
Might be the difference between "non-numeric assumptions" and "specific numbers used as examples".

In other words, what do you get if you use your specific numbers (instead of the ones in McQ's example), but adhere to the assumptions that you invest the RMD and don't withdraw converted funds from the Roth account?
Those numbers are not just random made up numbers.
Prof McQ delves into some detail to arrive at those number to meet the "non-numeric" assumptions he has laid out for the specific scenarios he is planning to model.

I agree that the framework can be used for other scenarios, I was commenting on the ones the Prof is planning to focus on for his analysis.

Using our numbers I don't get to anywhere near the the scenario where converting at 22% or 24% pays off handsomely.

That is drive by these differences from assumptions above:

We don't have much of a pension
Our Soc Sec is more modest (wife was a SAHM), and I retired well before getting 35 years of contributions.
Our TDA is not huge, and is invested in all fixed income.
Our taxable investement are invested tax-efficient (thanks to help from this forum!), low turnover
Due to quirks in our local ACA market, we get sizable tax credits at income levels higher than in most other places.
"Our Soc Sec is more modest (wife was a SAHM), and I retired well before getting 35 years of contributions."
FWIW I would double check that for sure , I have less than 30 years as well ...untill I count the years I worked in HS and college before graduation and getting a 'real job'. Due to the nature of SS that still leaves me with about $45K SS per year at 70 which I believe would then be 1-1/2 times that if my wife had not worked or $67.5K per year for both.

"Using our numbers I don't get to anywhere near the the scenario where converting at 22% or 24% pays off handsomely.
That is drive by these differences from assumptions above:
We don't have much of a pension.
Our TDA is not huge, and is invested in all fixed income.
Our taxable investement are invested tax-efficient (thanks to help from this forum!), low turnover
Due to quirks in our local ACA market, we get sizable tax credits at income levels higher than in most other places"

It would be great for you to explain your strategy to stay away from an eventual rise above the 22% tax rates.
With 1.5 million in TDA and maybe 4.5 million in after tax your accounts in total should be growing each year all along - correct?
What would your modeled portfolio performance ranges be? What happens with one spouses passing?
How does one eventually deplete the TDA and also utilize the funds within the aftertax along with the eventual SS without reaching 22% tax rates and above?
I don't expect to stay away from the 22% bracket, although not that far in to it.
That was not the criteria we were discussing. It was $313k in AGI in the future, the equivalent of about $200k today, capturing IRMAA.
In any case, we maintain all fixed income in our TDA, so expect close to 0% real growth. So, our first year RMD of about $55k, figure about $50k in Soc Sec, and some dividend being thrown off from the taxable account would likely keep our AGI in the $120k-$150k range, depending on how investments do before we get there.

We, in fact, do perform Roth Conversions at nearly 22% marginal rate, but that is actually in the 12% bracket due to loss of ACA tax credits.
It would not make sense for us to do conversions in the 22% bracket because that would be an effective rate of ~32%, which despite what the early versions of the spreadsheet shows, would not pay off when we will be withdrawing just a little into the 22% bracket, mainly because our taxable portfolio does NOT have 100% turnover each year.

An added note to our discussion about pensions.
Let's say 40% household have pensions, maybe 15% of them are over $100k (i think i am being very generous here, but not sure)
That is 6%.
My understanding is that at least some places that have pensions, do not also have generous 401k type plans, so potentially reduced savings in TDA for those people.
So, lets say 1/3 of those with $100k pensions also amass $4.5m in their TDA,
So, we are now at about 2% of households, which is roughly what FiveK and I came to from a different angle, as the percent of households that likely meet the criteria we are discussing for this analysis.

It is all subjective what we call "wealthy".
So, i will change my earlier statement to "This analysis is mainly for people in the top 2% or so of wealth in America"
Once in a while you get shown the light, in the strangest of places if you look at it right.
marcopolo
Posts: 8411
Joined: Sat Dec 03, 2016 9:22 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by marcopolo »

McQ wrote: Fri Oct 01, 2021 5:53 pm What should be the metric for Roth conversion outcomes?

[with a shoutout to Woodspinner’s thread on this topic]

Very important: this post assumes once again that the goal is to accumulate greater after-tax wealth some decades into the future, same as my first series of posts on this thread. I’ll be using younger ages in most of the examples (85 or 90, not 100+), but that still gives a conversion made while in the 60s some decades to play out.

Other goals will require other metrics. Some of the more ideological among taxpayers may set a goal of minimizing lifetime tax payments to the hated gumnint. As has been pointed out on other threads here at BH, one of the simplest ways to achieve that goal is to earn lousy returns on the funds in your TDA—leave it all in a money market fund, for instance. Silly, of course, but I try never to underestimate the prevalence and power of tax phobia. Many of the more manipulative sales pitches I see for Roth conversions try to aggravate that phobia, which lies latent in most of us.

The goal is greater after-tax wealth available for spending. Period. For the next portion of this thread, that will be my only metric for Roth conversions. Note the slight segue in this paragraph: greater after-tax spending power, for me and DW, says nothing about my heirs. So in this second series of posts, if a conversion doesn’t seem likely to pay off except for heirs, it does not pencil out.

Further: it’s not necessary for me to exercise the increased spending power; my consumption habits are pretty well set. And if I were to start spending down any converted funds, a new analysis would be required: spending at year n means less after-tax spending power at year n + 10. That analysis will be forthcoming; I just haven’t written my way to it yet.

My point: in this second series I will focus exclusively on in-life outcomes: the after-tax spending power available to the living at each evaluation, assuming no prior tapping of the funds. In these analyses there can be no step up in basis at death, because there have not yet been two deaths.

Choosing the metric

In the SSRN paper, I more or less threw everything against the wall, in the spirit of discovering what might stick. After reflection, and stimulated by BH criticism, here I will focus on a select few and add one new metric. Thus far the only metric used has been the additional after-tax, nominal dollars, achieved by converting versus not converting. That’s a very common metric in popular accounts: that after n years, the converter will be this many dollars ahead. But it’s a week reed, as we shall see, and not only because future inflated dollars are being used.

Every Roth conversion requires a voluntary tax debit paid right now: money sent to the government that the government did not demand, and would have been happy to receive only later, many years in the future, and under what could be a generous and will likely be a very prolonged repayment plan (=tax on RMDs at some percentage like 22% or less, on annual RMDs initially figured at about 4% of the balance, continuing for decades). Imagine a payment schedule set at double-digit basis points per year—would that my bank offered a mortgage on those terms! Then again, I can pay off a mortgage; but there is no way to pay off the tax liability on my tax-sheltered funds while maintaining that shelter.

Except through a Roth conversion.

A Roth conversion, then, is like the buyout of a contract. I accepted the government subsidy year after year for decades to build up my TDA more rapidly than I could otherwise have afforded. Now, in the form of taxable RMDs, the contract requires that I “pay back” that subsidy. That’s a metaphor; I may or may not pay back more dollars after starting RMDs than the subsidy dollars I received in the decades beforehand. Either way, having to take RMDs and having to pay tax on them at then-current rates was always part of the TDA “contract.”

When I make a Roth conversion, I remove all further tax liability on the sum that remains, now called Roth funds rather than TDA funds. If it’s 1929, and the market stays down for twenty years, I probably overpaid to get out of that contract [a hail to Lee_WSJ here]. If I paid 22% to avert RMDs whose tax liability would have been only 12%, I definitely overpaid.

Paying tax upfront for an uncertain future outcome is rather like making an investment. Given the uncertainty of future tax structures and portfolio returns, it could even be called a wager. I will call the conversion an investment; and as such, the metric must be return on investment, or ROI, with the investment equal to the tax debit that had to be paid upfront to buy out of the TDA “contract.”

But how to calculate the return on that investment?

My thinking on this count has evolved, as acknowledged above. But first, a question of framing. The day after the conversion, the after-tax value of that portion of my TDA, if I had not converted, must exactly equal the value of the Roth account, if I did convert, even though the pretax values are different, $100,000 versus $78,000 in the case of constant rates at 22%. So, where is the “investment?”

Answer: can’t frame the problem that way. The whole point of analyzing a conversion is to compare outcomes for writing a check for tax now, versus dribbling out much smaller tax payments, year after year, for what could be a long, long time. That framework blows up if you liquidate the TDA and pay tax now.

The investment is fact: you really do have to send the IRS a check for $22,000 right now. And they are not giving it back; that money is gone, except as it is earned back, slowly but surely, by the future tax payments you no longer have to make.

*and again: paying tax from outside the conversion will only have a small effect. The IRS still gets its check, paid from one of your accounts or another.

In sum: a Roth conversion requires an investment. The amount of the investment is the totally unnecessary, voluntarily undertaken tax debit that had to be paid up front.

But still, what is the return on that Roth conversion investment?

In the paper, one of my metrics was to take the nth geometric root of [1 + (real Roth surplus/tax debit)]. If the conversion was $12,000 ahead in real terms/constant dollars after 10 years, on an initial debit of $24,000, that means taking the 10th root of 1.5, or 1.04138. The ROI would then be a real 4.14% annualized, not too shabby.

But that formulation nagged at me. It was, after all, the geometric root of an arithmetic difference between two exponential series. Mmmnnh.

Here I’m going to introduce a simpler formulation: the nth geometric root of [Roth value / (combined after-tax value of counterfactual TDA + taxable account holding reinvested RMDs)]. What was an arithmetic difference—the dollar pay off to the Roth conversion cited in all previous posts—is now replaced by a simple ratio, which can be glossed as with-Roth wealth / no-conversion wealth. That’s consistent with what we agreed above : the goal is greater after-tax wealth. But now we express ‘greater’ by the degree to which the ratio exceeds 1.0, rather than the amount that remains after a subtraction.

Here is an example. In this spreadsheet I’ve added columns to the right, hidden other columns, and rounded to the dollar, hoping to keep it all within screen. It shows results through age 100 for conversion at a constant 22%. The newly visible columns, beginning with the yellow highlighted column Q, are the focus.

Image

For conversion under a constant tax rate of 22%, the Roth surplus at age 85 is $11,540. That’s the difference between Roth wealth of $296,205 and the after-tax wealth of (no-convert TDA + taxable) of $284,665. The ratio of those two is 1.0405. The 14th root is 1.00284, making the ROI under the new metric equal to an annualized 0.28%--as of age 85, after fourteen years.

That’s not a very large number. Rather puts the $11,540 “gain from conversion” into a different perspective, don’t you think?

We may have arrived at our F. Scott Fitzgerald moment:
1. Roth conversions, even at constant or lowered tax rates, do pay off, and in large dollar amounts at later ages;
2. But not by much, when the excess return is computed as a rate, and the evaluation occurs at younger ages

For comparison, the ROI metric in the SSRN paper would have been calculated this way: 1) Convert $11,540 into its real value of $7,629; 2) calculate (1 + (7629/24000)], which is 1.3468; 3) take the 14th root, which equals 1.0215; 4) making the “SSRN paper” ROI equal to 2.15% (rounded). Put another way, if you invested $22,000 at a real rate of 2.15%/year, after fourteen years you’d have a gain of $7,629 in real dollars.

Note that by contrast, the new metric does not require deflating either sum into constant dollars; both are expressed in the same units of inflated future dollars.

Let me stop here to see if there are comments, especially critiques of the new metric. There is one obvious objection which I would like to foreground for evaluation: why not lose the geometric roots and stick with the raw ratios?

-the new metric would then indicate that after fourteen years, you would have 1.0405X the after-tax wealth with conversion as without, given a constant tax rate of 22%;
-the SSRN paper metric would indicate that after fourteen years, you’d have gained a real, after-tax, 34.68% on the initial tax debit.

I’m stuck on the need for the geometric roots, in the spirit of recognizing the time value of money; but I never worked professionally in finance, and I am open to arguments for applying a different treatment to the raw ratios.

Thoughts?
I think either metric could be used as long as it is made clear either explicitly, or in context, which formulation is being considered.
I will once again say that it is still not very useful from a practical planning point of view with the taxable account being fully taxed on gains each year.
It falsely inflates the gains, regardless of which metric is used.
Once in a while you get shown the light, in the strangest of places if you look at it right.
User avatar
teen persuasion
Posts: 2319
Joined: Sun Oct 25, 2015 1:43 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by teen persuasion »

smitcat wrote: Fri Oct 01, 2021 6:28 pm
"Our Soc Sec is more modest (wife was a SAHM), and I retired well before getting 35 years of contributions."
FWIW I would double check that for sure , I have less than 30 years as well ...untill I count the years I worked in HS and college before graduation and getting a 'real job'. Due to the nature of SS that still leaves me with about $45K SS per year at 70 which I believe would then be 1-1/2 times that if my wife had not worked or $67.5K per year for both.
Spousal SS benefits are 1/2 of your SS at FRA (67?) not at 70. So if your SS is roughly $45k at 70, your SS at 67 is probably $36k; spousal benefits would be 1/2 of that, $18k, for a total around $63k. Possibly less, if the spouse claims prior to their FRA.
Chip
Posts: 3994
Joined: Wed Feb 21, 2007 3:57 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by Chip »

FiveK wrote: Fri Oct 01, 2021 2:43 pm
Chip wrote: Fri Oct 01, 2021 2:21 pm Having spendable funds isn't the same as actually spending those funds. And the government's share may change over time.

My heirs are charities. If I convert today and die tomorrow I never recover those prepaid taxes, nor do the charities. If my heirs weren't charities but I entered a LTC facility for a lengthy stay my future tax rate would decline significantly compared to my current projections.
No disagreement with those points, but they run counter to the condition in the original post:
FiveK wrote: Fri Oct 01, 2021 1:36 pm The payoff is always immediate when a conversion is done at a marginal rate lower than what the future holds.
Sorry, I'm still not getting your point here. Perhaps it's because of the the terms immediate and payoff. When I read payoff I think "money into my pocket" or "less money out of my pocket". When I read "immediate" I think "right now".

How about: "The payoff always occurs when a conversion is done at a marginal rate lower than what the future holds."?
Chip
Posts: 3994
Joined: Wed Feb 21, 2007 3:57 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by Chip »

cas wrote: Fri Oct 01, 2021 3:06 pm And that gets back to my interest in the "resilience" of Roth conversions to future unexpected life changes.

Stuff like:

Suppose: I don't think I'll need to use this money ever in my life. My primary beneficiary is a non-spouse with a higher expected tax rate than me, but my secondary beneficiary is a charity. Should I convert to Roth?

But what if I run through all my primary funds for long term care and end up having to dip into these funds? How old would I have to be before a Roth conversion wasn't a disaster (and maybe even a boon?)?

But I outlived my primary beneficiary, so now the Roth is going to go to a charity. Was the Roth conversion a disaster? Is it ever an advantage?

But my primary beneficiary had an unexpected life change and now their tax rate is going to be much lower than expected. Was the Roth conversion a disaster? Or was it fairly resilient to that type of change in fortune?
I like the term resilience. Though I am essentially finished with conversions, I probably should have extended my own analysis to attempt to answer those sorts of questions. Here's another one:

If I convert at a higher tax rate than my joint tax filing status analysis indicates, how many years would my surviving spouse have to pay single filer tax rates and IRMAA for it to pay off?
Edit: and to bring all those thoughts back to McQ, since this is his thread: McQ is trudging patiently through details, trying to explore a framework for looking at this whole "RMD effect" idea. It isn't quite fair for me to be flitting about like a hummingbird, settling only on surprising results and chattering about particularly tasty ones.
Nice phrasing. I'll try to keep those flitting urges in check as well. :beer
smitcat
Posts: 13227
Joined: Mon Nov 07, 2016 9:51 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by smitcat »

marcopolo wrote: Fri Oct 01, 2021 8:36 pm
smitcat wrote: Fri Oct 01, 2021 6:28 pm
marcopolo wrote: Fri Oct 01, 2021 1:36 pm
FiveK wrote: Fri Oct 01, 2021 1:06 pm
marcopolo wrote: Fri Oct 01, 2021 2:21 am

Well, quoting directly from Prof McQ post title Part II (describing the scenario he will be exploring):







We don't anticipate having $313k of taxable income in 15 years
We don't anticipate having $7.2M in a TDA
We don't have a $100k pension (more like 5k), nor likely to have a TDA of $4.5M

We have a little over $1.5M in our TDA, we keep only fixed income assets in the TDA, which will likely not do much more than keep up with inflation.
I suspect that is not that unusual for people in our ball park of net worth.
Might be the difference between "non-numeric assumptions" and "specific numbers used as examples".

In other words, what do you get if you use your specific numbers (instead of the ones in McQ's example), but adhere to the assumptions that you invest the RMD and don't withdraw converted funds from the Roth account?
Those numbers are not just random made up numbers.
Prof McQ delves into some detail to arrive at those number to meet the "non-numeric" assumptions he has laid out for the specific scenarios he is planning to model.

I agree that the framework can be used for other scenarios, I was commenting on the ones the Prof is planning to focus on for his analysis.

Using our numbers I don't get to anywhere near the the scenario where converting at 22% or 24% pays off handsomely.

That is drive by these differences from assumptions above:

We don't have much of a pension
Our Soc Sec is more modest (wife was a SAHM), and I retired well before getting 35 years of contributions.
Our TDA is not huge, and is invested in all fixed income.
Our taxable investement are invested tax-efficient (thanks to help from this forum!), low turnover
Due to quirks in our local ACA market, we get sizable tax credits at income levels higher than in most other places.
"Our Soc Sec is more modest (wife was a SAHM), and I retired well before getting 35 years of contributions."
FWIW I would double check that for sure , I have less than 30 years as well ...untill I count the years I worked in HS and college before graduation and getting a 'real job'. Due to the nature of SS that still leaves me with about $45K SS per year at 70 which I believe would then be 1-1/2 times that if my wife had not worked or $67.5K per year for both.

"Using our numbers I don't get to anywhere near the the scenario where converting at 22% or 24% pays off handsomely.
That is drive by these differences from assumptions above:
We don't have much of a pension.
Our TDA is not huge, and is invested in all fixed income.
Our taxable investement are invested tax-efficient (thanks to help from this forum!), low turnover
Due to quirks in our local ACA market, we get sizable tax credits at income levels higher than in most other places"

It would be great for you to explain your strategy to stay away from an eventual rise above the 22% tax rates.
With 1.5 million in TDA and maybe 4.5 million in after tax your accounts in total should be growing each year all along - correct?
What would your modeled portfolio performance ranges be? What happens with one spouses passing?
How does one eventually deplete the TDA and also utilize the funds within the aftertax along with the eventual SS without reaching 22% tax rates and above?
I don't expect to stay away from the 22% bracket, although not that far in to it.
That was not the criteria we were discussing. It was $313k in AGI in the future, the equivalent of about $200k today, capturing IRMAA.
In any case, we maintain all fixed income in our TDA, so expect close to 0% real growth. So, our first year RMD of about $55k, figure about $50k in Soc Sec, and some dividend being thrown off from the taxable account would likely keep our AGI in the $120k-$150k range, depending on how investments do before we get there.

We, in fact, do perform Roth Conversions at nearly 22% marginal rate, but that is actually in the 12% bracket due to loss of ACA tax credits.
It would not make sense for us to do conversions in the 22% bracket because that would be an effective rate of ~32%, which despite what the early versions of the spreadsheet shows, would not pay off when we will be withdrawing just a little into the 22% bracket, mainly because our taxable portfolio does NOT have 100% turnover each year.

An added note to our discussion about pensions.
Let's say 40% household have pensions, maybe 15% of them are over $100k (i think i am being very generous here, but not sure)
That is 6%.
My understanding is that at least some places that have pensions, do not also have generous 401k type plans, so potentially reduced savings in TDA for those people.
So, lets say 1/3 of those with $100k pensions also amass $4.5m in their TDA,
So, we are now at about 2% of households, which is roughly what FiveK and I came to from a different angle, as the percent of households that likely meet the criteria we are discussing for this analysis.

It is all subjective what we call "wealthy".
So, i will change my earlier statement to "This analysis is mainly for people in the top 2% or so of wealth in America"
"In any case, we maintain all fixed income in our TDA, so expect close to 0% real growth. So, our first year RMD of about $55k, figure about $50k in Soc Sec, and some dividend being thrown off from the taxable account would likely keep our AGI in the $120k-$150k range, depending on how investments do before we get there."
How do you remain at those low levels of taxable account incomes with an initial taxable account size of 4.5 million and the fact that the acount will grow over time?

"My understanding is that at least some places that have pensions, do not also have generous 401k type plans, so potentially reduced savings in TDA for those people."
The link posted has the number of folks participating in both pensions and retirement accounts at 29+ million. (not eligable but particpating)

"We, in fact, do perform Roth Conversions at nearly 22% marginal rate, but that is actually in the 12% bracket due to loss of ACA tax credits."
How do you get any room to do Roth conversions now with 4.5 million in taxable supplying income? Where do the current spendable funds come from without generating income?

What are your thought/plans for the Hawaii estate tax?
Last edited by smitcat on Sat Oct 02, 2021 8:10 am, edited 1 time in total.
smitcat
Posts: 13227
Joined: Mon Nov 07, 2016 9:51 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by smitcat »

teen persuasion wrote: Fri Oct 01, 2021 9:42 pm
smitcat wrote: Fri Oct 01, 2021 6:28 pm
"Our Soc Sec is more modest (wife was a SAHM), and I retired well before getting 35 years of contributions."
FWIW I would double check that for sure , I have less than 30 years as well ...untill I count the years I worked in HS and college before graduation and getting a 'real job'. Due to the nature of SS that still leaves me with about $45K SS per year at 70 which I believe would then be 1-1/2 times that if my wife had not worked or $67.5K per year for both.
Spousal SS benefits are 1/2 of your SS at FRA (67?) not at 70. So if your SS is roughly $45k at 70, your SS at 67 is probably $36k; spousal benefits would be 1/2 of that, $18k, for a total around $63k. Possibly less, if the spouse claims prior to their FRA.
Thank you - we do not have a SAH spouse so not familiar with those rules.
ThankYouJack
Posts: 5704
Joined: Wed Oct 08, 2014 7:27 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by ThankYouJack »

McQ wrote: Thu Sep 23, 2021 12:52 pm
Why are RMDs reinvested? Because of the underlying assumption about goals: that a conversion was contemplated precisely because these RMDs were expected to be surplus, i.e., not necessary to meet spending needs or living expenses. The conversion serves to reduce these surplus RMDs, thus preventing unwanted ongoing tax exposure, at the cost of a one-time tax payment up front. But if the conversion had not been done, the RMDs would still have been surplus, hence, the after-tax remainder of each RMD could have been immediately re-invested.
Incredible analysis, but I think "always" in the title could be misleading/confusing to some because of certain underlying assumptions like the one above.

I wonder how the numbers look if RMDs are not reinvested?


Another factor why "always" might not be the case. I have lumpy income and will be in the 12% tax bracket this year. Initial thought is I should do a conversion since I'll probably never be in a lower federal tax bracket, my portfolio is very pre-tax heavy, RMDs will likely come into play, but here are some conditions:

1) If I don't do the conversion, I may receive around a $3k for the earned income tax credit
2) If I do the conversion, I'll incur a long term capital gain because I'll need to pay the additional taxes from my taxable account
3) My state income tax rate will be gradually decreasing over the next few years

So it may cost me quite a bit to do the conversion. I'm leaning towards not doing the conversion this year assuming I can get the EITC
User avatar
teen persuasion
Posts: 2319
Joined: Sun Oct 25, 2015 1:43 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by teen persuasion »

ThankYouJack wrote: Sat Oct 02, 2021 9:33 am
McQ wrote: Thu Sep 23, 2021 12:52 pm
Why are RMDs reinvested? Because of the underlying assumption about goals: that a conversion was contemplated precisely because these RMDs were expected to be surplus, i.e., not necessary to meet spending needs or living expenses. The conversion serves to reduce these surplus RMDs, thus preventing unwanted ongoing tax exposure, at the cost of a one-time tax payment up front. But if the conversion had not been done, the RMDs would still have been surplus, hence, the after-tax remainder of each RMD could have been immediately re-invested.
Incredible analysis, but I think "always" in the title could be misleading/confusing to some because of certain underlying assumptions like the one above.

I wonder how the numbers look if RMDs are not reinvested?


Another factor why "always" might not be the case. I have lumpy income and will be in the 12% tax bracket this year. Initial thought is I should do a conversion since I'll probably never be in a lower federal tax bracket, my portfolio is very pre-tax heavy, RMDs will likely come into play, but here are some conditions:

1) If I don't do the conversion, I may receive around a $3k for the earned income tax credit
2) If I do the conversion, I'll incur a long term capital gain because I'll need to pay the additional taxes from my taxable account
3) My state income tax rate will be gradually decreasing over the next few years

So it may cost me quite a bit to do the conversion. I'm leaning towards not doing the conversion this year assuming I can get the EITC
Yeah, I'm looking at conversions vs EITC, too. I've realized you can't focus on tax bracket, it has to be marginal rate. And EITC's phaseout rate increases your marginal rate, as does pushing you out of the 0% LTCG bracket.

For us, when we had more than one dependent, EITC phaseout is 21%, plus a state 30% match (another 6.3%). As we go to one dependent, I believe that EITC phaseout is 16% (and state 30% match adds another 4.8%). Then add on federal and state tax brackets. So our marginal rate is probably 36% for any Roth conversions while we could be eligible for EITC. :annoyed
ThankYouJack
Posts: 5704
Joined: Wed Oct 08, 2014 7:27 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by ThankYouJack »

teen persuasion wrote: Sat Oct 02, 2021 10:32 am
ThankYouJack wrote: Sat Oct 02, 2021 9:33 am
McQ wrote: Thu Sep 23, 2021 12:52 pm
Why are RMDs reinvested? Because of the underlying assumption about goals: that a conversion was contemplated precisely because these RMDs were expected to be surplus, i.e., not necessary to meet spending needs or living expenses. The conversion serves to reduce these surplus RMDs, thus preventing unwanted ongoing tax exposure, at the cost of a one-time tax payment up front. But if the conversion had not been done, the RMDs would still have been surplus, hence, the after-tax remainder of each RMD could have been immediately re-invested.
Incredible analysis, but I think "always" in the title could be misleading/confusing to some because of certain underlying assumptions like the one above.

I wonder how the numbers look if RMDs are not reinvested?


Another factor why "always" might not be the case. I have lumpy income and will be in the 12% tax bracket this year. Initial thought is I should do a conversion since I'll probably never be in a lower federal tax bracket, my portfolio is very pre-tax heavy, RMDs will likely come into play, but here are some conditions:

1) If I don't do the conversion, I may receive around a $3k for the earned income tax credit
2) If I do the conversion, I'll incur a long term capital gain because I'll need to pay the additional taxes from my taxable account
3) My state income tax rate will be gradually decreasing over the next few years

So it may cost me quite a bit to do the conversion. I'm leaning towards not doing the conversion this year assuming I can get the EITC
Yeah, I'm looking at conversions vs EITC, too. I've realized you can't focus on tax bracket, it has to be marginal rate. And EITC's phaseout rate increases your marginal rate, as does pushing you out of the 0% LTCG bracket.

For us, when we had more than one dependent, EITC phaseout is 21%, plus a state 30% match (another 6.3%). As we go to one dependent, I believe that EITC phaseout is 16% (and state 30% match adds another 4.8%). Then add on federal and state tax brackets. So our marginal rate is probably 36% for any Roth conversions while we could be eligible for EITC. :annoyed
Good point but I think it's even more complex than just looking at the marginal rate. If I convert, I would fill up the top of the 12% bracket, far beyond the EITC limit. So the first $15k may be at the higher EITC phase out marginal rate, but the next $45k or so would be at a much lower rate.
marcopolo
Posts: 8411
Joined: Sat Dec 03, 2016 9:22 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by marcopolo »

smitcat wrote: Sat Oct 02, 2021 8:03 am
marcopolo wrote: Fri Oct 01, 2021 8:36 pm
smitcat wrote: Fri Oct 01, 2021 6:28 pm
marcopolo wrote: Fri Oct 01, 2021 1:36 pm
FiveK wrote: Fri Oct 01, 2021 1:06 pm
Might be the difference between "non-numeric assumptions" and "specific numbers used as examples".

In other words, what do you get if you use your specific numbers (instead of the ones in McQ's example), but adhere to the assumptions that you invest the RMD and don't withdraw converted funds from the Roth account?
Those numbers are not just random made up numbers.
Prof McQ delves into some detail to arrive at those number to meet the "non-numeric" assumptions he has laid out for the specific scenarios he is planning to model.

I agree that the framework can be used for other scenarios, I was commenting on the ones the Prof is planning to focus on for his analysis.

Using our numbers I don't get to anywhere near the the scenario where converting at 22% or 24% pays off handsomely.

That is drive by these differences from assumptions above:

We don't have much of a pension
Our Soc Sec is more modest (wife was a SAHM), and I retired well before getting 35 years of contributions.
Our TDA is not huge, and is invested in all fixed income.
Our taxable investement are invested tax-efficient (thanks to help from this forum!), low turnover
Due to quirks in our local ACA market, we get sizable tax credits at income levels higher than in most other places.
"Our Soc Sec is more modest (wife was a SAHM), and I retired well before getting 35 years of contributions."
FWIW I would double check that for sure , I have less than 30 years as well ...untill I count the years I worked in HS and college before graduation and getting a 'real job'. Due to the nature of SS that still leaves me with about $45K SS per year at 70 which I believe would then be 1-1/2 times that if my wife had not worked or $67.5K per year for both.

"Using our numbers I don't get to anywhere near the the scenario where converting at 22% or 24% pays off handsomely.
That is drive by these differences from assumptions above:
We don't have much of a pension.
Our TDA is not huge, and is invested in all fixed income.
Our taxable investement are invested tax-efficient (thanks to help from this forum!), low turnover
Due to quirks in our local ACA market, we get sizable tax credits at income levels higher than in most other places"

It would be great for you to explain your strategy to stay away from an eventual rise above the 22% tax rates.
With 1.5 million in TDA and maybe 4.5 million in after tax your accounts in total should be growing each year all along - correct?
What would your modeled portfolio performance ranges be? What happens with one spouses passing?
How does one eventually deplete the TDA and also utilize the funds within the aftertax along with the eventual SS without reaching 22% tax rates and above?
I don't expect to stay away from the 22% bracket, although not that far in to it.
That was not the criteria we were discussing. It was $313k in AGI in the future, the equivalent of about $200k today, capturing IRMAA.
In any case, we maintain all fixed income in our TDA, so expect close to 0% real growth. So, our first year RMD of about $55k, figure about $50k in Soc Sec, and some dividend being thrown off from the taxable account would likely keep our AGI in the $120k-$150k range, depending on how investments do before we get there.

We, in fact, do perform Roth Conversions at nearly 22% marginal rate, but that is actually in the 12% bracket due to loss of ACA tax credits.
It would not make sense for us to do conversions in the 22% bracket because that would be an effective rate of ~32%, which despite what the early versions of the spreadsheet shows, would not pay off when we will be withdrawing just a little into the 22% bracket, mainly because our taxable portfolio does NOT have 100% turnover each year.

An added note to our discussion about pensions.
Let's say 40% household have pensions, maybe 15% of them are over $100k (i think i am being very generous here, but not sure)
That is 6%.
My understanding is that at least some places that have pensions, do not also have generous 401k type plans, so potentially reduced savings in TDA for those people.
So, lets say 1/3 of those with $100k pensions also amass $4.5m in their TDA,
So, we are now at about 2% of households, which is roughly what FiveK and I came to from a different angle, as the percent of households that likely meet the criteria we are discussing for this analysis.

It is all subjective what we call "wealthy".
So, i will change my earlier statement to "This analysis is mainly for people in the top 2% or so of wealth in America"
"In any case, we maintain all fixed income in our TDA, so expect close to 0% real growth. So, our first year RMD of about $55k, figure about $50k in Soc Sec, and some dividend being thrown off from the taxable account would likely keep our AGI in the $120k-$150k range, depending on how investments do before we get there."
How do you remain at those low levels of taxable account incomes with an initial taxable account size of 4.5 million and the fact that the acount will grow over time?

"My understanding is that at least some places that have pensions, do not also have generous 401k type plans, so potentially reduced savings in TDA for those people."
The link posted has the number of folks participating in both pensions and retirement accounts at 29+ million. (not eligable but particpating)

"We, in fact, do perform Roth Conversions at nearly 22% marginal rate, but that is actually in the 12% bracket due to loss of ACA tax credits."
How do you get any room to do Roth conversions now with 4.5 million in taxable supplying income? Where do the current spendable funds come from without generating income?

What are your thought/plans for the Hawaii estate tax?
The simple answer is that we don't have $4.5m in a taxable account. Our ~$7m in net worth is roughly broken down as follows:
$2.5m in Taxable
$1.5m in TDA
$1.0m in Roth/HSA
$350k in 529 plans (mostly excess, 1 semester left to pay)
$2.0m in Home (as you noticed, we live in Hawaii)

We also gave a sizable DAF that we contributed to when we were accumulating that we use for our charitable giving that is not included here.

The taxable account throws off about $40k in dividends.
For now we get the rest of our spending (typically ~$120k, less during Covid) from selling in taxable, that still leaves about $30-50k a year or so to be in the 12% bracket and stay under ACA cliff (marginal rate of 20.5% in 2021/2, 21.8% in other years). The exact amount depends on whether our AA dictates to sell equities (with gains) or bonds from taxable. That may narrow a bit when we go down to family of 2.

We expect our taxable account to dwindle down a bit by the time we hit RMD age to get to numbers I stated above. We are drawing close to 5% from that account.

Do you think it would make sense for us to convert into the 22% bracket, paying over 30% marginal rates? I just don't see how that would pay off. Even for a surviving spouse, the marginal rates don't get that high until $165k.

The Hawaii estate tax exemption is $5.49m, with portability, so a total of $11m. If our assets keep growing despite our spending, and annual gifting, and get above that level, I will be happy to pay some estate taxes. I also expect the exemption amount to change over time. But, we can't discuss that here.
Once in a while you get shown the light, in the strangest of places if you look at it right.
User avatar
FiveK
Posts: 15692
Joined: Sun Mar 16, 2014 2:43 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by FiveK »

ThankYouJack wrote: Sat Oct 02, 2021 11:44 am Good point but I think it's even more complex than just looking at the marginal rate. If I convert, I would fill up the top of the 12% bracket, far beyond the EITC limit. So the first $15k may be at the higher EITC phase out marginal rate, but the next $45k or so would be at a much lower rate.
If one calculates Marginal tax rate for whatever (IRA contribution, IRA withdrawal, etc.) as (change in tax)/(change in whatever) then it all works out.

When the more "whatever" one does leads to a worse result at some point (e.g., making 401k contributions until dropping from the 22% to 12% bracket, doing Roth conversions until going from the 24% to 32% bracket, etc.) the analysis is straightforward: the different marginal rate effects are separate from each other.

When one has to go through a "worse" range to reach a "better" range (e.g., the example here of paying a higher rate on initial conversions but a lower rate on additional conversions) one has to combine the marginal rate effects into that (change in tax)/(change in whatever) calculation.
ThankYouJack
Posts: 5704
Joined: Wed Oct 08, 2014 7:27 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by ThankYouJack »

I'm still trying to wrap my head around this. I know some on here are ultra-wealthy and ultra-savers but how likely is it that one would reinvest 100% of their RMDs from the time they start until the day they die. If that's the plan, what sort of portfolio withdrawal rates are we talking -- extremely small or even negative? And wouldn't it make sense to give away more of the wealth while one is alive?

FiveK wrote: Sat Oct 02, 2021 1:28 pm
ThankYouJack wrote: Sat Oct 02, 2021 11:44 am Good point but I think it's even more complex than just looking at the marginal rate. If I convert, I would fill up the top of the 12% bracket, far beyond the EITC limit. So the first $15k may be at the higher EITC phase out marginal rate, but the next $45k or so would be at a much lower rate.
If one calculates Marginal tax rate for whatever (IRA contribution, IRA withdrawal, etc.) as (change in tax)/(change in whatever) then it all works out.

When the more "whatever" one does leads to a worse result at some point (e.g., making 401k contributions until dropping from the 22% to 12% bracket, doing Roth conversions until going from the 24% to 32% bracket, etc.) the analysis is straightforward: the different marginal rate effects are separate from each other.

When one has to go through a "worse" range to reach a "better" range (e.g., the example here of paying a higher rate on initial conversions but a lower rate on additional conversions) one has to combine the marginal rate effects into that (change in tax)/(change in whatever) calculation.
Thanks, that makes sense to me.
User avatar
FiveK
Posts: 15692
Joined: Sun Mar 16, 2014 2:43 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by FiveK »

ThankYouJack wrote: Sat Oct 02, 2021 5:06 pm I'm still trying to wrap my head around this. I know some on here are ultra-wealthy and ultra-savers but how likely is it that one would reinvest 100% of their RMDs from the time they start until the day they die.
Good question. Perhaps "not enough to spend much space in the initial parts of the Traditional versus Roth - Bogleheads and Roth IRA conversion - Bogleheads wiki articles," but "enough to mention the possibility in those articles and refer people to threads like this and tools mentioned herein for those who want to evaluate the applicability to their own situations"?
curmudgeon
Posts: 2630
Joined: Thu Jun 20, 2013 11:00 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by curmudgeon »

ThankYouJack wrote: Sat Oct 02, 2021 5:06 pm I'm still trying to wrap my head around this. I know some on here are ultra-wealthy and ultra-savers but how likely is it that one would reinvest 100% of their RMDs from the time they start until the day they die. If that's the plan, what sort of portfolio withdrawal rates are we talking -- extremely small or even negative? And wouldn't it make sense to give away more of the wealth while one is alive?
I have a relative who generally falls into that category. Single, not long after retirement sold her house and moved into a CCRC. Has both a state pension and social security. Thrifty, but not penny-pinching. The retirement community and volunteering filled much of her time. Enjoyed travelling some, but not into extensive travel. For the most part she lived off of pension/SS. Starting to need some substantial extra assistance at age 97, and so may finally start spending some of her RMD (at this point, much of the tIRA has already gone into taxable via RMDs).

For her, the tIRA was a reserve, not something she planned to spend. Her estate plans have a portion going to charity, and the rest going to relatives.
ChrisC
Posts: 1475
Joined: Tue Jun 19, 2012 9:10 am
Location: North Carolina

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by ChrisC »

ThankYouJack wrote: Sat Oct 02, 2021 5:06 pm I'm still trying to wrap my head around this. I know some on here are ultra-wealthy and ultra-savers but how likely is it that one would reinvest 100% of their RMDs from the time they start until the day they die?
I've seen many people who routinely post in this forum, who are Federal pensioners with generous steams of pension income, like myself. I know "ultra-wealthy" and I'm certain we're not even remotely within that space. The difference in net worth between me and Marcopolo, who posted upthread and apparently questions whether he can take meaningful advantage of Roth conversions, is negligible. But I have a Federal pension, COLA-adjusted, that annually provides me with a healthy income stream and my wife (who has a modest COLA-adjusted pension from local government service) and I were super-savers during employment, combined with employer matches to 401Ks during most of the last 2 decades of our employment. With additional Social Security streams of income, modest income from other taxable sources, we have not touched one cent of our TDAs (except for conversions) or Roths since we retired in 2013, and all of these accounts will likely be inherited by our children. Our pensions and SS income alone trips us into the 24% tax bracket. We began modest Roth conversions in 2013, and have become more aggressive in recent years, resulting in moving all my wife's TDAs to Roth before she reached 70 and started drawing on Social Security retirement benefits.

There are quite a few people here with Federal civilian or military pensions, who can live solely on those income streams and even enjoy a robust retirement, including extensive travel, funding grandchildren 529s, and gifting to children or charities. And they will never touch their TDAs or Roths for living expenses and will leave the accounts all to their children, which in my case are in higher brackets than my wife and I, MFJ. (Though it is conceivable for us to draw on our TDAs for LTC expenses, after we've exhausted a war chess of LTC resources, which include LTCi, $140K plus HSAs and possibly residence in a CCRC -- and those draws would certainly be tax advantaged to us.)

The same is likely true for a number of State and local government pensioners, who in some cases have even more generous pension income than us Feds because of the ability to "pension spike" income from bonuses or over-time pay, which adds to their retirement annuity base.
smitcat
Posts: 13227
Joined: Mon Nov 07, 2016 9:51 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by smitcat »

marcopolo wrote: Sat Oct 02, 2021 12:19 pm
smitcat wrote: Sat Oct 02, 2021 8:03 am
marcopolo wrote: Fri Oct 01, 2021 8:36 pm
smitcat wrote: Fri Oct 01, 2021 6:28 pm
marcopolo wrote: Fri Oct 01, 2021 1:36 pm

Those numbers are not just random made up numbers.
Prof McQ delves into some detail to arrive at those number to meet the "non-numeric" assumptions he has laid out for the specific scenarios he is planning to model.

I agree that the framework can be used for other scenarios, I was commenting on the ones the Prof is planning to focus on for his analysis.

Using our numbers I don't get to anywhere near the the scenario where converting at 22% or 24% pays off handsomely.

That is drive by these differences from assumptions above:

We don't have much of a pension
Our Soc Sec is more modest (wife was a SAHM), and I retired well before getting 35 years of contributions.
Our TDA is not huge, and is invested in all fixed income.
Our taxable investement are invested tax-efficient (thanks to help from this forum!), low turnover
Due to quirks in our local ACA market, we get sizable tax credits at income levels higher than in most other places.
"Our Soc Sec is more modest (wife was a SAHM), and I retired well before getting 35 years of contributions."
FWIW I would double check that for sure , I have less than 30 years as well ...untill I count the years I worked in HS and college before graduation and getting a 'real job'. Due to the nature of SS that still leaves me with about $45K SS per year at 70 which I believe would then be 1-1/2 times that if my wife had not worked or $67.5K per year for both.

"Using our numbers I don't get to anywhere near the the scenario where converting at 22% or 24% pays off handsomely.
That is drive by these differences from assumptions above:
We don't have much of a pension.
Our TDA is not huge, and is invested in all fixed income.
Our taxable investement are invested tax-efficient (thanks to help from this forum!), low turnover
Due to quirks in our local ACA market, we get sizable tax credits at income levels higher than in most other places"

It would be great for you to explain your strategy to stay away from an eventual rise above the 22% tax rates.
With 1.5 million in TDA and maybe 4.5 million in after tax your accounts in total should be growing each year all along - correct?
What would your modeled portfolio performance ranges be? What happens with one spouses passing?
How does one eventually deplete the TDA and also utilize the funds within the aftertax along with the eventual SS without reaching 22% tax rates and above?
I don't expect to stay away from the 22% bracket, although not that far in to it.
That was not the criteria we were discussing. It was $313k in AGI in the future, the equivalent of about $200k today, capturing IRMAA.
In any case, we maintain all fixed income in our TDA, so expect close to 0% real growth. So, our first year RMD of about $55k, figure about $50k in Soc Sec, and some dividend being thrown off from the taxable account would likely keep our AGI in the $120k-$150k range, depending on how investments do before we get there.

We, in fact, do perform Roth Conversions at nearly 22% marginal rate, but that is actually in the 12% bracket due to loss of ACA tax credits.
It would not make sense for us to do conversions in the 22% bracket because that would be an effective rate of ~32%, which despite what the early versions of the spreadsheet shows, would not pay off when we will be withdrawing just a little into the 22% bracket, mainly because our taxable portfolio does NOT have 100% turnover each year.

An added note to our discussion about pensions.
Let's say 40% household have pensions, maybe 15% of them are over $100k (i think i am being very generous here, but not sure)
That is 6%.
My understanding is that at least some places that have pensions, do not also have generous 401k type plans, so potentially reduced savings in TDA for those people.
So, lets say 1/3 of those with $100k pensions also amass $4.5m in their TDA,
So, we are now at about 2% of households, which is roughly what FiveK and I came to from a different angle, as the percent of households that likely meet the criteria we are discussing for this analysis.

It is all subjective what we call "wealthy".
So, i will change my earlier statement to "This analysis is mainly for people in the top 2% or so of wealth in America"
"In any case, we maintain all fixed income in our TDA, so expect close to 0% real growth. So, our first year RMD of about $55k, figure about $50k in Soc Sec, and some dividend being thrown off from the taxable account would likely keep our AGI in the $120k-$150k range, depending on how investments do before we get there."
How do you remain at those low levels of taxable account incomes with an initial taxable account size of 4.5 million and the fact that the acount will grow over time?

"My understanding is that at least some places that have pensions, do not also have generous 401k type plans, so potentially reduced savings in TDA for those people."
The link posted has the number of folks participating in both pensions and retirement accounts at 29+ million. (not eligable but particpating)

"We, in fact, do perform Roth Conversions at nearly 22% marginal rate, but that is actually in the 12% bracket due to loss of ACA tax credits."
How do you get any room to do Roth conversions now with 4.5 million in taxable supplying income? Where do the current spendable funds come from without generating income?

What are your thought/plans for the Hawaii estate tax?
The simple answer is that we don't have $4.5m in a taxable account. Our ~$7m in net worth is roughly broken down as follows:
$2.5m in Taxable
$1.5m in TDA
$1.0m in Roth/HSA
$350k in 529 plans (mostly excess, 1 semester left to pay)
$2.0m in Home (as you noticed, we live in Hawaii)

We also gave a sizable DAF that we contributed to when we were accumulating that we use for our charitable giving that is not included here.

The taxable account throws off about $40k in dividends.
For now we get the rest of our spending (typically ~$120k, less during Covid) from selling in taxable, that still leaves about $30-50k a year or so to be in the 12% bracket and stay under ACA cliff (marginal rate of 20.5% in 2021/2, 21.8% in other years). The exact amount depends on whether our AA dictates to sell equities (with gains) or bonds from taxable. That may narrow a bit when we go down to family of 2.

We expect our taxable account to dwindle down a bit by the time we hit RMD age to get to numbers I stated above. We are drawing close to 5% from that account.

Do you think it would make sense for us to convert into the 22% bracket, paying over 30% marginal rates? I just don't see how that would pay off. Even for a surviving spouse, the marginal rates don't get that high until $165k.

The Hawaii estate tax exemption is $5.49m, with portability, so a total of $11m. If our assets keep growing despite our spending, and annual gifting, and get above that level, I will be happy to pay some estate taxes. I also expect the exemption amount to change over time. But, we can't discuss that here.
"The taxable account throws off about $40k in dividends."
(2.5 miilion / $40K approx. 1.5%)
This sounds really very good and also really low. How do/did you manage to keep these so low? I could see maybe picking specific funds now that are some of the lowest divs but we have to account for some 'legacy' investments made much earlier on that would need to be 'traded out'. That of course would then generate taxes as well.

"Do you think it would make sense for us to convert into the 22% bracket, paying over 30% marginal rates? I just don't see how that would pay off. Even for a surviving spouse, the marginal rates don't get that high until $165k."
Any plan that has the most likelihood of maximizing your after tax spending dollars including the life of the TDA's would be the one I voted for. If you can stay below any tax rate and still do that its great.
ThankYouJack
Posts: 5704
Joined: Wed Oct 08, 2014 7:27 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by ThankYouJack »

ChrisC wrote: Sun Oct 03, 2021 12:05 am
ThankYouJack wrote: Sat Oct 02, 2021 5:06 pm I'm still trying to wrap my head around this. I know some on here are ultra-wealthy and ultra-savers but how likely is it that one would reinvest 100% of their RMDs from the time they start until the day they die?
I've seen many people who routinely post in this forum, who are Federal pensioners with generous steams of pension income, like myself. I know "ultra-wealthy" and I'm certain we're not even remotely within that space. The difference in net worth between me and Marcopolo, who posted upthread and apparently questions whether he can take meaningful advantage of Roth conversions, is negligible. But I have a Federal pension, COLA-adjusted, that annually provides me with a healthy income stream and my wife (who has a modest COLA-adjusted pension from local government service) and I were super-savers during employment, combined with employer matches to 401Ks during most of the last 2 decades of our employment. With additional Social Security streams of income, modest income from other taxable sources, we have not touched one cent of our TDAs (except for conversions) or Roths since we retired in 2013, and all of these accounts will likely be inherited by our children. Our pensions and SS income alone trips us into the 24% tax bracket. We began modest Roth conversions in 2013, and have become more aggressive in recent years, resulting in moving all my wife's TDAs to Roth before she reached 70 and started drawing on Social Security retirement benefits.

There are quite a few people here with Federal civilian or military pensions, who can live solely on those income streams and even enjoy a robust retirement, including extensive travel, funding grandchildren 529s, and gifting to children or charities. And they will never touch their TDAs or Roths for living expenses and will leave the accounts all to their children, which in my case are in higher brackets than my wife and I, MFJ. (Though it is conceivable for us to draw on our TDAs for LTC expenses, after we've exhausted a war chess of LTC resources, which include LTCi, $140K plus HSAs and possibly residence in a CCRC -- and those draws would certainly be tax advantaged to us.)

The same is likely true for a number of State and local government pensioners, who in some cases have even more generous pension income than us Feds because of the ability to "pension spike" income from bonuses or over-time pay, which adds to their retirement annuity base.

Thanks for the example. I know there are situations out there like this (especially on this forum) but it still seems quite unique.

I won't be in the situation (won't be getting a large pension) but even if I have to take RMDs that far exceed my rock solid, very safe and comfortable living expenses, I'm unlikely to reinvest 100% of it and as I'd start giving more to my family / friends / charity while I'm alive. I can only dream to one day be in that situation :)
cas
Posts: 2245
Joined: Wed Apr 26, 2017 8:41 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by cas »

FiveK wrote: Fri Oct 01, 2021 7:26 pm
McQ wrote: Fri Oct 01, 2021 5:53 pm
A Roth conversion, then, is like the buyout of a contract.
...
Paying tax upfront for an uncertain future outcome is rather like making an investment.
Between these two, the "buyout of a contract" analogy seems more apt.

With an investment, one hopes to get back gains but also the invested amount. With a Roth conversion the tax paid is gone to the government, the same as if one had been in business with a partner and bought out the partner to become the sole owner.
I hestitate to wade in here, because my knowledge in this area is minimal. I wouldn't be surprised if I make some mind-numbingly ignorant comment.

But ... since no one else is entering into the McQ <-> FiveK discussion, here goes ...

1) I'm still rather vague on what the implications of the outcome of the "contract" vs "investment" discussion are.

Something like:

"Investment" would imply that a ROI-type-metric (for which McQ is trying to figure out the appropriate calculation) is a more appropriate metric?

"Contract" would imply that BETR (Break Even Tax Rate, like Kitces, Vanguard, and Fidelity have used in various papers/tools) is a more apt metric?

Or something else entirely?


2) To my ignorant viewpoint, something that seems like it might weigh on the "contract" side of the scales:

[Given the assumption that the RMDs aren't needed for living expenses and will be reinvested...]

The following does not seem to be the *only* implication of "buying out" the contract or not:

"Paying tax upfront for an uncertain future outcome" (and "locking in" a tax rate, for better or worse)
-vs-
"dribbling out much smaller tax payments [on the annual RMDs themselves], year after year, for what could be a long, long time" (and having tax rate adjust to the uncertain future year-by-year, for better or worse).

An *additional* part of the "contract" is that at age 72 (or whenever RMDs start), your business partner is insisting that you *must* change the nature of the "business" in a way that will (gradually) change how the "business" is taxed.

If you disapprove of that required change in business operations, you are free to buy out your partner and avoid that change in business operations.

(That is ... The IRS says you must take RMDs. Those RMDs will move from being in a tax-deferred account to being in a taxable account. Tax drag will probably exist on annual distributions (dividends, capital gains, etc.) from the taxable account. (Here the "uncertain future" rears its head again, but under tax law for many decades it seems quite difficult to achieve ZERO tax drag on a taxable account (over the long term) once you are getting social security.)

However, you are free to "buy out" the IRS's share of the TDA, keep the remaining assets in a tax-advantaged account (Roth), and avoid the issue of tax-drag on a taxable account.

[Assuming everything else is equal...]
Making this aspect of the decision on whether to "buy out" the IRS' share of your "business" involves considering whether you think the tax-drag on the taxable account will hurt the value of your share of the TDA+taxable accounts enough to bother you.)
smitcat
Posts: 13227
Joined: Mon Nov 07, 2016 9:51 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by smitcat »

ChrisC wrote: Sun Oct 03, 2021 12:05 am
ThankYouJack wrote: Sat Oct 02, 2021 5:06 pm I'm still trying to wrap my head around this. I know some on here are ultra-wealthy and ultra-savers but how likely is it that one would reinvest 100% of their RMDs from the time they start until the day they die?
I've seen many people who routinely post in this forum, who are Federal pensioners with generous steams of pension income, like myself. I know "ultra-wealthy" and I'm certain we're not even remotely within that space. The difference in net worth between me and Marcopolo, who posted upthread and apparently questions whether he can take meaningful advantage of Roth conversions, is negligible. But I have a Federal pension, COLA-adjusted, that annually provides me with a healthy income stream and my wife (who has a modest COLA-adjusted pension from local government service) and I were super-savers during employment, combined with employer matches to 401Ks during most of the last 2 decades of our employment. With additional Social Security streams of income, modest income from other taxable sources, we have not touched one cent of our TDAs (except for conversions) or Roths since we retired in 2013, and all of these accounts will likely be inherited by our children. Our pensions and SS income alone trips us into the 24% tax bracket. We began modest Roth conversions in 2013, and have become more aggressive in recent years, resulting in moving all my wife's TDAs to Roth before she reached 70 and started drawing on Social Security retirement benefits.

There are quite a few people here with Federal civilian or military pensions, who can live solely on those income streams and even enjoy a robust retirement, including extensive travel, funding grandchildren 529s, and gifting to children or charities. And they will never touch their TDAs or Roths for living expenses and will leave the accounts all to their children, which in my case are in higher brackets than my wife and I, MFJ. (Though it is conceivable for us to draw on our TDAs for LTC expenses, after we've exhausted a war chess of LTC resources, which include LTCi, $140K plus HSAs and possibly residence in a CCRC -- and those draws would certainly be tax advantaged to us.)

The same is likely true for a number of State and local government pensioners, who in some cases have even more generous pension income than us Feds because of the ability to "pension spike" income from bonuses or over-time pay, which adds to their retirement annuity base.
Great post - thank you
Although we ourselves have no pension(s) (zero) we do know quite a few folks in your situation now and many others who are younger who will be there someday.
yog
Posts: 659
Joined: Wed Jan 15, 2020 11:57 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by yog »

OECD defines taxes as compulsory unrequited obligations to the government. There can be no ROI on fulfilling an obligation. There can be no 'tax diversification'. You fulfill your obligations according to the terms. Since taxes on deferred income will ultimately be fulfilled by us our our executors, then by solving for time & opportunity value we can optimally time the recognition of income when it is most advantageous to our metric of maximizing after-tax income.

This is already solved in commercial applications using Bellman optimality equations. Mathematically this is much more complicated than most realize for preserving accuracy.

When you respect time space to capture 100% of remaining available opportunity in actionable time space, there are 5 distinct tax outcomes where Roth conversions may provide opportunity in the present:
1) Savings from converting now at marginal tax rates lower than marginal tax rates at future withdrawals
2) Gains from reduced taxable drag & increased Roth contributions when paying conversion taxes with funds outside retirement accounts
3) Savings from eliminating or reducing rising ordinary income from mandatory RMDs
4) Savings from reduced state estate taxes
5) Enhanced liquidity prior to age 59.5 with penalty-free access to retirement accounts subject to 5yr rules

The key insight here is tax optimization, including Roth conversions, isn't actually a financial problem but a time scheduling problem. All Roth conversions do is accelerate ordinary income.

Though it addresses a slightly different aspect of retirement planning, here's a interesting recent scholarly paper written by someone McQ may know where Bellman equations are used extensively to support their research: Combining Investment and Tax Strategies for Optimizing Lifetime Solvency under Uncertain Returns and Mortality
Topic Author
McQ
Posts: 1414
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by McQ »

ThankYouJack wrote: Sat Oct 02, 2021 5:06 pm I'm still trying to wrap my head around this. I know some on here are ultra-wealthy and ultra-savers but how likely is it that one would reinvest 100% of their RMDs from the time they start until the day they die. If that's the plan, what sort of portfolio withdrawal rates are we talking -- extremely small or even negative? And wouldn't it make sense to give away more of the wealth while one is alive?
...
[quoting only ThankYouJack here, but the issue has come up several times, so here goes]

Keep in mind: in the spreadsheet for this thread, everything is off-camera except the results for converting / not-converting $100,000, assumed to be a fraction of the total TDA at the time. Something generated $108,750, to put the taxpayer into the 22% bracket when the RMDs begin; or, something generated the $200,450 to put the taxpayer into the 24% bracket; but those other sources of income, which might include many tens of thousands in RMDs on the remainder of the TDA, or much less, are purposely not visible.

The beauty of that approach: I don’t have to care whether the taxpayer is ChrisC, with the $200,450 coming mostly from social security and pensions, or Rob and Sue in the SSRN paper, with all of it coming from social security and RMDs on a $4 million TDA. To enter 22% as the tax on avoided RMDs—the RMDs reduced by the conversion—there just needs to be $108,750 (or more) coming from some income source(s)—whatever these happen to be.

Which brings us to “How could someone invest all their RMDs as surplus funds for the rest of their life? It makes no sense.” That misunderstands the spreadsheet set up. The spreadsheet simply says, if you don’t convert $100,000, you are going to have RMDs from that $100,000, and they are going to be taxed at the indicated rate. The first year RMD on a $100,000 portion of the TDA is $3,650. I can certainly imagine a family with $108,750 of other income, who needs every penny of that extra $3,650 to live (=total taxable income of $112,400 to scrape by). But really … remind me again why they wanted to reduce the amount of RMDs subject to tax?

Throughout I’ve made the tacit assumption, which I will have to make explicit when I edit the opening post, that the conversion is partial, so that it is made on the margin. Since $100,000 is only a portion of the TDA, the RMDs being reduced are the marginal RMDs, the last $3650 of RMDs, which if no conversion is done, will be taxed at 22% or 24%, the stated tax rate avoided via conversion in the case of most of the scenarios examined.

That’s all that “reinvested as surplus” implies. Tens of thousands of dollars of other RMDs might be taken and spent, as part of the $108,750 or $200,450 of income required to push the taxpayer into the 22% or 24% bracket prior to the marginal RMDs, which could have been removed by conversion.

PS. The dirty little secret of RMD reduction: it takes huge conversions to materially reduce the RMD dollar amount (during the 70s, when the withdrawal rate is about 4%). In the running example: it took a $100,000 conversion to reduce taxable income by ... wait for it ... all of $3650. To drop all the way down from top of the pre-IRMAA 22% bracket ($176,000) to the bottom ($108,750), would require conversions adding up to $1.842 million dollars. That's another reason why I like the new spreadsheet: a $100,000 conversion at the margin seems rather more realistic.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
User avatar
FiveK
Posts: 15692
Joined: Sun Mar 16, 2014 2:43 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by FiveK »

yog wrote: Sun Oct 03, 2021 3:48 pm Though it addresses a slightly different aspect of retirement planning, here's a interesting recent scholarly paper written by someone McQ may know where Bellman equations are used extensively to support their research: Combining Investment and Tax Strategies for Optimizing Lifetime Solvency under Uncertain Returns and Mortality
Thanks for posting the link - interesting paper.

Seems in accord with the wiki article on Tax-efficient fund placement:
Determination of your asset allocation (% stocks / % bonds), which sets your portfolio's level of acceptable risk, is the single most influential decision you can make on your portfolio's performance. Only consider taxes after you have configured your total portfolio.
If you (or anyone) can explain Figure 2 on p. 11 in terms actionable for investors, that would be good....
curmudgeon
Posts: 2630
Joined: Thu Jun 20, 2013 11:00 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by curmudgeon »

FiveK wrote: Sun Oct 03, 2021 10:34 pm
yog wrote: Sun Oct 03, 2021 3:48 pm Though it addresses a slightly different aspect of retirement planning, here's a interesting recent scholarly paper written by someone McQ may know where Bellman equations are used extensively to support their research: Combining Investment and Tax Strategies for Optimizing Lifetime Solvency under Uncertain Returns and Mortality
Thanks for posting the link - interesting paper.

Seems in accord with the wiki article on Tax-efficient fund placement:
Determination of your asset allocation (% stocks / % bonds), which sets your portfolio's level of acceptable risk, is the single most influential decision you can make on your portfolio's performance. Only consider taxes after you have configured your total portfolio.
If you (or anyone) can explain Figure 2 on p. 11 in terms actionable for investors, that would be good....
That paper seemed interesting at first, then I realized it was just using a fixed 20-year period for investment returns to feed into the analysis. Nearly all the "model portfolios" were short international (because international didn't do as well over that period). There may be interesting modeling techniques brought into play, but the applicability to an investor today (as opposed to one in 1998) seems missing.
yog
Posts: 659
Joined: Wed Jan 15, 2020 11:57 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by yog »

curmudgeon wrote: Sun Oct 03, 2021 11:34 pm
FiveK wrote: Sun Oct 03, 2021 10:34 pm
yog wrote: Sun Oct 03, 2021 3:48 pm Though it addresses a slightly different aspect of retirement planning, here's a interesting recent scholarly paper written by someone McQ may know where Bellman equations are used extensively to support their research: Combining Investment and Tax Strategies for Optimizing Lifetime Solvency under Uncertain Returns and Mortality
Thanks for posting the link - interesting paper.

Seems in accord with the wiki article on Tax-efficient fund placement:
Determination of your asset allocation (% stocks / % bonds), which sets your portfolio's level of acceptable risk, is the single most influential decision you can make on your portfolio's performance. Only consider taxes after you have configured your total portfolio.
If you (or anyone) can explain Figure 2 on p. 11 in terms actionable for investors, that would be good....
That paper seemed interesting at first, then I realized it was just using a fixed 20-year period for investment returns to feed into the analysis. Nearly all the "model portfolios" were short international (because international didn't do as well over that period). There may be interesting modeling techniques brought into play, but the applicability to an investor today (as opposed to one in 1998) seems missing.
The authors explore an age-based and resource-based AA glide path for investors at different resource levels and optimizing around solvency. The portfolio targets are shown in Table 2 on page 9, ranked from 0 (91% bonds/9% stocks) to 14 (7% bonds/93% stocks). This is approximate, they have more exact figures to include their international targets, and yes they do short international with higher stock allocations. I wouldn't get too hung up about the exact construction as long as you could maintain a portfolio along the efficient frontier. Figure 2 is then the graphical representation of the intersection for the efficient frontier at a given age and resource level to ensure solvency. Once you set your resource and age-based AA, then consider tax strategy.

Our process actually followed something along this route - first, set an age-based AA allocation (70/30 since I was 49); then optimize SS claiming strategy (early due to some plan complexities) & tax-optimized income sequencing, which due to my younger age at retirement and our state of residence is a bit closer to the informed path (TDA, taxable, Roth); finally, Roth conversions supporting our optimal income sequencing (22% or 24%, ACA makes 22% ugly); then I was stuck, and had to revisit and tweak our AA to fit the optimal Roth conversion path (24% requires more cash, so 90%/10% liquidity = 70/30). Not really BH, but sharing because it is informative. We are the 'special case'.

Another thing I found interesting about this paper is how similar their optimal strategy appears to Kitces model of a lifetime tax equilibrium rate or possibly what the performance may be of a tax-adjusted portfolio outcome. There are also a number of good citations for ancillary reading, though they will all suffer from certain assumptions we may find objectionable for our personal situation.
ThankYouJack
Posts: 5704
Joined: Wed Oct 08, 2014 7:27 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by ThankYouJack »

McQ wrote: Sun Oct 03, 2021 5:29 pm
ThankYouJack wrote: Sat Oct 02, 2021 5:06 pm I'm still trying to wrap my head around this. I know some on here are ultra-wealthy and ultra-savers but how likely is it that one would reinvest 100% of their RMDs from the time they start until the day they die. If that's the plan, what sort of portfolio withdrawal rates are we talking -- extremely small or even negative? And wouldn't it make sense to give away more of the wealth while one is alive?
...
[quoting only ThankYouJack here, but the issue has come up several times, so here goes]

Keep in mind: in the spreadsheet for this thread, everything is off-camera except the results for converting / not-converting $100,000, assumed to be a fraction of the total TDA at the time. Something generated $108,750, to put the taxpayer into the 22% bracket when the RMDs begin; or, something generated the $200,450 to put the taxpayer into the 24% bracket; but those other sources of income, which might include many tens of thousands in RMDs on the remainder of the TDA, or much less, are purposely not visible.

The beauty of that approach: I don’t have to care whether the taxpayer is ChrisC, with the $200,450 coming mostly from social security and pensions, or Rob and Sue in the SSRN paper, with all of it coming from social security and RMDs on a $4 million TDA. To enter 22% as the tax on avoided RMDs—the RMDs reduced by the conversion—there just needs to be $108,750 (or more) coming from some income source(s)—whatever these happen to be.

Which brings us to “How could someone invest all their RMDs as surplus funds for the rest of their life? It makes no sense.” That misunderstands the spreadsheet set up. The spreadsheet simply says, if you don’t convert $100,000, you are going to have RMDs from that $100,000, and they are going to be taxed at the indicated rate. The first year RMD on a $100,000 portion of the TDA is $3,650. I can certainly imagine a family with $108,750 of other income, who needs every penny of that extra $3,650 to live (=total taxable income of $112,400 to scrape by). But really … remind me again why they wanted to reduce the amount of RMDs subject to tax?

Throughout I’ve made the tacit assumption, which I will have to make explicit when I edit the opening post, that the conversion is partial, so that it is made on the margin. Since $100,000 is only a portion of the TDA, the RMDs being reduced are the marginal RMDs, the last $3650 of RMDs, which if no conversion is done, will be taxed at 22% or 24%, the stated tax rate avoided via conversion in the case of most of the scenarios examined.

That’s all that “reinvested as surplus” implies. Tens of thousands of dollars of other RMDs might be taken and spent, as part of the $108,750 or $200,450 of income required to push the taxpayer into the 22% or 24% bracket prior to the marginal RMDs, which could have been removed by conversion.

PS. The dirty little secret of RMD reduction: it takes huge conversions to materially reduce the RMD dollar amount (during the 70s, when the withdrawal rate is about 4%). In the running example: it took a $100,000 conversion to reduce taxable income by ... wait for it ... all of $3650. To drop all the way down from top of the pre-IRMAA 22% bracket ($176,000) to the bottom ($108,750), would require conversions adding up to $1.842 million dollars. That's another reason why I like the new spreadsheet: a $100,000 conversion at the margin seems rather more realistic.
Thanks for the follow up. I read the paper but haven't had a chance to go through the spreadsheet. I'm still torn on if I should do a conversion this year. I may have missed it, but do certain tax credits also get factored in?
Topic Author
McQ
Posts: 1414
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by McQ »

ThankYouJack wrote: Mon Oct 04, 2021 11:37 am
McQ wrote: Sun Oct 03, 2021 5:29 pm
ThankYouJack wrote: Sat Oct 02, 2021 5:06 pm I'm still trying to wrap my head around this. I know some on here are ultra-wealthy and ultra-savers but how likely is it that one would reinvest 100% of their RMDs from the time they start until the day they die. If that's the plan, what sort of portfolio withdrawal rates are we talking -- extremely small or even negative? And wouldn't it make sense to give away more of the wealth while one is alive?
...
[quoting only ThankYouJack here, but the issue has come up several times, so here goes]

Keep in mind: in the spreadsheet for this thread, everything is off-camera except the results for converting / not-converting $100,000, assumed to be a fraction of the total TDA at the time. Something generated $108,750, to put the taxpayer into the 22% bracket when the RMDs begin; or, something generated the $200,450 to put the taxpayer into the 24% bracket; but those other sources of income, which might include many tens of thousands in RMDs on the remainder of the TDA, or much less, are purposely not visible.

The beauty of that approach: I don’t have to care whether the taxpayer is ChrisC, with the $200,450 coming mostly from social security and pensions, or Rob and Sue in the SSRN paper, with all of it coming from social security and RMDs on a $4 million TDA. To enter 22% as the tax on avoided RMDs—the RMDs reduced by the conversion—there just needs to be $108,750 (or more) coming from some income source(s)—whatever these happen to be.

Which brings us to “How could someone invest all their RMDs as surplus funds for the rest of their life? It makes no sense.” That misunderstands the spreadsheet set up. The spreadsheet simply says, if you don’t convert $100,000, you are going to have RMDs from that $100,000, and they are going to be taxed at the indicated rate. The first year RMD on a $100,000 portion of the TDA is $3,650. I can certainly imagine a family with $108,750 of other income, who needs every penny of that extra $3,650 to live (=total taxable income of $112,400 to scrape by). But really … remind me again why they wanted to reduce the amount of RMDs subject to tax?

Throughout I’ve made the tacit assumption, which I will have to make explicit when I edit the opening post, that the conversion is partial, so that it is made on the margin. Since $100,000 is only a portion of the TDA, the RMDs being reduced are the marginal RMDs, the last $3650 of RMDs, which if no conversion is done, will be taxed at 22% or 24%, the stated tax rate avoided via conversion in the case of most of the scenarios examined.

That’s all that “reinvested as surplus” implies. Tens of thousands of dollars of other RMDs might be taken and spent, as part of the $108,750 or $200,450 of income required to push the taxpayer into the 22% or 24% bracket prior to the marginal RMDs, which could have been removed by conversion.

PS. The dirty little secret of RMD reduction: it takes huge conversions to materially reduce the RMD dollar amount (during the 70s, when the withdrawal rate is about 4%). In the running example: it took a $100,000 conversion to reduce taxable income by ... wait for it ... all of $3650. To drop all the way down from top of the pre-IRMAA 22% bracket ($176,000) to the bottom ($108,750), would require conversions adding up to $1.842 million dollars. That's another reason why I like the new spreadsheet: a $100,000 conversion at the margin seems rather more realistic.
Thanks for the follow up. I read the paper but haven't had a chance to go through the spreadsheet. I'm still torn on if I should do a conversion this year. I may have missed it, but do certain tax credits also get factored in?
Imperative to factor in all the credits, cliff-edges etc. that will be lost / triggered by the conversion. See the post by FiveK just a bit up thread for a practical approach to the calculations.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Topic Author
McQ
Posts: 1414
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by McQ »

Spreadsheet posted

I've made available a version of the spreadsheet I've been using. [Edit 12-8-2021] It can be downloaded here: https://github.com/emcquarrie/Roth-conv ... readsheets

It has the metrics I introduced a few posts back and that I will be using to assess "plausible best cases" for converting at 22% in the next planned post.

As always, corrections, issues, points of contention, and suggestions are all welcome. Future versions will add tabs to assess some of the topics I've been deferring, such as the impact of paying tax from outside the conversion, or the effect of not realizing capital gains every year.

This version is limited to addressing present versus future tax rates, and the effect of varying the rate of return (not yet discussed here, but the setup allows it).
Last edited by McQ on Wed Dec 08, 2021 1:25 pm, edited 1 time in total.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Topic Author
McQ
Posts: 1414
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by McQ »

How good can it get?

A few posts back I identified “plausible best cases” for outcomes from converting to Roth at a tax rate of 22%. These were:
1. Avoiding a post-TCJA tax rate of 25%
2. Avoiding a future rate of 28% (next bracket up, post TCJA)
3. Avoiding an IRMAA penalty combined rate (post-TCJA) of 32%
4. Avoiding an SS tax torpedo of 40.7% (or 46.25% post-TCJA).
The benchmark for these will be conversion at a constant tax rate of 22%. And the comparison case will be the “plausible worst case” outcome of converting at 22% to avert RMDs that were only going to be taxed at 12%, also known as a doing a Bob-and-Barb (i.e., failing to adjust future tax brackets for inflation).

I also introduced several metrics for evaluation, explained upthread; and I said I would stick to outcomes likely to be achieved while alive. That will be defined here as the outcome achieved at age 85 or 95.

If you want to follow along at home, the spreadsheet used to generate these outcomes can be downloaded here: [edit 12-8-2021] https://github.com/emcquarrie/Roth-conv ... readsheets

Here is the first code block showing outcomes at a constant rate of 22%.

Code: Select all

      Real      Ratio    Tax               Wealth
      Roth      to tax   debit    Wealth   ratio
Age   surplus   debit    ROI      ratio    ROI
------------------------------------------------
85    $ 7,629    0.35    2.15%    1.04     0.28%
95    $42,190    1.92    4.56%    1.13     0.49%
Consistent with my initial posts on this thread, even when rates stay constant a Roth conversion can pay off. But it’s a slow process, with not much return after fourteen years; only out in the 90s does the payoff start to amount to something.

The most likely “win”

The statute says that the 22% rate will lapse and return to its pre-TCJA value of 25% in 2026. Here are the metrics:

Code: Select all

      Real      Ratio    Tax               Wealth
      Roth      to tax   debit    Wealth   ratio
Age   surplus   debit    ROI      ratio    ROI
------------------------------------------------

85    $14,868    0.68    3.76%    1.08     0.57%
95    $55,103    2.50    5.36%    1.17     0.66%
This small change in future rate has a noticeable impact: the real Roth surplus at age 85 doubles relative to constant rates. Interestingly, the incremental impact is comparatively less at age 95, and this “not that much better” result at 95 holds across metrics. I’ll interpret it later.

A somewhat less likely but sweeter outcome

If you are converting during a temporary valley in your income, and you have been a doughty saver, you may reasonably fear that absent a conversion at 22%, you will be in the next bracket up on retirement, and post-TCJA that will be at 28% (we’ll ignore IRMAA, see my thread on pensions). That’s a six point rate hike, double the three points just investigated. Here are the metrics.

Code: Select all

      Real      Ratio    Tax               Wealth
      Roth      to tax   debit    Wealth   ratio
Age   surplus   debit    ROI      ratio    ROI
------------------------------------------------
85    $22,106    1.00    5.09%    1.13     0.86%
95    $68,017    3.09    6.05%    1.22     0.83%
Well, nothing has doubled since the prior scenario. And the improvement again is comparatively less at age 95 than at age 85, where it is up 50% from the 22-->25 case. Interestingly, the most stringent metric, the rightmost column, actually decreases slightly at age 95. Translation: the improvement in wealth in the ten years after 85 isn’t coming fast enough.

A not much less likely but sweeter still outcome

Post TCJA, IRMAA penalties might add 4% to the income tax rate. Who wouldn’t want to convert at 22% to save 32%? Here are the metrics.

Code: Select all

      Real      Ratio    Tax               Wealth
      Roth      to tax   debit    Wealth   ratio
Age   surplus   debit    ROI      ratio    ROI
------------------------------------------------
85    $31,757    1.44    6.59%    1.19     1.27%
95    $85,235    3.87    6.82%    1.29     1.07%
Hmmh. Results have improved once more, but the numbers are hardly … striking. Good news: the pattern at age 95 now has an explanation. Each 1% increase in tax rate relative to the base case of constant rates adds about $4333 to the age 95 real Roth surplus.* By this point the rate is up 10 points, and the age 95 surplus is up about $43,000 relative to the constant case. Bad news: according to the most stringent metric on the right, the ROI is noticeably lower at age 95 than 85.
*At 85, it is about $2400 per point of rate increase; remember, these are real dollars.

The Big One: Dodging the SS Tax Torpedo

Using post-TCJA rates, if you have enough social security, and just the right amount of other income, each dollar of income gets taxed itself plus brings $0.85 of your social security payment into the taxable fold. Your marginal tax rate becomes 1.85X the income tax rate of 25%, or 46.25% in all. Owww….

Code: Select all

      Real      Ratio    Tax               Wealth
      Roth      to tax   debit    Wealth   ratio
Age   surplus   debit    ROI      ratio    ROI
------------------------------------------------
85    $66,139    3.01   10.42%    1.51     2.99%
95   $146,574    6.66    8.86%    1.63     2.07%
And no surprise, these are the best results thus far. To put them in context, here is a summary of the raw portfolio values that generate the metrics (these are nominal values while the tables above gave real values).

Code: Select all

                 Roth       TDA pretax              TDA post-tax
                 balance    balance      Taxable    balance
22% constant     $296,205   $204,335     $125,284   $159,381
22% --> 46.25%   $296,205   $204,335     $ 86,333   $109,830
As can be seen, the Roth and pretax TDA values don’t change. Rather, the taxable balance is lower because so much more of each RMD is lost to tax and not available for reinvestment. Likewise, the post-tax value of the TDA is that much lower under higher rates of tax.

I should point out that these tax torpedo results are probably overly rosy. They assume you stay in the tax torpedo even as you liquidate hundreds of thousands of dollars of TDA for the evaluation—which is highly unlikely, because the torpedo is generally a window that gets exited once income hits a certain point.

Summary

Looking over the entire set, I draw the following conclusions about the expected payoff from Roth conversions under various tax rate scenarios:
1. If future rates move higher, the payoff becomes substantial much sooner, in the 80s.
2. But at later ages, 95 and up, the gain is less, because tax drag has had a chance to grow large even under the constant rate case. Tax drag on the reinvested RMDs is an independent source of payoff.
3. The payoff in the 80s remains modest even if future rates jump ten points. The wealth ratio (2nd column from the right) is still less than 1.2. Likewise, focusing on the tax debit ratio, to earn the debit as young as 85 requires a jump of 6 points in the rate.

Glass half full, or half empty?

In many cases a Roth conversion amounts to a simple bookkeeping change plus a check made out to the IRS. The VTI in one account is translated into a VTI in another account, minus a debit. Not a lot of effort, nor much in the way of risk other than the tax debit, which itself is just a change in the timing of tax. Almost any positive impact might be judged sufficient to motivate these bookkeeping and timing changes.

On the other hand, many folks contemplating a Roth conversion might look at these numbers and find themselves humming, Is that all there is? Yes there is a payoff; but in the context of income levels of $110,000 plus—the floor of the 22% bracket—and TDA assets of $1 million to $4 million, depending on pension—is the payoff from a $100,000 conversion enough to get excited about? Only you can decide.

Another way to approach the metrics would be to calculate the risk/reward matrix across the three most likely outcomes: 1) no change in rates; 2) post-TCJA increase to 25%; 3) miscalculation, where future rates drop to 12%. Here it is at age 85, sticking to the metric of the real dollar surplus.

Code: Select all

           Real Roth surplus 
Rates      at age 85:
22 --> 12  -$16,499
22 --> 22   $ 7,629
22 --> 25   $14,868
If you are careful about tax brackets and inflation, you can be reasonably certain of not doing a Bob-and-Barb, and thus not ending up in the 12% bracket and losing $16,499 on the conversion. Or you can say, “this conversion is risky, because on my current projection RMDs are only just barely going to lift me into the 22% bracket. Just a few tweaks downward in the variables, and I’m going to find myself like Bob and Barb.” In which case, don’t do the conversion.

If the negative outcome can be ruled out as unlikely, then the conversion decision is straightforward: in the not unlikely case of no change in rates, the $100,000 conversion will put you $7000 ahead after fourteen years; best probable case, $14,000; even as there is always the small chance that it could be $20,000 or $30,000, if the markets prove unexpectedly rewarding and you do get pushed into the next tax bracket up.

Likewise, the SS Tax Torpedo requires that you be comfortably into the 22% tax bracket. If the amount of your social security can be projected reasonably well, along with your expected pension if any, and those two are going to bring you to the 22% floor, then almost any non-zero amount of RMDs might fall into the tax torpedo, with a big payoff from the conversion.

But before you start licking your lips, remember: if your social security as a couple is too low (<$35,000 or so), it will hit the 85% taxation ceiling before other income gets you to the 22% bracket. You won’t be in the top half of the torpedo. And if your total income is too high, you’ll exit the torpedo on the other side and again, will not be paying a marginal rate that begins with 4.

Run the numbers.
Last edited by McQ on Wed Dec 08, 2021 1:26 pm, edited 1 time in total.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
User avatar
FiveK
Posts: 15692
Joined: Sun Mar 16, 2014 2:43 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by FiveK »

McQ wrote: Thu Oct 07, 2021 4:41 pm If you want to follow along at home, the spreadsheet used to generate these outcomes can be downloaded here: http://www.edwardfmcquarrie.com/wp-cont ... sheet.xlsx
Firefox displays "File not downloaded: Potential security risk." for this link when attempting to open in Excel. I suspect that warning is due to the stereotypical "(over)abundance of caution" but you might look at using Google drive or other hosting mechanism to eliminate that concern.
Here is the first code block showing outcomes at a constant rate of 22%.

Code: Select all

      Real      Ratio    Tax               Wealth
      Roth      to tax   debit    Wealth   ratio
Age   surplus   debit    ROI      ratio    ROI
------------------------------------------------
85    $ 7,629    0.35    2.15%    1.04     0.28%
95    $42,190    1.92    4.56%    1.13     0.49%
<large snip>

Run the numbers.
For more number running, it seems one can configure inputs on the spreadsheet at How do RMDs affect Roth conversion choices? (aka "MDM spreadsheet") to reproduce these results:

Code: Select all

+────────────────────+──────────+
| Original owner     |          |
+────────────────────+──────────+
| RMD Start age      | 72       |
| Roth conv. age     | 71       |
| Roth conv. amount  | $90,909  |
| 1st yr tax rate    | 22.0%    |
| Pay tax w/ cash    | 0        |
| Basis fract.       | 100%     |
| EOY conversion?    | 1        |
| EOY RMD?           | 1        |
|                    |          |
|                    | Original |
| Annual tax rate    | 22.00%   |
| Stocks             | 100%     |
| Bonds              | 0%       |
| Turnover           | 100%     |
| Cap gain           | 10%      |
| Dividend           | 0%       |
| Interest           | 0%       |
| Tax-ad gain        | 10%      |
| CG tax             | 15.0%    |
| Div. tax           | 15.0%    |
+────────────────────+──────────+
Some notes:
- the $90,909 is actually "=100000/1.1" to get things aligned at the start
- the "turnover" assumption has some effect
- assuming a 100% turnover for capital gains means the individual assumptions for cap gain and dividend are irrelevant: only the sum matters
- the "Pay tax w/ cash" assumption has a significant effect, especially with Basis fract = 100% but also if Basis fract = 0%
- the MDM spreadsheet seems to assume all inputs and results are "real". Thus one should compare the E2/I11 result there with column O in McQ's model.
hppycamper
Posts: 49
Joined: Sat May 22, 2021 8:41 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by hppycamper »

McQ wrote: Thu Oct 07, 2021 4:41 pm
If the negative outcome can be ruled out as unlikely, then the conversion decision is straightforward: in the not unlikely case of no change in rates, the $100,000 conversion will put you $7000 ahead after fourteen years; best probable case, $14,000;
Thank you, Prof. McQ, for another rule we can use for high level planning/projection, without spending hours upon hours sifting through details :D
Topic Author
McQ
Posts: 1414
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by McQ »

FiveK wrote: Thu Oct 07, 2021 5:45 pm
McQ wrote: Thu Oct 07, 2021 4:41 pm If you want to follow along at home, the spreadsheet used to generate these outcomes can be downloaded here: http://www.edwardfmcquarrie.com/wp-cont ... sheet.xlsx
Firefox displays "File not downloaded: Potential security risk." for this link when attempting to open in Excel. I suspect that warning is due to the stereotypical "(over)abundance of caution" but you might look at using Google drive or other hosting mechanism to eliminate that concern.
Well that's annoying--thank you for bringing it to my attention. My (university) copy of google drive doesn't let me open files to "anyone who has the link." That's why I switched to postimages.com early in the thread for ss images.

Is there an equivalent of postimages.com for ss files? Happy to post a copy there instead of on my web site.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
curmudgeon
Posts: 2630
Joined: Thu Jun 20, 2013 11:00 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by curmudgeon »

McQ wrote: Thu Oct 07, 2021 4:41 pm How good can it get?

The Big One: Dodging the SS Tax Torpedo

Using post-TCJA rates, if you have enough social security, and just the right amount of other income, each dollar of income gets taxed itself plus brings $0.85 of your social security payment into the taxable fold. Your marginal tax rate becomes 1.85X the income tax rate of 25%, or 46.25% in all. Owww….

Code: Select all

      Real      Ratio    Tax               Wealth
      Roth      to tax   debit    Wealth   ratio
Age   surplus   debit    ROI      ratio    ROI
------------------------------------------------
85    $66,139    3.01   10.42%    1.51     2.99%
95   $146,574    6.66    8.86%    1.63     2.07%
......

I should point out that these tax torpedo results are probably overly rosy. They assume you stay in the tax torpedo even as you liquidate hundreds of thousands of dollars of TDA for the evaluation—which is highly unlikely, because the torpedo is generally a window that gets exited once income hits a certain point.
.......

On the other hand, many folks contemplating a Roth conversion might look at these numbers and find themselves humming, Is that all there is? Yes there is a payoff; but in the context of income levels of $110,000 plus—the floor of the 22% bracket—and TDA assets of $1 million to $4 million, depending on pension—is the payoff from a $100,000 conversion enough to get excited about? Only you can decide.
......

If the negative outcome can be ruled out as unlikely, then the conversion decision is straightforward: in the not unlikely case of no change in rates, the $100,000 conversion will put you $7000 ahead after fourteen years; best probable case, $14,000; even as there is always the small chance that it could be $20,000 or $30,000, if the markets prove unexpectedly rewarding and you do get pushed into the next tax bracket up.

Likewise, the SS Tax Torpedo requires that you be comfortably into the 22% tax bracket. If the amount of your social security can be projected reasonably well, along with your expected pension if any, and those two are going to bring you to the 22% floor, then almost any non-zero amount of RMDs might fall into the tax torpedo, with a big payoff from the conversion.

But before you start licking your lips, remember: if your social security as a couple is too low (<$35,000 or so), it will hit the 85% taxation ceiling before other income gets you to the 22% bracket. You won’t be in the top half of the torpedo. And if your total income is too high, you’ll exit the torpedo on the other side and again, will not be paying a marginal rate that begins with 4.

Run the numbers.
I think this post hits the key points. The additional problem of a survivor having to take RMDs and file in a higher single bracket is real, but with so much uncertainty about how long that period might be I'm not sure it's worth going into the weeds there.

A few notes:

1) There can actually be a "window within the window" of the SS tax torpedo with even higher marginal rates. If you have qualified dividends/LTCG, the movement of $.85 of SS into taxation can also move $.85 of qualified dividends from 0% to 15% tax rates. This adds another 12.75% to the marginal rate, potentially taking it up to 59%. This tends to be a much smaller window, though, and depends on having just the right (wrong?) mix of income types and amounts.

2) The SS torpedo window is by no means intuitive, nor is the effect of inflation on it over time. Smaller SS benefits or larger pensions may make it irrelevant (and large enough RMDs will take the hit and blow right past it). I think the largest effect is on couples with larger SS amounts. I did a quick look at a couple with $90,000 combined SS, $2,000 interest, and $10,000 dividends. In this case, it looked like the SS torpedo window for additional income (whether RMD or other ordinary income) ran from $5,000 to $17,000 at 1.5X, and from $17,000 to $70,000 at 1.85X. Note that in this case the baseline for the multiplier starts below the 25/22% bracket, so some of the torpedo might only involve boosting the effective marginal rate by 8.5% or 12%. Still, for someone in the right circumstances, that's a $65,000 window.
marcopolo
Posts: 8411
Joined: Sat Dec 03, 2016 9:22 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by marcopolo »

hppycamper wrote: Thu Oct 07, 2021 10:40 pm
McQ wrote: Thu Oct 07, 2021 4:41 pm
If the negative outcome can be ruled out as unlikely, then the conversion decision is straightforward: in the not unlikely case of no change in rates, the $100,000 conversion will put you $7000 ahead after fourteen years; best probable case, $14,000;
Thank you, Prof. McQ, for another rule we can use for high level planning/projection, without spending hours upon hours sifting through details :D
Just remember that this is only true if you compare it to the case where you would turn over all your gains in the taxable account, and pay taxes on them, each year. I doubt very many people do that.

You may want to wait until Prof. McQ gets a chance to add more realistic assumption about portfolio turnover into his model before you run off and use this for planning, even at a high level.
Once in a while you get shown the light, in the strangest of places if you look at it right.
Topic Author
McQ
Posts: 1414
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by McQ »

curmudgeon wrote: Thu Oct 07, 2021 11:32 pm
McQ wrote: Thu Oct 07, 2021 4:41 pm How good can it get?

The Big One: Dodging the SS Tax Torpedo

Using post-TCJA rates, if you have enough social security, and just the right amount of other income, each dollar of income gets taxed itself plus brings $0.85 of your social security payment into the taxable fold. Your marginal tax rate becomes 1.85X the income tax rate of 25%, or 46.25% in all. Owww….

Code: Select all

      Real      Ratio    Tax               Wealth
      Roth      to tax   debit    Wealth   ratio
Age   surplus   debit    ROI      ratio    ROI
------------------------------------------------
85    $66,139    3.01   10.42%    1.51     2.99%
95   $146,574    6.66    8.86%    1.63     2.07%
......

I should point out that these tax torpedo results are probably overly rosy. They assume you stay in the tax torpedo even as you liquidate hundreds of thousands of dollars of TDA for the evaluation—which is highly unlikely, because the torpedo is generally a window that gets exited once income hits a certain point.
.......

On the other hand, many folks contemplating a Roth conversion might look at these numbers and find themselves humming, Is that all there is? Yes there is a payoff; but in the context of income levels of $110,000 plus—the floor of the 22% bracket—and TDA assets of $1 million to $4 million, depending on pension—is the payoff from a $100,000 conversion enough to get excited about? Only you can decide.
......

If the negative outcome can be ruled out as unlikely, then the conversion decision is straightforward: in the not unlikely case of no change in rates, the $100,000 conversion will put you $7000 ahead after fourteen years; best probable case, $14,000; even as there is always the small chance that it could be $20,000 or $30,000, if the markets prove unexpectedly rewarding and you do get pushed into the next tax bracket up.

Likewise, the SS Tax Torpedo requires that you be comfortably into the 22% tax bracket. If the amount of your social security can be projected reasonably well, along with your expected pension if any, and those two are going to bring you to the 22% floor, then almost any non-zero amount of RMDs might fall into the tax torpedo, with a big payoff from the conversion.

But before you start licking your lips, remember: if your social security as a couple is too low (<$35,000 or so), it will hit the 85% taxation ceiling before other income gets you to the 22% bracket. You won’t be in the top half of the torpedo. And if your total income is too high, you’ll exit the torpedo on the other side and again, will not be paying a marginal rate that begins with 4.

Run the numbers.
I think this post hits the key points. The additional problem of a survivor having to take RMDs and file in a higher single bracket is real, but with so much uncertainty about how long that period might be I'm not sure it's worth going into the weeds there.

A few notes:

1) There can actually be a "window within the window" of the SS tax torpedo with even higher marginal rates. If you have qualified dividends/LTCG, the movement of $.85 of SS into taxation can also move $.85 of qualified dividends from 0% to 15% tax rates. This adds another 12.75% to the marginal rate, potentially taking it up to 59%. This tends to be a much smaller window, though, and depends on having just the right (wrong?) mix of income types and amounts.

2) The SS torpedo window is by no means intuitive, nor is the effect of inflation on it over time. Smaller SS benefits or larger pensions may make it irrelevant (and large enough RMDs will take the hit and blow right past it). I think the largest effect is on couples with larger SS amounts. I did a quick look at a couple with $90,000 combined SS, $2,000 interest, and $10,000 dividends. In this case, it looked like the SS torpedo window for additional income (whether RMD or other ordinary income) ran from $5,000 to $17,000 at 1.5X, and from $17,000 to $70,000 at 1.85X. Note that in this case the baseline for the multiplier starts below the 25/22% bracket, so some of the torpedo might only involve boosting the effective marginal rate by 8.5% or 12%. Still, for someone in the right circumstances, that's a $65,000 window.
Hello Curmudgeon: I have to admit to playing catchup on social security taxation in the four months I have been on site. As you remark, “The SS torpedo window is by no means intuitive.” I have had to learn a lot.

But I would like to challenge you on the “window within the window.” I see the logic of your 59% marginal rate; but I am not able to accept it as a real case, absent a spreadsheet demonstration. I would need to see a column with raw social security as the top cell, then other ordinary income (or qualified dividends) sufficient to get the total taxable income up to the 22% tax bracket floor of 108,750 (an intermediate row would show how much of SS had become taxable at that point). Once the total taxable ordinary income hits 108,750, that triggers the 1.85X 22% top rate torpedo. But the SS income would have to not yet be 85% taxable, the max, for the next $1000 to be taxable at 1.85X.

Next cell can be qualified dividend income. Side by side columns to the right can step up one or another income $1000 at a time. Rows below can calculate tax.

You need to show a pair of adjacent columns where +1000 of income triggers +590 of tax. I am skeptical, but willing to be taught.
Perfectly okay for the window you hypothesize to be exceedingly narrow. Any window $1000 or more wide is scored as a win for your position re the 59% rate.

I’ll take up the realism (or lack thereof) of your second example, the couple who earned $90K in SS but had no TDA at all, in a later post.

PS: if on Windows, a snip screen of the spreadsheet, uploaded to postimages.com, is a simple way to display your ss results in a post where all can see.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
User avatar
FiveK
Posts: 15692
Joined: Sun Mar 16, 2014 2:43 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by FiveK »

McQ wrote: Thu Oct 07, 2021 10:55 pm Is there an equivalent of postimages.com for ss files? Happy to post a copy there instead of on my web site.
This may be a blind leading the blind situation, but I'll try. ;)

I believe that one can use a personal google drive to make files available "to anyone with the link". That's what I do when Posting images in the Bogleheads forum. You wouldn't need all the extra items specific to this forum and images, just the shareable link.

When I say "personal" I mean one under a "FiveK" google account, not my actual personal gmail.
User avatar
FiveK
Posts: 15692
Joined: Sun Mar 16, 2014 2:43 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by FiveK »

McQ wrote: Fri Oct 08, 2021 12:40 am But I would like to challenge you on the “window within the window.” I see the logic of your 59% marginal rate; but I am not able to accept it as a real case, absent a spreadsheet demonstration.
Don't know about 59%, but below is a quick picture of a >50% marginal rate:

Image

One should be able to reproduce it using the personal finance toolbox in Excel with the following entries in cells on the Calculations tab. The chart starts near cell F82.
B38: 40000
G2: 1
G9: 70
H35: KS
P83: 110
...and standard Excel changing of the y-axis scale.

No idea if the KS state tax is calculated accurately but it seems plausible after some quick internet searching....

ETA: adding capital gains or qualified dividends would likely push the marginal rate higher, as shown in a single filer, age 66, receiving $30K/yr SS benefits and $20K/yr qualified dividends.
curmudgeon
Posts: 2630
Joined: Thu Jun 20, 2013 11:00 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by curmudgeon »

McQ wrote: Fri Oct 08, 2021 12:40 am
curmudgeon wrote: Thu Oct 07, 2021 11:32 pm
1) There can actually be a "window within the window" of the SS tax torpedo with even higher marginal rates. If you have qualified dividends/LTCG, the movement of $.85 of SS into taxation can also move $.85 of qualified dividends from 0% to 15% tax rates. This adds another 12.75% to the marginal rate, potentially taking it up to 59%. This tends to be a much smaller window, though, and depends on having just the right (wrong?) mix of income types and amounts.

2) The SS torpedo window is by no means intuitive, nor is the effect of inflation on it over time. Smaller SS benefits or larger pensions may make it irrelevant (and large enough RMDs will take the hit and blow right past it). I think the largest effect is on couples with larger SS amounts. I did a quick look at a couple with $90,000 combined SS, $2,000 interest, and $10,000 dividends. In this case, it looked like the SS torpedo window for additional income (whether RMD or other ordinary income) ran from $5,000 to $17,000 at 1.5X, and from $17,000 to $70,000 at 1.85X. Note that in this case the baseline for the multiplier starts below the 25/22% bracket, so some of the torpedo might only involve boosting the effective marginal rate by 8.5% or 12%. Still, for someone in the right circumstances, that's a $65,000 window.
Hello Curmudgeon: I have to admit to playing catchup on social security taxation in the four months I have been on site. As you remark, “The SS torpedo window is by no means intuitive.” I have had to learn a lot.

But I would like to challenge you on the “window within the window.” I see the logic of your 59% marginal rate; but I am not able to accept it as a real case, absent a spreadsheet demonstration. I would need to see a column with raw social security as the top cell, then other ordinary income (or qualified dividends) sufficient to get the total taxable income up to the 22% tax bracket floor of 108,750 (an intermediate row would show how much of SS had become taxable at that point). Once the total taxable ordinary income hits 108,750, that triggers the 1.85X 22% top rate torpedo. But the SS income would have to not yet be 85% taxable, the max, for the next $1000 to be taxable at 1.85X.
......

I’ll take up the realism (or lack thereof) of your second example, the couple who earned $90K in SS but had no TDA at all, in a later post.
That's a good challenge, thanks. It made me go back and examine the potential scenarios more closely to come up with more than "I remember this from a few years ago" (my first encounter with this was, in fact, the result of plugging in some guesstimates into TurboTax a few years ago, but memories can be deceptive).

The situation is confused a bit by the fact that SS benefit taxation is phased in based on gross income, but the shift to higher dividend/LTCG taxes is made based on taxable income. The 22% tax bracket floor is actually $81050 for MFJ taxable income, and this is also the point at which dividends/LTCG move from 0% to 15%. So the conditions for hitting the inner window become: taxable income > $81000 and dividends/LTCG > (taxable income - 81000) and SS still phasing in at 85%.

If we posit $80000 of combined SS benefits (slightly lower than I had used earlier), then the other income needed to meet the constraints tends to come in a range around $60000 to $77000. The mix of qualified dividends and RMD/other income is important (to meet the other constraints). If I plug in $30000 of qualified dividends, for RMDs in the range from about $30000 to $40000 this creates the appropriate scenario. In the current tax rates, this ought to give marginal rates of 1.85X 22% (for the extra RMD), plus .85X 15% (for the dividends moving to the 15% rate), for an overall marginal rate of 53.5%. However, when I plug these numbers into my favorite tax estimator ( https://www.mortgagecalculator.org/calc ... ulator.php), I only come out with 50% marginal rate across that range. If I translate that to the post-TCJA 25% bracket we were discussing, it comes to 56% marginal, not the 59% I expected. I have some more homework to do ...

--------------------------
edited to add:

So FiveK did my homework for me, and pointed out that the ordinary income (RMD etc) was NOT actually in the 22% bracket. The scenario comes out as an artifact of the way income is stacked and where the tax is computed. My head still spins a little, but the calculation comes out at 1.85X (12%+15%) to give 50% marginal at current rates. Post-TCJA gives 1.85X (15%+15%), or 55.5% marginal in this zone.

So McQ was correct in his challenge, but it does still turn out that there is a "window within the window", it's just that my methodology (and final marginal rate) were off.
Last edited by curmudgeon on Fri Oct 08, 2021 11:11 pm, edited 1 time in total.
User avatar
FiveK
Posts: 15692
Joined: Sun Mar 16, 2014 2:43 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by FiveK »

curmudgeon wrote: Fri Oct 08, 2021 7:15 pm If we posit $80000 of combined SS benefits (slightly lower than I had used earlier), then the other income needed to meet the constraints tends to come in a range around $60000 to $77000. The mix of qualified dividends and RMD/other income is important (to meet the other constraints). If I plug in $30000 of qualified dividends, for RMDs in the range from about $30000 to $40000 this creates the appropriate scenario. In the current tax rates, this ought to give marginal rates of 1.85X 22% (for the extra RMD), plus .85X 15% (for the dividends moving to the 15% rate), for an overall marginal rate of 53.5%.
In that scenario, ordinary income is still in the 12% bracket.

Thus one gets 1.85 * 12% + 1.85 * 15% = 49.95% for the maximum marginal federal tax rate in that area.

Change the 12% bracket to 15% and it's a 55.5% marginal rate. Of course, state taxes could come into play....
curmudgeon
Posts: 2630
Joined: Thu Jun 20, 2013 11:00 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by curmudgeon »

FiveK wrote: Fri Oct 08, 2021 7:56 pm
curmudgeon wrote: Fri Oct 08, 2021 7:15 pm If we posit $80000 of combined SS benefits (slightly lower than I had used earlier), then the other income needed to meet the constraints tends to come in a range around $60000 to $77000. The mix of qualified dividends and RMD/other income is important (to meet the other constraints). If I plug in $30000 of qualified dividends, for RMDs in the range from about $30000 to $40000 this creates the appropriate scenario. In the current tax rates, this ought to give marginal rates of 1.85X 22% (for the extra RMD), plus .85X 15% (for the dividends moving to the 15% rate), for an overall marginal rate of 53.5%.
In that scenario, ordinary income is still in the 12% bracket.

Thus one gets 1.85 * 12% + 1.85 * 15% = 49.95% for the maximum marginal federal tax rate in that area.

Change the 12% bracket to 15% and it's a 55.5% marginal rate. Of course, state taxes could come into play....
Thanks, I can see I was misinterpreting the way income stacks when the brackets are applied. As McQ was pointing out, you can't really be in the 22% bracket on ordinary income while still having some qualified dividends in the 0% bracket. I'll append a not on that in my prior response.
cas
Posts: 2245
Joined: Wed Apr 26, 2017 8:41 am

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by cas »

curmudgeon wrote: Fri Oct 08, 2021 7:15 pm My head still spins a little, but the calculation comes out at 1.85X (12%+15%) to give 50% marginal at current rates. Post-TCJA gives 1.85X (15%+15%), or 55.5% marginal in this zone.
I think of it with a "bumper car" or "falling domino" analogy.

- You add $1 of RMD/Roth conversion. (ordinary income; taxed in nominal 12% bracket = $1 x 12% = 12 cents tax)
- That bumps a new $0.85 of SS into being taxed (ordinary income; taxed in nominal 12% bracket = $0.85 x 12% = 10.2 cents tax)
- Then that $1.85 of cumulative new ordinary income bumps $1.85 of QD/LTCG from the 0% special QD/LTCG bracket to the 15% special QD/LTCG bracket (1.85 x 15% = 27.75 cents tax)

That means that (12 + 10.2 + 27.75) = 49.95 cents of new tax was added to your tax bill as a result of adding $1 in ordinary income via RMD or Roth conversion. That is a 49.95% marginal rate on that $1 of income.

You have to get just the right mix of RMD/Roth conversion + SS + QD/LTCG to get a stretch of 49.95% marginal rate, and it usually isn't all that wide and doesn't result in all that much actual additional tax, dollar-wise. No doubt exceptions exist.
Topic Author
McQ
Posts: 1414
Joined: Fri Jun 18, 2021 12:21 am
Location: California

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by McQ »

FiveK wrote: Fri Oct 08, 2021 7:56 pm
curmudgeon wrote: Fri Oct 08, 2021 7:15 pm If we posit $80000 of combined SS benefits (slightly lower than I had used earlier), then the other income needed to meet the constraints tends to come in a range around $60000 to $77000. The mix of qualified dividends and RMD/other income is important (to meet the other constraints). If I plug in $30000 of qualified dividends, for RMDs in the range from about $30000 to $40000 this creates the appropriate scenario. In the current tax rates, this ought to give marginal rates of 1.85X 22% (for the extra RMD), plus .85X 15% (for the dividends moving to the 15% rate), for an overall marginal rate of 53.5%.
In that scenario, ordinary income is still in the 12% bracket.

Thus one gets 1.85 * 12% + 1.85 * 15% = 49.95% for the maximum marginal federal tax rate in that area.

Change the 12% bracket to 15% and it's a 55.5% marginal rate. Of course, state taxes could come into play....
Thank you for the wiki link, FiveK, you are teaching me again. I would have had the window as (1.85 X ordinary rate) + dividend rate, since dividends go into MAGI even when not taxed. But I set up your wiki scenario in TurboTax, and it is as you have it: 1.85X (ordinary rate + dividend rate). Married couples may enjoy this Kitces post: https://www.kitces.com/blog/long-term-c ... in-0-rate/.

Curmudgeon, if you'll accept 55.5% as the post-TCJA rate on this piece of the SS torpedo, I'll score the exchange for you. I did not think it possible on Federal alone.

Next, contra cas, the window is more than wide enough to matter within the Roth conversion space. TurboTax had the additional income window running from just over $16,000 to just under $24,000, per the wiki, an $8000 band. Age 72 RMD on a $100,000 conversion is only $3650. So it's easy to imagine a super savvy Roth conversion that reduced RMDs within that window: thus converting at 22% to save $55.5% post-TCJA--even better than the more typical SS tax torpedo examples.

But it will take a sharp pencil, extremely accurate income projections, and just the right personal situation to harvest this gain.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
hppycamper
Posts: 49
Joined: Sat May 22, 2021 8:41 pm

Re: Why Roth conversions always pay off—if you can hold on long enough

Post by hppycamper »

marcopolo wrote: Fri Oct 08, 2021 12:26 am
hppycamper wrote: Thu Oct 07, 2021 10:40 pm
McQ wrote: Thu Oct 07, 2021 4:41 pm
If the negative outcome can be ruled out as unlikely, then the conversion decision is straightforward: in the not unlikely case of no change in rates, the $100,000 conversion will put you $7000 ahead after fourteen years; best probable case, $14,000;
Thank you, Prof. McQ, for another rule we can use for high level planning/projection, without spending hours upon hours sifting through details :D
Just remember that this is only true if you compare it to the case where you would turn over all your gains in the taxable account, and pay taxes on them, each year. I doubt very many people do that.

You may want to wait until Prof. McQ gets a chance to add more realistic assumption about portfolio turnover into his model before you run off and use this for planning, even at a high level.
:oops: Great point! Thank you for pointing it out!
Post Reply