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.
wrongfunds
Posts: 3187
Joined: Tue Dec 21, 2010 2:55 pm

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

Post by wrongfunds »

Chip wrote: Mon Oct 18, 2021 12:26 pm
wrongfunds wrote: Mon Oct 18, 2021 8:45 am
McQ wrote: Sun Oct 17, 2021 3:59 pm Next planned post: What if I pay the tax from outside the conversion?
I have fundamental problem with this issue. Why is there "outside" money in the first place? Did it have a specific reason to be there? As somebody else aptly put that money did NOT fall from the sky! What is individual giving up by depleting the "outside" money and how would you account for that?

Frankly, analysis based upon the above assumption is close to basing it on "assume you have unlimited money ..."
I don't understand the concern. Many people max out their tax-deferred options and save additional money in taxable accounts. That's where all the tax payments came from for our own Roth conversions.
But the taxable money need to have the taxes paid on (aka capital gains). My comment was regarding the "outside money" was for the cash.
User avatar
iceport
Posts: 6054
Joined: Sat Apr 07, 2007 4:29 pm

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

Post by iceport »

wrongfunds wrote: Mon Oct 18, 2021 12:45 pm
Chip wrote: Mon Oct 18, 2021 12:26 pm
wrongfunds wrote: Mon Oct 18, 2021 8:45 am
McQ wrote: Sun Oct 17, 2021 3:59 pm Next planned post: What if I pay the tax from outside the conversion?
I have fundamental problem with this issue. Why is there "outside" money in the first place? Did it have a specific reason to be there? As somebody else aptly put that money did NOT fall from the sky! What is individual giving up by depleting the "outside" money and how would you account for that?

Frankly, analysis based upon the above assumption is close to basing it on "assume you have unlimited money ..."
I don't understand the concern. Many people max out their tax-deferred options and save additional money in taxable accounts. That's where all the tax payments came from for our own Roth conversions.
But the taxable money need to have the taxes paid on (aka capital gains). My comment was regarding the "outside money" was for the cash.
I don't know how typical it is to do this, but I've maintained a "sinking fund" of sorts for a long time. While I was working, that sinking fund served dual roles as both an emergency fund and as a fund source for expected large expenses in the near term, like car replacements, major home repairs, etc. Now retired, I still hold that sinking fund, and it's even grown to serve as the savings I intend to use for major home improvements. I still make small monthly contributions to this fund from my main source of retirement income, a DB pension.

You may (correctly) call it mental accounting, but I don't count these savings as part of my retirement portfolio. Some of it is held in a credit union savings account, but most of it is in a short term bond fund. Capital gains taxes are negligible, but I pay taxes on interest and dividends (also, unfortunately, fairly negligible :( ).

When I do a Roth conversion, the taxes are paid from this sinking fund. That's my source of "outside money."
"Discipline matters more than allocation.” |—| "In finance, if you’re certain of anything, you’re out of your mind." ─William Bernstein
marcopolo
Posts: 8445
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 »

iceport wrote: Mon Oct 18, 2021 1:35 pm
wrongfunds wrote: Mon Oct 18, 2021 12:45 pm
Chip wrote: Mon Oct 18, 2021 12:26 pm
wrongfunds wrote: Mon Oct 18, 2021 8:45 am
McQ wrote: Sun Oct 17, 2021 3:59 pm Next planned post: What if I pay the tax from outside the conversion?
I have fundamental problem with this issue. Why is there "outside" money in the first place? Did it have a specific reason to be there? As somebody else aptly put that money did NOT fall from the sky! What is individual giving up by depleting the "outside" money and how would you account for that?

Frankly, analysis based upon the above assumption is close to basing it on "assume you have unlimited money ..."
I don't understand the concern. Many people max out their tax-deferred options and save additional money in taxable accounts. That's where all the tax payments came from for our own Roth conversions.
But the taxable money need to have the taxes paid on (aka capital gains). My comment was regarding the "outside money" was for the cash.
I don't know how typical it is to do this, but I've maintained a "sinking fund" of sorts for a long time. While I was working, that sinking fund served dual roles as both an emergency fund and as a fund source for expected large expenses in the near term, like car replacements, major home repairs, etc. Now retired, I still hold that sinking fund, and it's even grown to serve as the savings I intend to use for major home improvements. I still make small monthly contributions to this fund from my main source of retirement income, a DB pension.

You may (correctly) call it mental accounting, but I don't count these savings as part of my retirement portfolio. Some of it is held in a credit union savings account, but most of it is in a short term bond fund. Capital gains taxes are negligible, but I pay taxes on interest and dividends (also, unfortunately, fairly negligible :( ).

When I do a Roth conversion, the taxes are paid from this sinking fund. That's my source of "outside money."
You may not consider it a part of your portfolio (your personal mental accountingis fine as long as you are happy with it :beer )
But, for an analysis of the various benefits of Roth Conversions, it should be accounted for as the alternate investment if not paying those taxes, rather than magically falling out of the sky. I have no doubt Prof. McQ will treat it appropriately in his analysis.
Once in a while you get shown the light, in the strangest of places if you look at it right.
User avatar
iceport
Posts: 6054
Joined: Sat Apr 07, 2007 4:29 pm

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

Post by iceport »

marcopolo wrote: Mon Oct 18, 2021 1:51 pm
iceport wrote: Mon Oct 18, 2021 1:35 pm
wrongfunds wrote: Mon Oct 18, 2021 12:45 pm
Chip wrote: Mon Oct 18, 2021 12:26 pm
wrongfunds wrote: Mon Oct 18, 2021 8:45 am
I have fundamental problem with this issue. Why is there "outside" money in the first place? Did it have a specific reason to be there? As somebody else aptly put that money did NOT fall from the sky! What is individual giving up by depleting the "outside" money and how would you account for that?

Frankly, analysis based upon the above assumption is close to basing it on "assume you have unlimited money ..."
I don't understand the concern. Many people max out their tax-deferred options and save additional money in taxable accounts. That's where all the tax payments came from for our own Roth conversions.
But the taxable money need to have the taxes paid on (aka capital gains). My comment was regarding the "outside money" was for the cash.
I don't know how typical it is to do this, but I've maintained a "sinking fund" of sorts for a long time. While I was working, that sinking fund served dual roles as both an emergency fund and as a fund source for expected large expenses in the near term, like car replacements, major home repairs, etc. Now retired, I still hold that sinking fund, and it's even grown to serve as the savings I intend to use for major home improvements. I still make small monthly contributions to this fund from my main source of retirement income, a DB pension.

You may (correctly) call it mental accounting, but I don't count these savings as part of my retirement portfolio. Some of it is held in a credit union savings account, but most of it is in a short term bond fund. Capital gains taxes are negligible, but I pay taxes on interest and dividends (also, unfortunately, fairly negligible :( ).

When I do a Roth conversion, the taxes are paid from this sinking fund. That's my source of "outside money."
You may not consider it a part of your portfolio (your personal mental accountingis fine as long as you are happy with it :beer )
But, for an analysis of the various benefits of Roth Conversions, it should be accounted for as the alternate investment if not paying those taxes, rather than magically falling out of the sky. I have no doubt Prof. McQ will treat it appropriately in his analysis.
Yes, I hope so. The key points are that no Roth space is sacrificed, and the alternative investment is a very low-returning taxable asset. (Just can't get away from dealing with those problematic return assumptions...)
"Discipline matters more than allocation.” |—| "In finance, if you’re certain of anything, you’re out of your mind." ─William Bernstein
wrongfunds
Posts: 3187
Joined: Tue Dec 21, 2010 2:55 pm

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

Post by wrongfunds »

The reason one has "cash" could be to have the peace of mind and if you are giving that up, then should that be accounted somehow? To put it differently, if X% of your portfolio is in cash for being able to sleep well at night, you absolutely need to assign value for it when it is consumed. It is NOT the miniscule interest on that cash that I am talking about. But rather the mental value of *having* that cash. Claiming that cash as "Free" means you just fooling yourself. Where does the replenishment comes from and what is the cost of that?
marcopolo
Posts: 8445
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 »

iceport wrote: Mon Oct 18, 2021 1:57 pm
marcopolo wrote: Mon Oct 18, 2021 1:51 pm
iceport wrote: Mon Oct 18, 2021 1:35 pm
wrongfunds wrote: Mon Oct 18, 2021 12:45 pm
Chip wrote: Mon Oct 18, 2021 12:26 pm

I don't understand the concern. Many people max out their tax-deferred options and save additional money in taxable accounts. That's where all the tax payments came from for our own Roth conversions.
But the taxable money need to have the taxes paid on (aka capital gains). My comment was regarding the "outside money" was for the cash.
I don't know how typical it is to do this, but I've maintained a "sinking fund" of sorts for a long time. While I was working, that sinking fund served dual roles as both an emergency fund and as a fund source for expected large expenses in the near term, like car replacements, major home repairs, etc. Now retired, I still hold that sinking fund, and it's even grown to serve as the savings I intend to use for major home improvements. I still make small monthly contributions to this fund from my main source of retirement income, a DB pension.

You may (correctly) call it mental accounting, but I don't count these savings as part of my retirement portfolio. Some of it is held in a credit union savings account, but most of it is in a short term bond fund. Capital gains taxes are negligible, but I pay taxes on interest and dividends (also, unfortunately, fairly negligible :( ).

When I do a Roth conversion, the taxes are paid from this sinking fund. That's my source of "outside money."
You may not consider it a part of your portfolio (your personal mental accountingis fine as long as you are happy with it :beer )
But, for an analysis of the various benefits of Roth Conversions, it should be accounted for as the alternate investment if not paying those taxes, rather than magically falling out of the sky. I have no doubt Prof. McQ will treat it appropriately in his analysis.
Yes, I hope so. The key points are that no Roth space is sacrificed, and the alternative investment is a very low-returning taxable asset. (Just can't get away from dealing with those problematic return assumptions...)
Unless you are adjusting your asset mix elsewhere, you are adding risk to your portfolio by doing this (assuming you hold mostly equities in your Roth). Which may be what you intended, but that is not a fair way to assess Roth Conversions, because some of the perceived benefit woukd cone from the changing asset mix, rather than the tax arbitrage of the Roth conversion.
Once in a while you get shown the light, in the strangest of places if you look at it right.
User avatar
iceport
Posts: 6054
Joined: Sat Apr 07, 2007 4:29 pm

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

Post by iceport »

marcopolo wrote: Mon Oct 18, 2021 2:21 pm Unless you are adjusting your asset mix elsewhere, you are adding risk to your portfolio by doing this (assuming you hold mostly equities in your Roth). Which may be what you intended, but that is not a fair way to assess Roth Conversions, because some of the perceived benefit woukd cone from the changing asset mix, rather than the tax arbitrage of the Roth conversion.
Well, that's exactly what I'm doing — i.e. *not* making any other AA adjustments. So any analysis that fails to accurately model what I'm actually doing in real life seems "unfair" to me.

In fact, the effect on the overall asset mix is negligible. My Roth conversions essentially move bond funds in a rollover IRA to bond funds in the Roth. And the dollar values involved are relatively minor. Filling up the 22% tax bracket with Roth conversions only requires high 4-figure tax payments for me. That really doesn't move the overall asset mix needle in any meaningful way.

Replenishment of the sinking fund is done over time, with monthly contributions from an outside income source.
"Discipline matters more than allocation.” |—| "In finance, if you’re certain of anything, you’re out of your mind." ─William Bernstein
User avatar
iceport
Posts: 6054
Joined: Sat Apr 07, 2007 4:29 pm

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

Post by iceport »

marcopolo wrote: Mon Oct 18, 2021 2:52 pm You are keeping bond funds in your Roth account? If so, then you are right the change in asset is fairly small.
My comment assumed those funds get moved to equities when going to the Roth account. Obviously, biased biased by what i do.
Yes, we're all biased by our own circumstances!

The thing is, with <28% in TDAs, I needed to put more bonds someplace. I chose to put them in the Roth, so that I could put more tax-efficient equities in the taxable account. I've come to doubt the wisdom of that strategy. I never really saw it coming, but with the high equity returns since the spring of 2009, the taxable account has ballooned far beyond expectations. It's to the point where I'm living with more risk than I target just to avoid selling greatly appreciated equities in taxable. Although the Roth account is still mostly filled with equities, I sure wish I had some US equities in there for the past 13 years. :oops:
"Discipline matters more than allocation.” |—| "In finance, if you’re certain of anything, you’re out of your mind." ─William Bernstein
marcopolo
Posts: 8445
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 »

iceport wrote: Mon Oct 18, 2021 3:09 pm
marcopolo wrote: Mon Oct 18, 2021 2:52 pm You are keeping bond funds in your Roth account? If so, then you are right the change in asset is fairly small.
My comment assumed those funds get moved to equities when going to the Roth account. Obviously, biased biased by what i do.
Yes, we're all biased by our own circumstances!

The thing is, with <28% in TDAs, I needed to put more bonds someplace. I chose to put them in the Roth, so that I could put more tax-efficient equities in the taxable account. I've come to doubt the wisdom of that strategy. I never really saw it coming, but with the high equity returns since the spring of 2009, the taxable account has ballooned far beyond expectations. It's to the point where I'm living with more risk than I target just to avoid selling greatly appreciated equities in taxable. Although the Roth account is still mostly filled with equities, I sure wish I had some US equities in there for the past 13 years. :oops:
I feel your pain. While I am all equities in our Roth accounts, it is all in International and some SCV :oops:
Once in a while you get shown the light, in the strangest of places if you look at it right.
Topic Author
McQ
Posts: 1425
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 »

What if I paid the tax on the conversion from other funds?

*The exchange initiated by wrongfunds just upthread makes a nice segue into this post. I will try to live up to marcopolo’s confidence, but that is for all of you to judge.

This is a staple of conventional wisdom about Roth conversions: that they will be more successful if the tax is paid from outside funds rather than from the conversion. At one level this is trivially, almost tautologically true. If, in our running example, the full $100,000 is moved into a Roth account, rather than $78,000; and if the metric is, How big is my ending Roth balance; then by definition, moving $100,000 under the Roth shelter will produce a bigger Roth account value than moving only $78,000 into the Roth. Doh!

But that’s not what enthusiasts really mean. They claim, rather, that the conversion will be more lucrative, have a higher payoff, be more likely to succeed, do better by some strictly appropriate metric. They don’t mean that the Roth accumulation with tax paid outside will be 100/78 as great; they intuit something more.

My answer in the SSRN paper took the form, Nah—payment of tax outside will do somewhat better, probably about 100/78ths as well. More tax-free must be better than less tax-free; but not necessarily by much.

This post and spreadsheet will refine that intuition. The fraction “100/78” or [conversion / (conversion – tax)] proved not to be as central as I initially thought. In fact, this post will end up more supportive of conventional wisdom than many posts in this thread. The surprise, or novel contribution, lies in the explanation of just why payment of tax outside boosts expected payoffs from the conversion.

The trap

[Here I want to excuse any BH from the criticism that follows; it is aimed at the more lame among popular accounts, along with the more manipulative sales pitches that I have seen.]

Sometimes, the analysis treats the $22,000 withdrawn from taxable accounts to pay the tax on the conversion rather cavalierly.

Of course, it is convenient to assume that it has a 100% cost basis. Hence, no additional tax will be due if liquidated to pay conversion tax.

Of course, we can assume these funds are available in unlimited amounts. Everyone has as many tens of thousands or hundreds of thousands of dollars sitting in a 100% cost basis account as needed to fund tax on conversions of arbitrarily large size. That’s you, right?

Now where would such 100% cost basis funds be found? Why sitting in a bank account, of course (looking at you, iceport). It follows that the taxpayer willingly chose to leave $22,000, $44,000, $110,000, whatever, in a bank account, over some arbitrary but long interval, to make sure it was always at the ready to fund a conversion should the time come. With $110,000 sitting in the bank earning 1 or 2 basis points, you could fund tax on up to five $100,000 conversions at 22%, without incurring any additional tax obligation.

Or, if you find yourself on a desert island as in that famous economist joke: just assume you receive a timely inheritance in the needed amount before undertaking the conversion.

Next, if you hadn’t used these bank deposits to pay tax on a Roth conversion, these accounts would have just sat there forever, earning that same 1 or 2 basis points, right? Which means you can let the Roth funds double four times over the course of a 30-year retirement (1.10 to the 30th power), and then subtract the $22,000 initial tax payment to estimate the value of paying tax outside of conversion, true?

No, no, no. Wrong, wrong, wrong. That $22,000 has a negative future value. It’s gone, once paid to the IRS; but what might that future value have become if the funds were kept?

Now in passing I want to acknowledge that most affluent 60-something couples contemplating a $100,000 Roth conversion in the 22% bracket will have $22,000 in the bank (but maybe not $44,000 … $110,000). These might have been emergency funds from before age 59.5 that can be repurposed now that TDA funds, often partly invested in fixed income, can be tapped at will without penalty. There’s nothing foolish about keeping 20% of your annual income set aside in a bank account for use as emergency funds, and nothing untoward about repurposing these funds when you no longer depend on wage income.

But, for spreadsheet work, it is completely inappropriate to enter that $22,000 as an unchanging fixed cost, deducted in the same dollar amount no matter how many decades have elapsed following the conversion. Instead, the proper counterfactual is to assume that $22,000 was taken out of the bank and put in the same asset as all the other accounts (stock, in this thread thus far; but I'll vary that later). Since in fact the money was sent to the IRS, in the spreadsheet the counterfactual has a negative sign: the amount you would have had if you hadn’t sent that check to the IRS, but instead accepted a smaller conversion along with an investment of the $22,000 into stock in a taxable account.

In short: the correct formulation is whether: A) to pay tax on the conversion using the $22,000 in the bank, so that a larger stock investment could be made under the Roth umbrella; or B) to take that $22,000 out of the bank and put it into stock in a taxable account, leaving the Roth account smaller because funds were paid out of the conversion, but the taxable account larger.

Let’s be clear: a taxable account subject to non-zero tax must underperform a Roth account when both are invested in the same asset. But when the only tax drag is 15% applied to capital gains and dividends, underperformance is slow at first, so that there may only be a small contribution to conversion outcomes initially.

*In the first iteration I’ll stick to the initial mark-to-market formulation; later I will redo it the marcopolo way, with all capital gains unrealized and stepped up at death, so that there is only a 30 basis point tax drag.

In sum: when tax is paid outside the conversion, there are actually two conversions under way:
1. The conversion of TDA funds, which carries a tax obligation; and
2. payment of that tax obligation by a second “conversion,” wherein funds that could have been invested in stock in a taxable account are instead sent to the IRS so that the Roth account created by the main conversion can be exactly that much larger.
3. The payoff from this dual conversion will be the sum of the two conversion payoffs.

For example, if we set aside the main TDA conversion for the moment, and simply invested $22,000 in a Roth account and left it alone for fourteen years, age 71 to 85, we’d have $22,000 X 1.1014, or $83,545. If we instead put $22,000 into a taxable stock investment, paying in tax 15% of each year’s 10% appreciation, we’d have $22,000 X 1.08514, or $68,935. Roth would win by $83,545 - $68,935 = $14,610 (nominal).

Accordingly, relative to the base case of paying tax from conversion, there should be an additional Roth surplus of exactly $14,610. The extra payoff comes from putting an extra $22,000 to work under the Roth umbrella.

In short: doing a larger conversion produces more of the payoffs generated by any conversion. But as we’ll see in a minute, the piggyback conversion turns out to be unexpectedly powerful.

It is always a good idea to convert taxable funds into Roth funds when this can be done at 0% tax (=100% cost basis, as in a bank account). The totally tax free Roth account must outperform any account subject to any amount of tax whatsoever, ceteris paribus.

Of course, if you had accumulated gains of 50% in that taxable account, subject to tax at 15%, you’d need to liquidate $23,650 to pay tax of $22,000 plus tax on that gain. The negative future values will be higher, and the payoff from the second conversion won’t be as lucrative. Hence the claim I have made again and again, here and in the SSRN paper: paying tax from outside the conversion must boost the conversion payoff. But by how much?

More than I initially expected, under the mark to market formulation; about what I expected, if done marcopolo style (next post).

Let’s take a look at the spreadsheet. I’ll focus on the age 85 and 95 results. There are new columns on the right side. As you suss them out, note the blue formula cells for the new columns.

Image

If we look at the real Roth surplus at age 85, at first glance, it appears to have gone way up: it’s now $17,288 rather than $7,629 in the base case of 22% --> 22%, rather more than a ratio of 100/78. That had me scratching my head.

The key is to decompose that number into its two components. The first flows from the $78,000 we would have had in the Roth in the base case. The second is what we get when we invest the extra $22,000 into a Roth account paying exactly 10%, and compare that to what we would have if we took the $22,000 and invested it in a taxable account at 8.5% net of taxes. That calculation was done a few paragraphs back: the difference after fourteen years, at age 85, was $14,610 nominal. Applying the inflation divisor, that’s a real Roth surplus, on the piggyback conversion, of $9,659. And there you have it:

$ 7,629 (payoff from main conversion, removing $100,000 from the TDA)
+
$ 9,659 (payoff from the 2nd or piggyback conversion, moving $22,000 more under the Roth shelter)
=
$17,288 (total real Roth surplus)

The surprise (to me): The payoff for putting $22,000 into a Roth account, rather than investing those same dollars in stock in a taxable account, is greater than the pay off from removing $100,000 from the TDA and putting the after-tax portion in a Roth. Even though $22,000 is a fraction of $78,000.

But that’s only true at younger ages. If you look down to age 95, the incremental contribution from the $22,000 second conversion, while still substantial, and far more than 22/78, is not larger in dollars than the base case.

What’s going on?

The miniscule tax exposure of early RMDs

Here is one possible explanation for the disproportionate contribution of conversion #2: the counterfactual taxable account (the negative future value of the $22,000 mailed to the IRS, had it been invested in stocks, aka piggyback conversion) must have been taxed at a higher rate than the ensemble of the TDA counterfactual and reinvested after-tax RMDs (=main conversion). Therefore the corresponding portion of the Roth pulls ahead more rapidly.

Put another way: RMDs taken in the early 70s are tiny. Taxing a tiny RMD at 22% produces a comparatively tiny dollar tax burden ($883 at age 73). Looking at the reinvested RMDs at age 73, these are still a small amount, producing a tax burden of $43. So at age 73 we had $116,985 in the TDA, and $2847 in the reinvested RMDs; we paid tax of $883 + $43. The total tax burden is 77 basis points. Since everything is invested in stocks and returning 10%, the after-tax rate of return is 9.23%. On conversion #2, the after-tax rate is 8.5%. That’s a bigger advantage for the Roth paying an untaxed 10%; and that’s why conversion #2 contributes so much to the total surplus.

Conversely, by age 95, RMDs are no longer tiny: over 10% must be withdrawn each year; plus the reinvested RMD account is much bigger. By this point, the annual tax burden on the TDA/reinvested RMDs is over 400 basis points and climbing. The $22,000 tax payment counterfactual is still growing at the same 8.5% rate after tax; so now its contribution to the overall Roth surplus is relatively less.

Sound good? Yeah, but turns out: the formulation just given is incorrect in the letter even if true in the spirit. It was me thinking out loud before I thought better.

Remember, this is the constant rate scenario: tax at conversion is the same rate as tax on RMDs. Way up thread at the very beginning I showed that under constant rates, the tax on RMDs didn’t count, didn’t contribute to the Roth conversion payoff. The $883 tax on that year’s RMD is matched by the Roth account having to start at $78,000 not $100,000. Only the tax drag on the reinvested RMDs contributes to the payoff.

But the reinvested RMDs account starts at zero. Its value does not equal the future value of the tax payment until age 79 (compare Column L to Column P). So its contribution to tax drag is still comparatively small at an age of 85. That’s why the contribution of the piggyback conversion seems disproportionately large relative to the dollar amounts involved.

Action implications

Should you pay conversion tax from outside the conversion? Yes, on two conditions: that you have surplus taxable funds available, and that the tax burden of cashing out these investments, so that you can write a check to the IRS, is not great.

As suggested above, most couples able to contemplate a $100,000 conversion in the 22% tax bracket should have $22,000 in cash-equivalents available somewhere. And there will be some who have $110,000, which can handle tax on five conversions. But at that point, the careful converter will check to see if too much is being converted; or in any case, the window for converting at a favorable rate may have closed.

As described above, paying tax from outside is essentially the same as making a magna-backdoor Roth contribution (I needed a word smaller than mega). There aren’t many opportunities to contribute $22,000 to a Roth in one fell swoop. And over a five year period, moving $500,000 from a TDA and $110,000 out of your taxable funds is going to put $500,000 under the Roth shelter; as opposed to moving the same $500,000 out of the TDA to have $390,000 under the Roth shelter while retaining $110,000 in taxable.

The shortfall for keeping $110,000 in stocks in taxable, relative to keeping those stocks under the Roth shelter, will quickly grow quite large—if all appreciation is taxed each year. And if the asset appreciates strongly, as in the running 10% assumption.

More Roth = more shelter from tax.

But more to the point: Backdoor Roth = great! The analysis shows the power of compounding even a small sum at 8.5% versus a totally tax free account compounding at 10%.

In sum: Backdoor Roth = great & conversion tax paid outside = backdoor Roth.

That’s enough for one post. Next one will explore questions such as these:
1. What about the marcopolo way, with only dividends taxed annually, and capital gains stepped up at death, thus tax free forever, same as the Roth? Dollar payoffs will be dramatically reduced, but will the pattern change?
2. Could the piggyback conversion, from paying the tax outside, ameliorate the outcome of guessing wrong about future tax rates, and seeing rates drop? How big is the effect?
3. And if future rates go up, as generally planned/hoped when converting, does that change the contribution of paying tax outside? Of course it does; but by how much?
4. Finally: if converting early raises payoffs; and paying tax outside raises payoffs; how good does it get if we convert early and pay tax outside?
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.
marcopolo
Posts: 8445
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: Mon Oct 18, 2021 4:28 pm What if I paid the tax on the conversion from other funds?

*The exchange initiated by wrongfunds just upthread makes a nice segue into this post. I will try to live up to marcopolo’s confidence, but that is for all of you to judge.

This is a staple of conventional wisdom about Roth conversions: that they will be more successful if the tax is paid from outside funds rather than from the conversion. At one level this is trivially, almost tautologically true. If, in our running example, the full $100,000 is moved into a Roth account, rather than $78,000; and if the metric is, How big is my ending Roth balance; then by definition, moving $100,000 under the Roth shelter will produce a bigger Roth account value than moving only $78,000 into the Roth. Doh!

But that’s not what enthusiasts really mean. They claim, rather, that the conversion will be more lucrative, have a higher payoff, be more likely to succeed, do better by some strictly appropriate metric. They don’t mean that the Roth accumulation with tax paid outside will be 100/78 as great; they intuit something more.

My answer in the SSRN paper took the form, Nah—payment of tax outside will do somewhat better, probably about 100/78ths as well. More tax-free must be better than less tax-free; but not necessarily by much.

This post and spreadsheet will refine that intuition. The fraction “100/78” or [conversion / (conversion – tax)] proved not to be as central as I initially thought. In fact, this post will end up more supportive of conventional wisdom than many posts in this thread. The surprise, or novel contribution, lies in the explanation of just why payment of tax outside boosts expected payoffs from the conversion.

The trap

[Here I want to excuse any BH from the criticism that follows; it is aimed at the more lame among popular accounts, along with the more manipulative sales pitches that I have seen.]

Sometimes, the analysis treats the $22,000 withdrawn from taxable accounts to pay the tax on the conversion rather cavalierly.

Of course, it is convenient to assume that it has a 100% cost basis. Hence, no additional tax will be due if liquidated to pay conversion tax.

Of course, we can assume these funds are available in unlimited amounts. Everyone has as many tens of thousands or hundreds of thousands of dollars sitting in a 100% cost basis account as needed to fund tax on conversions of arbitrarily large size. That’s you, right?

Now where would such 100% cost basis funds be found? Why sitting in a bank account, of course (looking at you, iceport). It follows that the taxpayer willingly chose to leave $22,000, $44,000, $110,000, whatever, in a bank account, over some arbitrary but long interval, to make sure it was always at the ready to fund a conversion should the time come. With $110,000 sitting in the bank earning 1 or 2 basis points, you could fund tax on up to five $100,000 conversions at 22%, without incurring any additional tax obligation.

Or, if you find yourself on a desert island as in that famous economist joke: just assume you receive a timely inheritance in the needed amount before undertaking the conversion.

Next, if you hadn’t used these bank deposits to pay tax on a Roth conversion, these accounts would have just sat there forever, earning that same 1 or 2 basis points, right? Which means you can let the Roth funds double four times over the course of a 30-year retirement (1.10 to the 30th power), and then subtract the $22,000 initial tax payment to estimate the value of paying tax outside of conversion, true?

No, no, no. Wrong, wrong, wrong. That $22,000 has a negative future value. It’s gone, once paid to the IRS; but what might that future value have become if the funds were kept?

Now in passing I want to acknowledge that most affluent 60-something couples contemplating a $100,000 Roth conversion in the 22% bracket will have $22,000 in the bank (but maybe not $44,000 … $110,000). These might have been emergency funds from before age 59.5 that can be repurposed now that TDA funds, often partly invested in fixed income, can be tapped at will without penalty. There’s nothing foolish about keeping 20% of your annual income set aside in a bank account for use as emergency funds, and nothing untoward about repurposing these funds when you no longer depend on wage income.

But, for spreadsheet work, it is completely inappropriate to enter that $22,000 as an unchanging fixed cost, deducted in the same dollar amount no matter how many decades have elapsed following the conversion. Instead, the proper counterfactual is to assume that $22,000 was taken out of the bank and put in the same asset as all the other accounts (stock, in this thread thus far; but I'll vary that later). Since in fact the money was sent to the IRS, in the spreadsheet the counterfactual has a negative sign: the amount you would have had if you hadn’t sent that check to the IRS, but instead accepted a smaller conversion along with an investment of the $22,000 into stock in a taxable account.

In short: the correct formulation is whether: A) to pay tax on the conversion using the $22,000 in the bank, so that a larger stock investment could be made under the Roth umbrella; or B) to take that $22,000 out of the bank and put it into stock in a taxable account, leaving the Roth account smaller because funds were paid out of the conversion, but the taxable account larger.

Let’s be clear: a taxable account subject to non-zero tax must underperform a Roth account when both are invested in the same asset. But when the only tax drag is 15% applied to capital gains and dividends, underperformance is slow at first, so that there may only be a small contribution to conversion outcomes initially.

*In the first iteration I’ll stick to the initial mark-to-market formulation; later I will redo it the marcopolo way, with all capital gains unrealized and stepped up at death, so that there is only a 30 basis point tax drag.

In sum: when tax is paid outside the conversion, there are actually two conversions under way:
1. The conversion of TDA funds, which carries a tax obligation; and
2. payment of that tax obligation by a second “conversion,” wherein funds that could have been invested in stock in a taxable account are instead sent to the IRS so that the Roth account created by the main conversion can be exactly that much larger.
3. The payoff from this dual conversion will be the sum of the two conversion payoffs.

For example, if we set aside the main TDA conversion for the moment, and simply invested $22,000 in a Roth account and left it alone for fourteen years, age 71 to 85, we’d have $22,000 X 1.1014, or $83,545. If we instead put $22,000 into a taxable stock investment, paying in tax 15% of each year’s 10% appreciation, we’d have $22,000 X 1.08514, or $68,935. Roth would win by $83,545 - $68,935 = $14,610 (nominal).

Accordingly, relative to the base case of paying tax from conversion, there should be an additional Roth surplus of exactly $14,610. The extra payoff comes from putting an extra $22,000 to work under the Roth umbrella.

In short: doing a larger conversion produces more of the payoffs generated by any conversion. But as we’ll see in a minute, the piggyback conversion turns out to be unexpectedly powerful.

It is always a good idea to convert taxable funds into Roth funds when this can be done at 0% tax (=100% cost basis, as in a bank account). The totally tax free Roth account must outperform any account subject to any amount of tax whatsoever, ceteris paribus.

Of course, if you had accumulated gains of 50% in that taxable account, subject to tax at 15%, you’d need to liquidate $23,650 to pay tax of $22,000 plus tax on that gain. The negative future values will be higher, and the payoff from the second conversion won’t be as lucrative. Hence the claim I have made again and again, here and in the SSRN paper: paying tax from outside the conversion must boost the conversion payoff. But by how much?

More than I initially expected, under the mark to market formulation; about what I expected, if done marcopolo style (next post).

Let’s take a look at the spreadsheet. I’ll focus on the age 85 and 95 results. There are new columns on the right side. As you suss them out, note the blue formula cells for the new columns.

Image

If we look at the real Roth surplus at age 85, at first glance, it appears to have gone way up: it’s now $17,288 rather than $7,629 in the base case of 22% --> 22%, rather more than a ratio of 100/78. That had me scratching my head.

The key is to decompose that number into its two components. The first flows from the $78,000 we would have had in the Roth in the base case. The second is what we get when we invest the extra $22,000 into a Roth account paying exactly 10%, and compare that to what we would have if we took the $22,000 and invested it in a taxable account at 8.5% net of taxes. That calculation was done a few paragraphs back: the difference after fourteen years, at age 85, was $14,610 nominal. Applying the inflation divisor, that’s a real Roth surplus, on the piggyback conversion, of $9,659. And there you have it:

$ 7,629 (payoff from main conversion, removing $100,000 from the TDA)
+
$ 9,659 (payoff from the 2nd or piggyback conversion, moving $22,000 more under the Roth shelter)
=
$17,288 (total real Roth surplus)

The surprise (to me): The payoff for putting $22,000 into a Roth account, rather than investing those same dollars in stock in a taxable account, is greater than the pay off from removing $100,000 from the TDA and putting the after-tax portion in a Roth. Even though $22,000 is a fraction of $78,000.

But that’s only true at younger ages. If you look down to age 95, the incremental contribution from the $22,000 second conversion, while still substantial, and far more than 22/78, is not larger in dollars than the base case.

What’s going on?

The miniscule tax exposure of early RMDs

Here is one possible explanation for the disproportionate contribution of conversion #2: the counterfactual taxable account (the negative future value of the $22,000 mailed to the IRS, had it been invested in stocks, aka piggyback conversion) must have been taxed at a higher rate than the ensemble of the TDA counterfactual and reinvested after-tax RMDs (=main conversion). Therefore the corresponding portion of the Roth pulls ahead more rapidly.

Put another way: RMDs taken in the early 70s are tiny. Taxing a tiny RMD at 22% produces a comparatively tiny dollar tax burden ($883 at age 73). Looking at the reinvested RMDs at age 73, these are still a small amount, producing a tax burden of $43. So at age 73 we had $116,985 in the TDA, and $2847 in the reinvested RMDs; we paid tax of $883 + $43. The total tax burden is 77 basis points. Since everything is invested in stocks and returning 10%, the after-tax rate of return is 9.23%. On conversion #2, the after-tax rate is 8.5%. That’s a bigger advantage for the Roth paying an untaxed 10%; and that’s why conversion #2 contributes so much to the total surplus.

Conversely, by age 95, RMDs are no longer tiny: over 10% must be withdrawn each year; plus the reinvested RMD account is much bigger. By this point, the annual tax burden on the TDA/reinvested RMDs is over 400 basis points and climbing. The $22,000 tax payment counterfactual is still growing at the same 8.5% rate after tax; so now its contribution to the overall Roth surplus is relatively less.

Sound good? Yeah, but turns out: the formulation just given is incorrect in the letter even if true in the spirit. It was me thinking out loud before I thought better.

Remember, this is the constant rate scenario: tax at conversion is the same rate as tax on RMDs. Way up thread at the very beginning I showed that under constant rates, the tax on RMDs didn’t count, didn’t contribute to the Roth conversion payoff. The $883 tax on that year’s RMD is matched by the Roth account having to start at $78,000 not $100,000. Only the tax drag on the reinvested RMDs contributes to the payoff.

But the reinvested RMDs account starts at zero. Its value does not equal the future value of the tax payment until age 79 (compare Column L to Column P). So its contribution to tax drag is still comparatively small at an age of 85. That’s why the contribution of the piggyback conversion seems disproportionately large relative to the dollar amounts involved.

Action implications

Should you pay conversion tax from outside the conversion? Yes, on two conditions: that you have surplus taxable funds available, and that the tax burden of cashing out these investments, so that you can write a check to the IRS, is not great.

As suggested above, most couples able to contemplate a $100,000 conversion in the 22% tax bracket should have $22,000 in cash-equivalents available somewhere. And there will be some who have $110,000, which can handle tax on five conversions. But at that point, the careful converter will check to see if too much is being converted; or in any case, the window for converting at a favorable rate may have closed.

As described above, paying tax from outside is essentially the same as making a magna-backdoor Roth contribution (I needed a word smaller than mega). There aren’t many opportunities to contribute $22,000 to a Roth in one fell swoop. And over a five year period, moving $500,000 from a TDA and $110,000 out of your taxable funds is going to put $500,000 under the Roth shelter; as opposed to moving the same $500,000 out of the TDA to have $390,000 under the Roth shelter while retaining $110,000 in taxable.

The shortfall for keeping $110,000 in stocks in taxable, relative to keeping those stocks under the Roth shelter, will quickly grow quite large—if all appreciation is taxed each year. And if the asset appreciates strongly, as in the running 10% assumption.

More Roth = more shelter from tax.

But more to the point: Backdoor Roth = great! The analysis shows the power of compounding even a small sum at 8.5% versus a totally tax free account compounding at 10%.

In sum: Backdoor Roth = great & conversion tax paid outside = backdoor Roth.

That’s enough for one post. Next one will explore questions such as these:
1. What about the marcopolo way, with only dividends taxed annually, and capital gains stepped up at death, thus tax free forever, same as the Roth? Dollar payoffs will be dramatically reduced, but will the pattern change?
2. Could the piggyback conversion, from paying the tax outside, ameliorate the outcome of guessing wrong about future tax rates, and seeing rates drop? How big is the effect?
3. And if future rates go up, as generally planned/hoped when converting, does that change the contribution of paying tax outside? Of course it does; but by how much?
4. Finally: if converting early raises payoffs; and paying tax outside raises payoffs; how good does it get if we convert early and pay tax outside?
Thank you for getting us back on topic!
I will try to digest this over the next day or two.
Once in a while you get shown the light, in the strangest of places if you look at it right.
User avatar
iceport
Posts: 6054
Joined: Sat Apr 07, 2007 4:29 pm

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

Post by iceport »

McQ wrote: Mon Oct 18, 2021 4:28 pm Now where would such 100% cost basis funds be found? Why sitting in a bank account, of course (looking at you, iceport). It follows that the taxpayer willingly chose to leave $22,000, $44,000, $110,000, whatever, in a bank account, over some arbitrary but long interval, to make sure it was always at the ready to fund a conversion should the time come. With $110,000 sitting in the bank earning 1 or 2 basis points, you could fund tax on up to five $100,000 conversions at 22%, without incurring any additional tax obligation.

Or, if you find yourself on a desert island as in that famous economist joke: just assume you receive a timely inheritance in the needed amount before undertaking the conversion.

Next, if you hadn’t used these bank deposits to pay tax on a Roth conversion, these accounts would have just sat there forever, earning that same 1 or 2 basis points, right? Which means you can let the Roth funds double four times over the course of a 30-year retirement (1.10 to the 30th power), and then subtract the $22,000 initial tax payment to estimate the value of paying tax outside of conversion, true?

No, no, no. Wrong, wrong, wrong. That $22,000 has a negative future value. It’s gone, once paid to the IRS; but what might that future value have become if the funds were kept?
Heh, heh... thanks for giving me a good chuckle, McQ!

Seriously, though, the more I think about this analysis, the more convinced I become that generalizations or assumptions about what are hoped to be common or typical circumstances might fall short for many of us. In my case, a mere $40k conversion fills the 22% (single) tax bracket. The tax on that is $8,800 federal.

I've come to appreciate the practical financial benefits as well as the emotional benefits of maintaining a sinking fund, or even a truly purposeless "slush fund" of cash or reasonably cash-like assets. While there are limits to how much money it might make sense to hold in reserve like that, I expect that I would always have enough of a "cash" buffer to absorb an $8,800 hit. It might not be able to fund a long series of conversions condensed to a single lump sum, but over time the fund is replenished, in an ongoing basis, with outside income. There's no reason to assume years of conversion taxes are sitting idle (with tiny interest rates) in a big lump sum for years on end, nor that it is alternatively invested in equities.

So at first glance, I don't think assuming a significant opportunity cost for paying the conversion tax from my slush fund is accurate for my circumstances. But I'll need to think through your explanations more carefully.
"Discipline matters more than allocation.” |—| "In finance, if you’re certain of anything, you’re out of your mind." ─William Bernstein
marcopolo
Posts: 8445
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 »

iceport wrote: Mon Oct 18, 2021 6:38 pm
McQ wrote: Mon Oct 18, 2021 4:28 pm Now where would such 100% cost basis funds be found? Why sitting in a bank account, of course (looking at you, iceport). It follows that the taxpayer willingly chose to leave $22,000, $44,000, $110,000, whatever, in a bank account, over some arbitrary but long interval, to make sure it was always at the ready to fund a conversion should the time come. With $110,000 sitting in the bank earning 1 or 2 basis points, you could fund tax on up to five $100,000 conversions at 22%, without incurring any additional tax obligation.

Or, if you find yourself on a desert island as in that famous economist joke: just assume you receive a timely inheritance in the needed amount before undertaking the conversion.

Next, if you hadn’t used these bank deposits to pay tax on a Roth conversion, these accounts would have just sat there forever, earning that same 1 or 2 basis points, right? Which means you can let the Roth funds double four times over the course of a 30-year retirement (1.10 to the 30th power), and then subtract the $22,000 initial tax payment to estimate the value of paying tax outside of conversion, true?

No, no, no. Wrong, wrong, wrong. That $22,000 has a negative future value. It’s gone, once paid to the IRS; but what might that future value have become if the funds were kept?
Heh, heh... thanks for giving me a good chuckle, McQ!

Seriously, though, the more I think about this analysis, the more convinced I become that generalizations or assumptions about what are hoped to be common or typical circumstances might fall short for many of us. In my case, a mere $40k conversion fills the 22% (single) tax bracket. The tax on that is $8,800 federal.

I've come to appreciate the practical financial benefits as well as the emotional benefits of maintaining a sinking fund, or even a truly purposeless "slush fund" of cash or reasonably cash-like assets. While there are limits to how much money it might make sense to hold in reserve like that, I expect that I would always have enough of a "cash" buffer to absorb an $8,800 hit. It might not be able to fund a long series of conversions condensed to a single lump sum, but over time the fund is replenished, in an ongoing basis, with outside income. There's no reason to assume years of conversion taxes are sitting idle (with tiny interest rates) in a big lump sum for years on end, nor that it is alternatively invested in equities.

So at first glance, I don't think assuming a significant opportunity cost for paying the conversion tax from my slush fund is accurate for my circumstances. But I'll need to think through your explanations more carefully.
I think it probably depends on what you would do with that money if you didn't have to replenish your slush fund. Would you still keep adding to it, letting it get ever larger, or would you invest some of It? If you would invest some of it once the slush fund got to a certain size, then I think anything that comes out of that slush fund then has that opportunity cost attached to it.
Once in a while you get shown the light, in the strangest of places if you look at it right.
Topic Author
McQ
Posts: 1425
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: Sun Oct 17, 2021 9:24 pm
Once McQ has a link to his spreadsheet that doesn't cause a security warning, the strategy outlined in this post ("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") will likely be implemented. Having two spreadsheet references (McQ's and MDM's) seems worthwhile.
Hello FiveK: are you able to securely access the file at this link?
https://github.com/emcquarrie/Roth-conv ... sheet.xlsx
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.
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 »

^^^
The file download worked for me.
User avatar
FiveK
Posts: 15742
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: Mon Oct 18, 2021 10:27 pm Hello FiveK: are you able to securely access the file at this link?
https://github.com/emcquarrie/Roth-conv ... sheet.xlsx
Yes. It's not as straightforward as a direct link, but it does work - thanks!

The one there now is not the most recent, correct?
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: Mon Oct 18, 2021 4:28 pm

Action implications

Should you pay conversion tax from outside the conversion? Yes, on two conditions: that you have surplus taxable funds available, and that the tax burden of cashing out these investments, so that you can write a check to the IRS, is not great.

As suggested above, most couples able to contemplate a $100,000 conversion in the 22% tax bracket should have $22,000 in cash-equivalents available somewhere. And there will be some who have $110,000, which can handle tax on five conversions. But at that point, the careful converter will check to see if too much is being converted; or in any case, the window for converting at a favorable rate may have closed.

As described above, paying tax from outside is essentially the same as making a magna-backdoor Roth contribution (I needed a word smaller than mega). There aren’t many opportunities to contribute $22,000 to a Roth in one fell swoop. And over a five year period, moving $500,000 from a TDA and $110,000 out of your taxable funds is going to put $500,000 under the Roth shelter; as opposed to moving the same $500,000 out of the TDA to have $390,000 under the Roth shelter while retaining $110,000 in taxable.
................

In sum: Backdoor Roth = great & conversion tax paid outside = backdoor Roth.

That’s enough for one post. Next one will explore questions such as these:
1. What about the marcopolo way, with only dividends taxed annually, and capital gains stepped up at death, thus tax free forever, same as the Roth? Dollar payoffs will be dramatically reduced, but will the pattern change?
An interesting exercise here. I've tended to view paying the tax on Roth conversions from outside funds as a "nice to have" type of matter. As you point out, it's effectively a way of moving the tax amount from taxable to Roth. It's possible to overstate the tax drag, however; if the retiree has moved enough of their accounts to Roth, they may be able to stay in the 0% tax bracket for qualified dividends.

It's not unusual for someone coming up towards early retirement to start stashing a bit more in cash equivalents on the taxable side of their investments even though that might not on the surface be the most tax-efficient location. That can be part of what they plan to use for living expenses prior to claiming SS. Of course, what's used for one purpose won't be available for the other, so some level of tax cost or foregone return is normally implied. For myself, I managed to raise a large pile of cash on the taxable side the year I retired by downsizing our home, but that is very much a corner case.

I think there is room for some interesting analysis of the effects of somewhat higher inflation in the scenarios as well. This would be in the case where you are actually spending down the accounts to some extent. In taxable accounts, a few years of 5%+ inflation can really hurt as you are effectively paying tax just to get your money back.
smitcat
Posts: 13304
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 »

curmudgeon wrote: Tue Oct 19, 2021 12:01 am
McQ wrote: Mon Oct 18, 2021 4:28 pm

Action implications

Should you pay conversion tax from outside the conversion? Yes, on two conditions: that you have surplus taxable funds available, and that the tax burden of cashing out these investments, so that you can write a check to the IRS, is not great.

As suggested above, most couples able to contemplate a $100,000 conversion in the 22% tax bracket should have $22,000 in cash-equivalents available somewhere. And there will be some who have $110,000, which can handle tax on five conversions. But at that point, the careful converter will check to see if too much is being converted; or in any case, the window for converting at a favorable rate may have closed.

As described above, paying tax from outside is essentially the same as making a magna-backdoor Roth contribution (I needed a word smaller than mega). There aren’t many opportunities to contribute $22,000 to a Roth in one fell swoop. And over a five year period, moving $500,000 from a TDA and $110,000 out of your taxable funds is going to put $500,000 under the Roth shelter; as opposed to moving the same $500,000 out of the TDA to have $390,000 under the Roth shelter while retaining $110,000 in taxable.
................

In sum: Backdoor Roth = great & conversion tax paid outside = backdoor Roth.

That’s enough for one post. Next one will explore questions such as these:
1. What about the marcopolo way, with only dividends taxed annually, and capital gains stepped up at death, thus tax free forever, same as the Roth? Dollar payoffs will be dramatically reduced, but will the pattern change?
An interesting exercise here. I've tended to view paying the tax on Roth conversions from outside funds as a "nice to have" type of matter. As you point out, it's effectively a way of moving the tax amount from taxable to Roth. It's possible to overstate the tax drag, however; if the retiree has moved enough of their accounts to Roth, they may be able to stay in the 0% tax bracket for qualified dividends.

It's not unusual for someone coming up towards early retirement to start stashing a bit more in cash equivalents on the taxable side of their investments even though that might not on the surface be the most tax-efficient location. That can be part of what they plan to use for living expenses prior to claiming SS. Of course, what's used for one purpose won't be available for the other, so some level of tax cost or foregone return is normally implied. For myself, I managed to raise a large pile of cash on the taxable side the year I retired by downsizing our home, but that is very much a corner case.

I think there is room for some interesting analysis of the effects of somewhat higher inflation in the scenarios as well. This would be in the case where you are actually spending down the accounts to some extent. In taxable accounts, a few years of 5%+ inflation can really hurt as you are effectively paying tax just to get your money back.
"For myself, I managed to raise a large pile of cash on the taxable side the year I retired by downsizing our home, but that is very much a corner case."
FWIW - I believe that this corner case may be much more common then you believe for folks in these categories. Here are just a few ways that someone coming up on early retirement may have 'newer' cash available...
- sold primary home
- sold primary home, owns another
- inheritance
- seperation bonus/payments
- sold business
- insurance
User avatar
iceport
Posts: 6054
Joined: Sat Apr 07, 2007 4:29 pm

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

Post by iceport »

marcopolo wrote: Mon Oct 18, 2021 7:08 pm
iceport wrote: Mon Oct 18, 2021 6:38 pm
McQ wrote: Mon Oct 18, 2021 4:28 pm Now where would such 100% cost basis funds be found? Why sitting in a bank account, of course (looking at you, iceport). It follows that the taxpayer willingly chose to leave $22,000, $44,000, $110,000, whatever, in a bank account, over some arbitrary but long interval, to make sure it was always at the ready to fund a conversion should the time come. With $110,000 sitting in the bank earning 1 or 2 basis points, you could fund tax on up to five $100,000 conversions at 22%, without incurring any additional tax obligation.

Or, if you find yourself on a desert island as in that famous economist joke: just assume you receive a timely inheritance in the needed amount before undertaking the conversion.

Next, if you hadn’t used these bank deposits to pay tax on a Roth conversion, these accounts would have just sat there forever, earning that same 1 or 2 basis points, right? Which means you can let the Roth funds double four times over the course of a 30-year retirement (1.10 to the 30th power), and then subtract the $22,000 initial tax payment to estimate the value of paying tax outside of conversion, true?

No, no, no. Wrong, wrong, wrong. That $22,000 has a negative future value. It’s gone, once paid to the IRS; but what might that future value have become if the funds were kept?
Heh, heh... thanks for giving me a good chuckle, McQ!

Seriously, though, the more I think about this analysis, the more convinced I become that generalizations or assumptions about what are hoped to be common or typical circumstances might fall short for many of us. In my case, a mere $40k conversion fills the 22% (single) tax bracket. The tax on that is $8,800 federal.

I've come to appreciate the practical financial benefits as well as the emotional benefits of maintaining a sinking fund, or even a truly purposeless "slush fund" of cash or reasonably cash-like assets. While there are limits to how much money it might make sense to hold in reserve like that, I expect that I would always have enough of a "cash" buffer to absorb an $8,800 hit. It might not be able to fund a long series of conversions condensed to a single lump sum, but over time the fund is replenished, in an ongoing basis, with outside income. There's no reason to assume years of conversion taxes are sitting idle (with tiny interest rates) in a big lump sum for years on end, nor that it is alternatively invested in equities.

So at first glance, I don't think assuming a significant opportunity cost for paying the conversion tax from my slush fund is accurate for my circumstances. But I'll need to think through your explanations more carefully.
I think it probably depends on what you would do with that money if you didn't have to replenish your slush fund. Would you still keep adding to it, letting it get ever larger, or would you invest some of It? If you would invest some of it once the slush fund got to a certain size, then I think anything that comes out of that slush fund then has that opportunity cost attached to it.
This is another insightful comment, marcopolo. You are correct, the theoretical lines between the slush fund and the "portfolio" would get very blurry over time. Money is fungible, and there are limits to mental accounting. So now that you've forced me to drop the convenience (or really just emotional comfort) of mental accounting, I can consider the more accurate, more realistic model for my real life behavior. (Ironically, it's actually easier or more convenient to drop the mental accounting.)

So what does real life look like for me? A formal, dedicated allocation to cash. If all funds are treated as part of the whole, there will always be an allocation to cash.

So to the extent that the cash position is replenished from portfolio re-balancing, there would correctly be an opportunity cost associated with using the cash to pay the Roth conversion tax.

However, this misses the current status of my portfolio. Currently, my taxable equities are sitting on mountains of unrealized capital gains, and I would never (well, up to a point...) liquidate them just to re-balance. Beyond the capital gains issue, the taxable equities are all well over their target allocations. So in real life, the correct answer is that, yes, I would just keep adding to the cash position, letting it get larger and larger. I already own too much in equities. At some point, too much cash will get invested in bonds — but *not* more equities. (Honestly, I think this could be an accurate assumption for lots of people!)

Therefor, in my current circumstances, the correct assumption for the opportunity cost of paying the Roth conversion tax from the taxable account is to use the expected return of short- to mid-term bonds, *not* equity returns.

This thought process really drives home to me how extremely variable an accurate analysis is to individual circumstances — both in general *and* at specific points in time.

If I've said it before, I'll reiterate: I don't think attempting to formulate generalized Roth conversion recommendations is possible. It all comes down to the specific details of individual investors' lives. Those specific details have the potential to completely upend the analyses of the (theoretically) "typical" circumstances.
"Discipline matters more than allocation.” |—| "In finance, if you’re certain of anything, you’re out of your mind." ─William Bernstein
marcopolo
Posts: 8445
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 »

iceport wrote: Tue Oct 19, 2021 1:57 pm
marcopolo wrote: Mon Oct 18, 2021 7:08 pm
iceport wrote: Mon Oct 18, 2021 6:38 pm
McQ wrote: Mon Oct 18, 2021 4:28 pm Now where would such 100% cost basis funds be found? Why sitting in a bank account, of course (looking at you, iceport). It follows that the taxpayer willingly chose to leave $22,000, $44,000, $110,000, whatever, in a bank account, over some arbitrary but long interval, to make sure it was always at the ready to fund a conversion should the time come. With $110,000 sitting in the bank earning 1 or 2 basis points, you could fund tax on up to five $100,000 conversions at 22%, without incurring any additional tax obligation.

Or, if you find yourself on a desert island as in that famous economist joke: just assume you receive a timely inheritance in the needed amount before undertaking the conversion.

Next, if you hadn’t used these bank deposits to pay tax on a Roth conversion, these accounts would have just sat there forever, earning that same 1 or 2 basis points, right? Which means you can let the Roth funds double four times over the course of a 30-year retirement (1.10 to the 30th power), and then subtract the $22,000 initial tax payment to estimate the value of paying tax outside of conversion, true?

No, no, no. Wrong, wrong, wrong. That $22,000 has a negative future value. It’s gone, once paid to the IRS; but what might that future value have become if the funds were kept?
Heh, heh... thanks for giving me a good chuckle, McQ!

Seriously, though, the more I think about this analysis, the more convinced I become that generalizations or assumptions about what are hoped to be common or typical circumstances might fall short for many of us. In my case, a mere $40k conversion fills the 22% (single) tax bracket. The tax on that is $8,800 federal.

I've come to appreciate the practical financial benefits as well as the emotional benefits of maintaining a sinking fund, or even a truly purposeless "slush fund" of cash or reasonably cash-like assets. While there are limits to how much money it might make sense to hold in reserve like that, I expect that I would always have enough of a "cash" buffer to absorb an $8,800 hit. It might not be able to fund a long series of conversions condensed to a single lump sum, but over time the fund is replenished, in an ongoing basis, with outside income. There's no reason to assume years of conversion taxes are sitting idle (with tiny interest rates) in a big lump sum for years on end, nor that it is alternatively invested in equities.

So at first glance, I don't think assuming a significant opportunity cost for paying the conversion tax from my slush fund is accurate for my circumstances. But I'll need to think through your explanations more carefully.
I think it probably depends on what you would do with that money if you didn't have to replenish your slush fund. Would you still keep adding to it, letting it get ever larger, or would you invest some of It? If you would invest some of it once the slush fund got to a certain size, then I think anything that comes out of that slush fund then has that opportunity cost attached to it.
This is another insightful comment, marcopolo. You are correct, the theoretical lines between the slush fund and the "portfolio" would get very blurry over time. Money is fungible, and there are limits to mental accounting. So now that you've forced me to drop the convenience (or really just emotional comfort) of mental accounting, I can consider the more accurate, more realistic model for my real life behavior. (Ironically, it's actually easier or more convenient to drop the mental accounting.)

So what does real life look like for me? A formal, dedicated allocation to cash. If all funds are treated as part of the whole, there will always be an allocation to cash.

So to the extent that the cash position is replenished from portfolio re-balancing, there would correctly be an opportunity cost associated with using the cash to pay the Roth conversion tax.

However, this misses the current status of my portfolio. Currently, my taxable equities are sitting on mountains of unrealized capital gains, and I would never (well, up to a point...) liquidate them just to re-balance. Beyond the capital gains issue, the taxable equities are all well over their target allocations. So in real life, the correct answer is that, yes, I would just keep adding to the cash position, letting it get larger and larger. I already own too much in equities. At some point, too much cash will get invested in bonds — but *not* more equities. (Honestly, I think this could be an accurate assumption for lots of people!)

Therefor, in my current circumstances, the correct assumption for the opportunity cost of paying the Roth conversion tax from the taxable account is to use the expected return of short- to mid-term bonds, *not* equity returns.

This thought process really drives home to me how extremely variable an accurate analysis is to individual circumstances — both in general *and* at specific points in time.

If I've said it before, I'll reiterate: I don't think attempting to formulate generalized Roth conversion recommendations is possible. It all comes down to the specific details of individual investors' lives. Those specific details have the potential to completely upend the analyses of the (theoretically) "typical" circumstances.
Someone upthread asked if there was a simple summary for this thread. I think this is probably it.
Once in a while you get shown the light, in the strangest of places if you look at it right.
HomeStretch
Posts: 11419
Joined: Thu Dec 27, 2018 2:06 pm

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

Post by HomeStretch »

iceport wrote: Tue Oct 19, 2021 1:57 pm … I don't think attempting to formulate generalized Roth conversion recommendations is possible. It all comes down to the specific details of individual investors' lives. Those specific details have the potential to completely upend the analyses of the (theoretically) "typical" circumstances.
I tend to agree with this statement.

McQ, the iterations of your spreadsheet posted on this thread likely take significant time. But, is it all feasible to:
1) create say five profiles of the most common investors contemplating Roth conversions,
2) add an initial “interview” with a set of questions intended to categorize an investor into one of those profiles based on their responses, and
3) point the investor to one of five spreadsheets, each one customized to one of the profiles?

Perhaps five profiles are too many/too few, if so pick a number. But semi-customized spreadsheets might be sufficient to give an investor a better tool for their circumstances to evaluate Roth conversions.
Topic Author
McQ
Posts: 1425
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: Mon Oct 18, 2021 11:14 pm
McQ wrote: Mon Oct 18, 2021 10:27 pm Hello FiveK: are you able to securely access the file at this link?
https://github.com/emcquarrie/Roth-conv ... sheet.xlsx
Yes. It's not as straightforward as a direct link, but it does work - thanks!

The one there now is not the most recent, correct?
Glad it worked. When I revisited Google drive to make another attempt, the support thread reminded me that "if you have your account through a school or university, public linking may be disabled." So, github for now.
The ss at github and on my website is the same and it is the basic spreadsheet. It allows the user to vary the tax rate at conversion and at RMDs, and the return rates for the accounts. It does not have the permutations investigated over the last week or two (pay from taxable, convert early, track cost basis per marcopolo). I probably won't update the basic spreadsheet until all the permutations are out, still a month or two to go. But any updates will be put in the same repositories.
Up to you whether to hold off linking for now, or to link with some caveat about "basic, see thread for updates as they appear"
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: 1425
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 »

iceport wrote: Tue Oct 19, 2021 1:57 pm ...
Therefore in my current circumstances, the correct assumption for the opportunity cost of paying the Roth conversion tax from the taxable account is to use the expected return of short- to mid-term bonds, *not* equity returns.

This thought process really drives home to me how extremely variable an accurate analysis is to individual circumstances — both in general *and* at specific points in time.

If I've said it before, I'll reiterate: I don't think attempting to formulate generalized Roth conversion recommendations is possible. It all comes down to the specific details of individual investors' lives. Those specific details have the potential to completely upend the analyses of the (theoretically) "typical" circumstances.
That is a defensible position, iceport, and you may be right. But that would leave me as an author and scholar, who must be concerned with the general case, with nothing to do. Conversely, the very difficulty of coming up with anything general in the Roth conversion space makes the project challenging and interesting, even if doomed to failure. Put another way, if it was a simple problem, I wouldn't be interested in it. But because it is not simple, my efforts may come to naught.
Scarcely matters, since I do what I want now that I'm retired, and this is one of the problems I find interesting at the moment.
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: 1425
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 »

HomeStretch wrote: Tue Oct 19, 2021 3:17 pm
iceport wrote: Tue Oct 19, 2021 1:57 pm … I don't think attempting to formulate generalized Roth conversion recommendations is possible. It all comes down to the specific details of individual investors' lives. Those specific details have the potential to completely upend the analyses of the (theoretically) "typical" circumstances.
I tend to agree with this statement.

McQ, the iterations of your spreadsheet posted on this thread likely take significant time. But, is it all feasible to:
1) create say five profiles of the most common investors contemplating Roth conversions,
2) add an initial “interview” with a set of questions intended to categorize an investor into one of those profiles based on their responses, and
3) point the investor to one of five spreadsheets, each one customized to one of the profiles?

Perhaps five profiles are too many/too few, if so pick a number. But semi-customized spreadsheets might be sufficient to give an investor a better tool for their circumstances to evaluate Roth conversions.
Homestretch, that is an excellent suggestion that I will have to consider down the road. The SSRN paper made a stab in that direction, but only had a single profile, one found questionable by too many BH. And of course that's one potential problem I see with the suggestion: whatever number I picked, 3, 5, or 11, would soon get the same response: "You left out the case of X and I wonder if you could add that scenario, which is of personal interest to me." Nor would any manageable set really address iceport's observation.

The second problem: I am probably not the right person to assemble that eminently practical and useful set of scenarios. It seems more the kind of thing done on the wiki or by individuals who are professionally employed as financial planners (which I am not). My strengths lie elsewhere. I don't advise individuals or create software tools. I write papers about data manipulations, along with my conceptualization of the underlying problem the data are designed to address, that I hope some readers will find interesting.

Nonetheless, I repeat: that's an excellent suggestion and food for thought. If I follow up on it down the road, I'll be acknowledging you by user handle. That may happen: if five examples serve my expository aim at some future point, I'll be all over it and asking for feedback on the scenarios initially selected. :sharebeer
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.
marcopolo
Posts: 8445
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: Tue Oct 19, 2021 6:09 pm
iceport wrote: Tue Oct 19, 2021 1:57 pm ...
Therefore in my current circumstances, the correct assumption for the opportunity cost of paying the Roth conversion tax from the taxable account is to use the expected return of short- to mid-term bonds, *not* equity returns.

This thought process really drives home to me how extremely variable an accurate analysis is to individual circumstances — both in general *and* at specific points in time.

If I've said it before, I'll reiterate: I don't think attempting to formulate generalized Roth conversion recommendations is possible. It all comes down to the specific details of individual investors' lives. Those specific details have the potential to completely upend the analyses of the (theoretically) "typical" circumstances.
That is a defensible position, iceport, and you may be right. But that would leave me as an author and scholar, who must be concerned with the general case, with nothing to do. Conversely, the very difficulty of coming up with anything general in the Roth conversion space makes the project challenging and interesting, even if doomed to failure. Put another way, if it was a simple problem, I wouldn't be interested in it. But because it is not simple, my efforts may come to naught.
Scarcely matters, since I do what I want now that I'm retired, and this is one of the problems I find interesting at the moment.
I think both of these viewpoints can, and probably are true.
It is a very complex problem, which makes it interesting, but also means there probably is not a simple genaralizable answer.

But, what i think can be done is to create a frame work with enough flexibility that individuals can take the output of this effort, modify inputs and parameters to fit their situation (mix of tax rates, asset allocation, location, etc.) and then evaluate various output scenarios (spend down, surviving spouse, heirs, etc.). It is indeed a tall order. But, if it were easy, it would have been done already.

I am a bit skeptical if all of that can be done in a spreadsheet without introducing more complex macros, but we shall see where this goes.
Once in a while you get shown the light, in the strangest of places if you look at it right.
User avatar
iceport
Posts: 6054
Joined: Sat Apr 07, 2007 4:29 pm

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

Post by iceport »

McQ wrote: Tue Oct 19, 2021 6:09 pm
iceport wrote: Tue Oct 19, 2021 1:57 pm ...
Therefore in my current circumstances, the correct assumption for the opportunity cost of paying the Roth conversion tax from the taxable account is to use the expected return of short- to mid-term bonds, *not* equity returns.

This thought process really drives home to me how extremely variable an accurate analysis is to individual circumstances — both in general *and* at specific points in time.

If I've said it before, I'll reiterate: I don't think attempting to formulate generalized Roth conversion recommendations is possible. It all comes down to the specific details of individual investors' lives. Those specific details have the potential to completely upend the analyses of the (theoretically) "typical" circumstances.
That is a defensible position, iceport, and you may be right. But that would leave me as an author and scholar, who must be concerned with the general case, with nothing to do. Conversely, the very difficulty of coming up with anything general in the Roth conversion space makes the project challenging and interesting, even if doomed to failure. Put another way, if it was a simple problem, I wouldn't be interested in it. But because it is not simple, my efforts may come to naught.
Scarcely matters, since I do what I want now that I'm retired, and this is one of the problems I find interesting at the moment.
I wasn't trying to imply that your efforts are bound to fail or are of limited value. I think the more constructive interpretation is that the analyses need to build in enough flexibility to handle a very wide range of circumstances. I've just been dutifully serving as a Devil's advocate, to demonstrate how different real life can be from typical assumptions. :twisted: It's really just an extension of the flexibility needed to try out the effects of various tax risks — which I really consider to be the biggest source of uncertainty in the whole endeavor.

At times I struggle to follow all the intricate details of your work, but I do find it interesting!
"Discipline matters more than allocation.” |—| "In finance, if you’re certain of anything, you’re out of your mind." ─William Bernstein
User avatar
sapphire96
Posts: 193
Joined: Fri Jun 16, 2017 8:08 pm

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

Post by sapphire96 »

Very intriguing thread - thank you for the effort you have put into this.

Sorry if I missed this, but did you perform an analysis on if one of the spouses dies (say, at age 80) and the surviving spouse is pushed into the single tax brackets? I suspect this would be favorable for Roth conversions.
Keep interest as your friend, not your foe. | Use money as a tool for bettering your life, not squandering it. | Stay the course, don’t deviate from it.
User avatar
FiveK
Posts: 15742
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: Tue Oct 19, 2021 6:02 pm Glad it worked. When I revisited Google drive to make another attempt, the support thread reminded me that "if you have your account through a school or university, public linking may be disabled." So, github for now.
The ss at github and on my website is the same and it is the basic spreadsheet. It allows the user to vary the tax rate at conversion and at RMDs, and the return rates for the accounts. It does not have the permutations investigated over the last week or two (pay from taxable, convert early, track cost basis per marcopolo). I probably won't update the basic spreadsheet until all the permutations are out, still a month or two to go. But any updates will be put in the same repositories.
Up to you whether to hold off linking for now, or to link with some caveat about "basic, see thread for updates as they appear"
The latter suggestion is appealing. With any luck it can be done in less than a month or two, but the best laid plans....
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 »

sapphire96 wrote: Tue Oct 19, 2021 8:29 pm Very intriguing thread - thank you for the effort you have put into this.

Sorry if I missed this, but did you perform an analysis on if one of the spouses dies (say, at age 80) and the surviving spouse is pushed into the single tax brackets? I suspect this would be favorable for Roth conversions.
This topic gets mentioned from time to time (including in this thread), but I haven't seen detailed analysis. Part of the problem is you are dealing with two overlapping distributions for ages of death in a couple, and so there are a very wide range of possible outcomes.

I found the following paper discussing life expectancies in couples starting from around age 60 (much more meaningful than general life expectancy).
https://www.nber.org/papers/w25009

An overly simplistic summary might be to expect 18 years of MFJ after the early 60's, followed by 12 years of filing single.
User avatar
LadyGeek
Site Admin
Posts: 95696
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 »

A discussion between marcopolo and smitcat regarding investment choices has been moved into a new thread. It was derailing this discussion. See: [Why Roth conversions always pay off - if you can hold on long enough - Investment side discussion]
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.
Topic Author
McQ
Posts: 1425
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 »

iceport wrote: Tue Oct 19, 2021 6:37 pm
McQ wrote: Tue Oct 19, 2021 6:09 pm
iceport wrote: Tue Oct 19, 2021 1:57 pm ...
Therefore in my current circumstances, the correct assumption for the opportunity cost of paying the Roth conversion tax from the taxable account is to use the expected return of short- to mid-term bonds, *not* equity returns.

This thought process really drives home to me how extremely variable an accurate analysis is to individual circumstances — both in general *and* at specific points in time.

If I've said it before, I'll reiterate: I don't think attempting to formulate generalized Roth conversion recommendations is possible. It all comes down to the specific details of individual investors' lives. Those specific details have the potential to completely upend the analyses of the (theoretically) "typical" circumstances.
That is a defensible position, iceport, and you may be right. But that would leave me as an author and scholar, who must be concerned with the general case, with nothing to do. Conversely, the very difficulty of coming up with anything general in the Roth conversion space makes the project challenging and interesting, even if doomed to failure. Put another way, if it was a simple problem, I wouldn't be interested in it. But because it is not simple, my efforts may come to naught.
Scarcely matters, since I do what I want now that I'm retired, and this is one of the problems I find interesting at the moment.
I wasn't trying to imply that your efforts are bound to fail or are of limited value. I think the more constructive interpretation is that the analyses need to build in enough flexibility to handle a very wide range of circumstances. I've just been dutifully serving as a Devil's advocate, to demonstrate how different real life can be from typical assumptions. :twisted: It's really just an extension of the flexibility needed to try out the effects of various tax risks — which I really consider to be the biggest source of uncertainty in the whole endeavor.

At times I struggle to follow all the intricate details of your work, but I do find it interesting!
Don't change a thing, iceport, your feedback has been constructive, here and in prior threads, and I have learned from it.
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: 1425
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: Tue Oct 19, 2021 10:46 pm
sapphire96 wrote: Tue Oct 19, 2021 8:29 pm Very intriguing thread - thank you for the effort you have put into this.

Sorry if I missed this, but did you perform an analysis on if one of the spouses dies (say, at age 80) and the surviving spouse is pushed into the single tax brackets? I suspect this would be favorable for Roth conversions.
This topic gets mentioned from time to time (including in this thread), but I haven't seen detailed analysis. Part of the problem is you are dealing with two overlapping distributions for ages of death in a couple, and so there are a very wide range of possible outcomes.

I found the following paper discussing life expectancies in couples starting from around age 60 (much more meaningful than general life expectancy).
https://www.nber.org/papers/w25009

An overly simplistic summary might be to expect 18 years of MFJ after the early 60's, followed by 12 years of filing single.
On my list of future posts: the widowed survivor. Taxes will go up for the widow/er. And whenever future taxes go up, a Roth conversion looks better, ceteris paribus.

Interesting question, from the standpoint of conversion planning: "Odds are, one of us will die before the other. Therefore, any Roth conversion will be more successful than it appears on a "die-the-same-year" analysis. Therefore, Roth conversions should always be given a thumb on the scale. Outcomes that look skimpy on a die-together analysis will look better, odds-on, if death does not come on the same schedule for both spouses."

Is that analysis correct? Should it be weighted by the odds, that one spouse will live significantly longer than the other? Does it/should it tip the scales toward convert / don't convert? Or is is simply distant future frosting on a cake already baked?

And if the conversion involved surplus RMDs, aimed at maximizing future after-tax wealth in the distant future; then by definition the widowed survivor didn't need these funds while both were alive; so what do they care if these surplus funds are taxed at a somewhat higher rate?

sapphire96, you may enjoy this thread: viewtopic.php?f=2&t=355111
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
iceport
Posts: 6054
Joined: Sat Apr 07, 2007 4:29 pm

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

Post by iceport »

HomeStretch wrote: Tue Oct 19, 2021 3:17 pm
iceport wrote: Tue Oct 19, 2021 1:57 pm … I don't think attempting to formulate generalized Roth conversion recommendations is possible. It all comes down to the specific details of individual investors' lives. Those specific details have the potential to completely upend the analyses of the (theoretically) "typical" circumstances.
I tend to agree with this statement.

McQ, the iterations of your spreadsheet posted on this thread likely take significant time. But, is it all feasible to:
1) create say five profiles of the most common investors contemplating Roth conversions,
2) add an initial “interview” with a set of questions intended to categorize an investor into one of those profiles based on their responses, and
3) point the investor to one of five spreadsheets, each one customized to one of the profiles?

Perhaps five profiles are too many/too few, if so pick a number. But semi-customized spreadsheets might be sufficient to give an investor a better tool for their circumstances to evaluate Roth conversions.
HomeStretch, your suggestion has a lot of merit, IMHO. Personal finances are enormously complex — particularly when taxes are involved!

Your idea here reminds me a lot of how The Mortgage Professor handles his website. The site has gotten much slicker over the years, and he's added a lot more "calculators," to his arsenal, but his overall approach is to design calculators that allow practically all parameters to be adjusted to accommodate individual circumstances, but that are also tailored to specific base-level conditions, presumably to keep the calculations from becoming unreasonably complex (and possibly also for ease of use).

Just take a quick peek at his list of calculators, peruse a few of them, and you'll see what I mean: The Mortgage Professor: Mortgage Calculators and the Questions They Address



Professor McQ, I don't know if you are only interested in identifying general rules that pertain to general cases, or if you might also be interested in developing tools to help individual investors navigate the complexities of Roth conversions that factor in all the warts and wrinkles of their specific circumstances. If the latter is something that interests you at all, I think it could be enormously valuable.

In the words of marcopolo:
marcopolo wrote: Tue Oct 19, 2021 6:36 pm It is indeed a tall order. But, if it were easy, it would have been done already.
"Discipline matters more than allocation.” |—| "In finance, if you’re certain of anything, you’re out of your mind." ─William Bernstein
Topic Author
McQ
Posts: 1425
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 »

iceport wrote: Thu Oct 21, 2021 12:24 pm
HomeStretch wrote: Tue Oct 19, 2021 3:17 pm
iceport wrote: Tue Oct 19, 2021 1:57 pm … I don't think attempting to formulate generalized Roth conversion recommendations is possible. It all comes down to the specific details of individual investors' lives. Those specific details have the potential to completely upend the analyses of the (theoretically) "typical" circumstances.
I tend to agree with this statement.

McQ, the iterations of your spreadsheet posted on this thread likely take significant time. But, is it all feasible to:
1) create say five profiles of the most common investors contemplating Roth conversions,
2) add an initial “interview” with a set of questions intended to categorize an investor into one of those profiles based on their responses, and
3) point the investor to one of five spreadsheets, each one customized to one of the profiles?

Perhaps five profiles are too many/too few, if so pick a number. But semi-customized spreadsheets might be sufficient to give an investor a better tool for their circumstances to evaluate Roth conversions.
HomeStretch, your suggestion has a lot of merit, IMHO. Personal finances are enormously complex — particularly when taxes are involved!

Your idea here reminds me a lot of how The Mortgage Professor handles his website. The site has gotten much slicker over the years, and he's added a lot more "calculators," to his arsenal, but his overall approach is to design calculators that allow practically all parameters to be adjusted to accommodate individual circumstances, but that are also tailored to specific base-level conditions, presumably to keep the calculations from becoming unreasonably complex (and possibly also for ease of use).

Just take a quick peek at his list of calculators, peruse a few of them, and you'll see what I mean: The Mortgage Professor: Mortgage Calculators and the Questions They Address



Professor McQ, I don't know if you are only interested in identifying general rules that pertain to general cases, or if you might also be interested in developing tools to help individual investors navigate the complexities of Roth conversions that factor in all the warts and wrinkles of their specific circumstances. If the latter is something that interests you at all, I think it could be enormously valuable.

In the words of marcopolo:
marcopolo wrote: Tue Oct 19, 2021 6:36 pm It is indeed a tall order. But, if it were easy, it would have been done already.
I may or may not go that route, iceport, but if i do, now I have a specific role model for implementing HomeStretch's suggestion. Thanks!
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: 1425
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 »

What if I pay tax from outside the conversion? Part II

[This post may not make much sense if you haven’t read the preceding October 18th post (i.e., Part I)]

If it is more lucrative to do two conversions, i.e., by paying tax outside the conversion, which creates a second, piggyback conversion, then piggyback conversions may also protect against a bad guess about future tax rates.

Upthread, I described an innocent Roth conversion error: converting at 28% pre-TCJA in 2016, only to find oneself taking RMDs taxed at the post-TCJA rate of 24%. I called it innocent because the Roth-minded are always being told, “Taxes can only go up—you must convert today!”

The same sort of error can be made in good faith in 2021. Example: you pay income tax at 24% on the conversion, plus IRMAA, for a combined rate of about 28% today; only to find in 2026 and thereafter that TCJA has not lapsed and IRMAA brackets have been reconfigured, so that your RMDs are only taxed at 24% (looking at you, Woodspinner).

Paying tax from outside should ameliorate the impact of a bad tax forecast; but by how much?

Here is the spreadsheet, showing tax at 28% / 24%, with corresponding changes to the future value of the tax payment.

Image

Good news: in the original scenario, guessing wrong by four points meant a breakeven age of 88, with breakeven meaning: “how long it will take for tax drag to overcome the tax penalty unintentionally paid at conversion?” But adding a piggyback conversion cuts the breakeven age to 80—a more tolerable risk for many, I should think. The more so, because Woodspinner and others in the same boat make the conversion in good faith, expecting that TCJA will not lapse, and IRMAA can never go away, so that the conversion is designed to reduce RMDs that might be taxed at 32% or more. Big payoff, if correctly guessed!

Good to know that a bad guess of this magnitude can be overcome in less than ten years with the aid of a piggyback conversion, i.e., tax paid outside.

Speaking of which, suppose you do guess right about future rates, and convert today at 22% to save post-TCJA tax at 25%. How much does paying tax outside help an already successful conversion?

Image

It helps by exactly the same dollar amount. All conversions at 22%, if tax is paid outside, and if returns are as estimated, and if mark-to-market taxes are as estimated, and if inflation is 3% as estimated, produce the same increment at a given age: $9,659 real at 85, and $29,926 real at 95, against an initial tax payment of $22,000 at age 71.

In the case where the future tax rate goes up, there was already a more substantial payoff than in the constant rate case. Hence, the incremental payoff from paying tax outside is a lower percentage. But $9,659 real, after fourteen years, on a piggyback conversion of $22,000 … well, of course, you could have done so much better in Gamestop. But still: for a paperwork reshuffling of account designations, and a check written to the IRS from one account rather than another, that ain’t half bad.

Next, it is now time to follow the marcopolo way (which is not dissimilar to the procedure used in the SSRN paper). Nobody who buys VTI, or any similar market index ETF, expects to realize capital gains tax annually. No: only qualified dividends are expected to be distributed. In all my spreadsheets, these dividends are assumed to be 2%, taxed at 15%, for tax drag of 30 basis points per year.

*as so often in this thread, I pass over today’s reality: the VTI doesn’t pay 2% in dividends right now. But just maybe, over the many decades of your retirement, that might still be a good estimate?

Under constant present and future tax rates, only tax drag contributes to Roth conversion payoffs. If tax drag is limited to 30 basis points (2% dividend yield X 15% tax rate), then conversion payoffs must be lower than those reviewed thus far. But does the pattern change? How does the relative contribution, of paying tax outside, change in percentage terms—if at all?

In this spreadsheet I assume that all annual evaluations occur at death. Therefore, the capital gains portion is never taxed—it is stepped up. That leads to a simple ss set up: both the future value of the tax payment, and the future value of the reinvested RMDs, appreciate at 9.70%. Recap: only dividends are taxed, these are taxed at the qualified rate, and dividends are only 2 percent. So the stock asset appreciating at 10% pretax gives returns of 9.70% after tax.

Here is the ss.

Image

Roth conversion payoffs are much, much lower, per marcopolo. But the pattern doesn’t change. At age 85, the contribution from paying conversion tax outside exceeds the payoff from the main conversion, same as before, and by about the same proportion. Age 94 results are also parallel.
In short: no matter the accounting, paying the tax from outside the conversion is powerful. Because tax free is good. And more tax free is better.

Now in what follows, I will mostly revert from the marcopolo way to the annual mark-to-market way. The marcopolo way is not a bad metric after age 95; quite a few of us will be dead by that age, even our spouses, so that the step up will be received, unless, ahem, Congress in its infinite wisdom decides to revoke the privilege of a step up at death.

But it is not a good metric at age 85, because for affluent baby boomer couples, second to die at age 85 is a mishap, an unexpected tragedy. In between 85 and 95, the needle gradually shifts toward expecting a step up. But at the same time, the NIIT rears its ugly head, if the funds are tapped while alive. So mark to market is slightly favorable toward Roth conversions; but not by so much, in the early years, and net of uncertainty.

Next post seeks out the Optimum: convert early and pay tax outside the conversion. How good can it get?
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.
marcopolo
Posts: 8445
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: Thu Oct 21, 2021 6:01 pm What if I pay tax from outside the conversion? Part II

[This post may not make much sense if you haven’t read the preceding October 18th post (i.e., Part I)]

If it is more lucrative to do two conversions, i.e., by paying tax outside the conversion, which creates a second, piggyback conversion, then piggyback conversions may also protect against a bad guess about future tax rates.

Upthread, I described an innocent Roth conversion error: converting at 28% pre-TCJA in 2016, only to find oneself taking RMDs taxed at the post-TCJA rate of 24%. I called it innocent because the Roth-minded are always being told, “Taxes can only go up—you must convert today!”

The same sort of error can be made in good faith in 2021. Example: you pay income tax at 24% on the conversion, plus IRMAA, for a combined rate of about 28% today; only to find in 2026 and thereafter that TCJA has not lapsed and IRMAA brackets have been reconfigured, so that your RMDs are only taxed at 24% (looking at you, Woodspinner).

Paying tax from outside should ameliorate the impact of a bad tax forecast; but by how much?

Here is the spreadsheet, showing tax at 28% / 24%, with corresponding changes to the future value of the tax payment.

Image

Good news: in the original scenario, guessing wrong by four points meant a breakeven age of 88, with breakeven meaning: “how long it will take for tax drag to overcome the tax penalty unintentionally paid at conversion?” But adding a piggyback conversion cuts the breakeven age to 80—a more tolerable risk for many, I should think. The more so, because Woodspinner and others in the same boat make the conversion in good faith, expecting that TCJA will not lapse, and IRMAA can never go away, so that the conversion is designed to reduce RMDs that might be taxed at 32% or more. Big payoff, if correctly guessed!

Good to know that a bad guess of this magnitude can be overcome in less than ten years with the aid of a piggyback conversion, i.e., tax paid outside.

Speaking of which, suppose you do guess right about future rates, and convert today at 22% to save post-TCJA tax at 25%. How much does paying tax outside help an already successful conversion?

Image

It helps by exactly the same dollar amount. All conversions at 22%, if tax is paid outside, and if returns are as estimated, and if mark-to-market taxes are as estimated, and if inflation is 3% as estimated, produce the same increment at a given age: $9,659 real at 85, and $29,926 real at 95, against an initial tax payment of $22,000 at age 71.

In the case where the future tax rate goes up, there was already a more substantial payoff than in the constant rate case. Hence, the incremental payoff from paying tax outside is a lower percentage. But $9,659 real, after fourteen years, on a piggyback conversion of $22,000 … well, of course, you could have done so much better in Gamestop. But still: for a paperwork reshuffling of account designations, and a check written to the IRS from one account rather than another, that ain’t half bad.

Next, it is now time to follow the marcopolo way (which is not dissimilar to the procedure used in the SSRN paper). Nobody who buys VTI, or any similar market index ETF, expects to realize capital gains tax annually. No: only qualified dividends are expected to be distributed. In all my spreadsheets, these dividends are assumed to be 2%, taxed at 15%, for tax drag of 30 basis points per year.

*as so often in this thread, I pass over today’s reality: the VTI doesn’t pay 2% in dividends right now. But just maybe, over the many decades of your retirement, that might still be a good estimate?

Under constant present and future tax rates, only tax drag contributes to Roth conversion payoffs. If tax drag is limited to 30 basis points (2% dividend yield X 15% tax rate), then conversion payoffs must be lower than those reviewed thus far. But does the pattern change? How does the relative contribution, of paying tax outside, change in percentage terms—if at all?

In this spreadsheet I assume that all annual evaluations occur at death. Therefore, the capital gains portion is never taxed—it is stepped up. That leads to a simple ss set up: both the future value of the tax payment, and the future value of the reinvested RMDs, appreciate at 9.70%. Recap: only dividends are taxed, these are taxed at the qualified rate, and dividends are only 2 percent. So the stock asset appreciating at 10% pretax gives returns of 9.70% after tax.

Here is the ss.

Image

Roth conversion payoffs are much, much lower, per marcopolo. But the pattern doesn’t change. At age 85, the contribution from paying conversion tax outside exceeds the payoff from the main conversion, same as before, and by about the same proportion. Age 94 results are also parallel.
In short: no matter the accounting, paying the tax from outside the conversion is powerful. Because tax free is good. And more tax free is better.

Now in what follows, I will mostly revert from the marcopolo way to the annual mark-to-market way. The marcopolo way is not a bad metric after age 95; quite a few of us will be dead by that age, even our spouses, so that the step up will be received, unless, ahem, Congress in its infinite wisdom decides to revoke the privilege of a step up at death.

But it is not a good metric at age 85, because for affluent baby boomer couples, second to die at age 85 is a mishap, an unexpected tragedy. In between 85 and 95, the needle gradually shifts toward expecting a step up. But at the same time, the NIIT rears its ugly head, if the funds are tapped while alive. So mark to market is slightly favorable toward Roth conversions; but not by so much, in the early years, and net of uncertainty.

Next post seeks out the Optimum: convert early and pay tax outside the conversion. How good can it get?
This all makes sense.
It might be interesting to take this approach, and look at the less affluent case of 12% - 12% equal rate case, as well as the dreaded 22% - 12% misfire case. In that case instead of the whole 2% dividend being taxed, only the protion of it that not qualified dividends gets taxed. For tax-efficient investments, that might be something like 5% of the distribution. That would apply to both the outside money as well as re-invested RMDs.
Once in a while you get shown the light, in the strangest of places if you look at it right.
User avatar
dodecahedron
Posts: 6607
Joined: Tue Nov 12, 2013 11:28 am

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

Post by dodecahedron »

smitcat wrote: Tue Oct 19, 2021 8:41 am "For myself, I managed to raise a large pile of cash on the taxable side the year I retired by downsizing our home, but that is very much a corner case."
FWIW - I believe that this corner case may be much more common then you believe for folks in these categories. Here are just a few ways that someone coming up on early retirement may have 'newer' cash available...
- sold primary home
- sold primary home, owns another
- inheritance
- seperation bonus/payments
- sold business
- insurance
I agree with smitcat!

Also:

dividends and interest from taxable accounts
random tax rebates (a lot of these in 2020 and 2021!)
refunds from cancelled trips and big events (also a lot of these in Pandemic times!)
realizing the household has too many cars and selling one of them (again, a lot of folks had this experience in Pandemic times!)
cash bonuses from credit card signups, brokerage transfers, and similar
tax loss harvesting (not so common recently but opportunities could arise in the future)
tax gain harvesting
Topic Author
McQ
Posts: 1425
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 »

dodecahedron wrote: Fri Oct 22, 2021 12:19 pm
smitcat wrote: Tue Oct 19, 2021 8:41 am "For myself, I managed to raise a large pile of cash on the taxable side the year I retired by downsizing our home, but that is very much a corner case."
FWIW - I believe that this corner case may be much more common then you believe for folks in these categories. Here are just a few ways that someone coming up on early retirement may have 'newer' cash available...
- sold primary home
- sold primary home, owns another
- inheritance
- seperation bonus/payments
- sold business
- insurance
I agree with smitcat!

Also:

dividends and interest from taxable accounts
random tax rebates (a lot of these in 2020 and 2021!)
refunds from cancelled trips and big events (also a lot of these in Pandemic times!)
realizing the household has too many cars and selling one of them (again, a lot of folks had this experience in Pandemic times!)
cash bonuses from credit card signups, brokerage transfers, and similar
tax loss harvesting (not so common recently but opportunities could arise in the future)
tax gain harvesting
Actually, I agree with smitcat too, and with your expanded list--both good reminders to me that there are quite a few ways that someone might end up with tens of thousands of dollars of 100% cost basis funds, without having had to leave them in a bank account for years earning almost nothing.

Each one of these routes may have a low incidence (the well-timed inheritance), but collectively, they add up, thus making the advice to pay tax outside the conversion more widely practicable than I realized when I started writing the paper.
You can take the academic out of the classroom by retirement, but you can't ever take the classroom out of his tone, style, and manner of approach.
Topic Author
McQ
Posts: 1425
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, with early conversion & tax paid outside?

To review, here is a capsule summary of how Roth conversion outcomes improve by converting early, or paying tax outside, relative to the base case of converting $100,000 at 22% with RMDs also taxed at 22%.

Code: Select all

Age:                        85             95
------------------------------------------------------
Base case age 71           $ 7,629        $ 42,190
[22% --> 22%]               
Convert 6 years early      $11,319        $ 62,595
Convert 12 years early     $16,794        $ 92,870
Pay tax outside 
(age 71)                   $17,288        $ 72,116
Note the power of paying tax outside: at age 85, it outperforms even a conversion 12 years early (but not at age 95).
Next, here are the results for converting early and paying tax outside (ss not shown).

Code: Select all

Age:                        85             95
------------------------------------------------------

Convert 6 years early 
& tax outside              $30,997        $115,991
Convert 12 years early 
& tax outside              $53,226        $184,260
Wow.

$53,226 is way, way more than $7,629—the amount gained from tax drag fourteen years after an age 71 conversion with constant tax rate of 22%.
$53,226 is also a lot more than $17,288—the amount gained from paying tax outside on an age 71 conversion. And it is a lot more than $16,794—the amount gained from converting 12 years early.

And the age 95 results … speak for themselves.

What’s going on here?

Three words: power of compounding. That was a key contribution of the SSRN paper: Roth conversions can pay, in the long run, independent of future versus current tax rates, because tax drag compounds. And compounds. And compounds.

When a conversion takes place twelve years early, that’s twelve extra years of compounding. The age 85 result comes after 26 years, not 14 years. When tax is paid outside, that’s a second source of tax drag, beyond that suffered by the reinvested RMDs. Early conversion means greater compounding, and two sources of tax drag also lead to greater compounding.

That’s why the conversion payoffs get so large.

Tax drag versus tax rates

Before I came on the scene, Roth conversion advice in popular accounts had locked onto a simple rubric: conversions make sense when future tax rates are projected to be higher. Dispute, such as it was, centered on how likely it was that some individual client’s future tax rate would or would not be higher in retirement.

Here is another row in the table, to be compared to the bottom row just above ($53,226 / $184,260). The new row shows the results for a favorable movement in tax rates, in this case, converting at 22% to avert tax on RMDs at the post-TCJA rate of 25%.

Code: Select all

Age:                        85             95
------------------------------------------------------

Compare: good tax rate     $69,162        $212,690   
forecast (22% -- 25%)
& 12 years early & tax
paid outside
Big-a deal, huh? Yes, you do better if future rates go up. But the increments—about $16,000 at age 85 and $28,000 at age 95—are small relative to what has already been gained from compounding tax drag.

An advantageous movement in tax rate contributes arithmetically to Roth conversion payoffs.
Tax drag contributes exponentially.

As a final demonstration of the power of early conversions with a piggyback, let’s return to poor Bob and Barb, who converted at 22% to save tax on their RMDs of 12% (while still somehow being subject to a 15% rate on qualified dividends and LTCG, per my discussion with cas upthread).

Code: Select all

Age:                        85             95
------------------------------------------------------
Poor Bob and Barb:         $   107        $ 89,492
Convert at 22%, 
RMDs at 12%--but early 
& tax outside
When first examined, their age 71 conversion did not breakeven until age 96. But if they convert 12 years early, and pay the tax outside, breakeven moves up to age 85, and results by age 95 are not bad.

Albert Einstein didn’t say it, but: compounding truly is one of the wonders of the universe.

Actionable advice

This is one of the few points in the thread where the analysis produces a specific, practical recommendation relevant across a wide swath of prospective converters.
1. Convert early &
2. Pay tax outside

Unfortunately, the first part of the advice will be hard to implement for many. Only a few will retire early enough to be in a temporarily low tax bracket and able to convert at 59. It’s probably better rephrased as, convert as early as you can. Many will already have decided they need to beat IRMAA, which means conversion before 65 for most. The advice then reads, “conversion at 62 is better than at 64, and 61 better still.”

The second piece of advice is more widely applicable, at least for a single conversion of $100,000 or less, in a tax bracket like 22% or less.

One final point. Although the tables in this post show some (to me) stunningly positive results, in a meaningful sense these remain “best case” scenarios. They assume a retiree has at least $100,000 to allocate to stocks, and they assume that stocks perform at their post-1926 US level, i.e., 7% real. That’s a little higher than Jeremy Siegel’s 200-year history shows; and it’s way higher than the 4.5% real seen since 1900 in the World ex-USA research by Dimson, Marsh and Staunton.

Both are long-term multi-decade averages, but the possibility still has to be considered that your multi-decade retirement will be, ah, how shall I say, poorly timed. What if stocks don’t return 10% nominal for you?

Next post: How does the rate of return affect Roth conversion outcomes? (Looking at you, iceport and LilyFleur.)
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.
cas
Posts: 2258
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 »

McQ wrote: Fri Oct 22, 2021 2:07 pm
Many will already have decided they need to beat IRMAA, which means conversion before 65 for most. The advice then reads, “conversion at 62 is better than at 64, and 61 better still.”
... unless ACA premium tax credits pop into the picture. :wink:

There is a whole sub-genre of "Roth conversion" threads on bogleheads that involves the interaction of Roth conversions and the ACA PTCs. This sub-genre wasn't relevant when you were looking at conversions for retirees who were already post-Medicare. But now that you are working backwards to younger ages...

If I recall correctly, marcopolo is more familiar with this sub-genre than I am.

(And ... I think this was probably mostly a typo on your part, but it involves a real-life mistake that I see people making in boglehead's threads, so I'll point it out anyway: IRMAA when Medicare starts at 65 is determined by MAGI at age 63. So, conversions that avoid IRMAA considerations need to happen before age 63.)
MattB
Posts: 1228
Joined: Fri May 28, 2021 12:27 am

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

Post by MattB »

McQ wrote: Tue Oct 19, 2021 6:09 pm
iceport wrote: Tue Oct 19, 2021 1:57 pm ...
Therefore in my current circumstances, the correct assumption for the opportunity cost of paying the Roth conversion tax from the taxable account is to use the expected return of short- to mid-term bonds, *not* equity returns.

This thought process really drives home to me how extremely variable an accurate analysis is to individual circumstances — both in general *and* at specific points in time.

If I've said it before, I'll reiterate: I don't think attempting to formulate generalized Roth conversion recommendations is possible. It all comes down to the specific details of individual investors' lives. Those specific details have the potential to completely upend the analyses of the (theoretically) "typical" circumstances.
That is a defensible position, iceport, and you may be right. But that would leave me as an author and scholar, who must be concerned with the general case, with nothing to do. Conversely, the very difficulty of coming up with anything general in the Roth conversion space makes the project challenging and interesting, even if doomed to failure. Put another way, if it was a simple problem, I wouldn't be interested in it. But because it is not simple, my efforts may come to naught.
Scarcely matters, since I do what I want now that I'm retired, and this is one of the problems I find interesting at the moment.
I appreciate the work you've been trying to do here. But I also have to laugh at your sense of the situation. Namely, you seem to recognize that there is no general solution here. Yet reject that conclusion because it leaves you with nothing to do.

I think my question was skipped over upthread. So, I'll repeated it: If Roth conversions always paid off (if you can hold long enough), then what would be the point of traditional 401k accounts in the first place?

You solution suggests it would almost always be better to save money in Roth 401k accounts than traditional. But we know that is not the case.
cas
Posts: 2258
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 »

MattB wrote: Fri Oct 22, 2021 4:02 pm
I think my question was skipped over upthread. So, I'll repeated it: If Roth conversions always paid off (if you can hold long enough), then what would be the point of traditional 401k accounts in the first place?
Your question (routed through Roothy asking it again) got 2 responses upthread (page 3 of thread)

McQ here
FiveK here
marcopolo
Posts: 8445
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 »

cas wrote: Fri Oct 22, 2021 3:17 pm
McQ wrote: Fri Oct 22, 2021 2:07 pm
Many will already have decided they need to beat IRMAA, which means conversion before 65 for most. The advice then reads, “conversion at 62 is better than at 64, and 61 better still.”
... unless ACA premium tax credits pop into the picture. :wink:

There is a whole sub-genre of "Roth conversion" threads on bogleheads that involves the interaction of Roth conversions and the ACA PTCs. This sub-genre wasn't relevant when you were looking at conversions for retirees who were already post-Medicare. But now that you are working backwards to younger ages...

If I recall correctly, marcopolo is more familiar with this sub-genre than I am.

(And ... I think this was probably mostly a typo on your part, but it involves a real-life mistake that I see people making in boglehead's threads, so I'll point it out anyway: IRMAA when Medicare starts at 65 is determined by MAGI at age 63. So, conversions that avoid IRMAA considerations need to happen before age 63.)
Well, remember, despite some heated discussion about how to define "wealthy", this analysis is about people with a lot of income. Something like $100k+ of taxable income "off camera". By the time all your conversions are in the 22% bracket, the ACA tax credits are a moot point (with the exception of 2021/22, when even relatively high incomes can get tax credits).

For, more mere mortals, they are quite significant, adding approximately a 10% surcharge to one's Roth Conversions, even more if crossing over the "cliff"

For us, it still makes sense to convert at the 12% rate (effectively at 22% with loss of subsidy) because we will likely be in the 22.2% bracket (SS hump) later in life. Even with this analysis, I can not justify converting into the 22% bracket, costing us ~32% now.
Once in a while you get shown the light, in the strangest of places if you look at it right.
marcopolo
Posts: 8445
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 »

cas wrote: Fri Oct 22, 2021 4:26 pm
MattB wrote: Fri Oct 22, 2021 4:02 pm
I think my question was skipped over upthread. So, I'll repeated it: If Roth conversions always paid off (if you can hold long enough), then what would be the point of traditional 401k accounts in the first place?
Your question (routed through Roothy asking it again) got 2 responses upthread (page 3 of thread)

McQ here
FiveK here
The first of those answers seems to be at odds with the rest of the formulation where there is so much "off camera" income (pensions, Soc Sec) such that all RMDs are assumed to be in the 22% or higher bracket. So, I don't see how the 0, 10, amd 12% space could now be used to justify trad 401k in the first place, under those assumptions.
Once in a while you get shown the light, in the strangest of places if you look at it right.
MattB
Posts: 1228
Joined: Fri May 28, 2021 12:27 am

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

Post by MattB »

cas wrote: Fri Oct 22, 2021 4:26 pm
MattB wrote: Fri Oct 22, 2021 4:02 pm
I think my question was skipped over upthread. So, I'll repeated it: If Roth conversions always paid off (if you can hold long enough), then what would be the point of traditional 401k accounts in the first place?
Your question (routed through Roothy asking it again) got 2 responses upthread (page 3 of thread)

McQ here
FiveK here
Thank you for pointing this out. I saw it was initially passed over and never bothered to wade back into the fray.

McQ's answer is more slight of hand.

FiveK's, that: "The primary answer: conversions don't always pay out," seems to be on the ball.

Thanks again.
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 »

cas wrote: Fri Oct 22, 2021 4:26 pm
MattB wrote: Fri Oct 22, 2021 4:02 pm
I think my question was skipped over upthread. So, I'll repeated it: If Roth conversions always paid off (if you can hold long enough), then what would be the point of traditional 401k accounts in the first place?
Your question (routed through Roothy asking it again) got 2 responses upthread (page 3 of thread)

McQ here
FiveK here
I think of Roth conversions and Roth 401K contributions as decisions worth looking into for those who may have accumulated too much in TDAs or who are interested in the "right" balance between Roth accounts and TDAs.

Regarding Roth 401K v. 401K, one opportunity period to evaluate the Roth 401K vs. 401K decision for a dual income couple may be when one spouse retires early while the other continues working for a few more years. If the marginal tax rate is the same or lower than the estimated post RMD marginal rate, Roth 401K contributions by the working spouse seems to extend the "Roth conversion" even earlier and with outside money to pay taxes.

Another opportunity period is for young people with high earning potentials to contribute to Roth 401K while doing pre-professional-career jobs/paid internships or at the beginning stage of their professional career.

Newbie here, just started learning these concepts/techniques a few months ago. Is my understanding on the right track?
MattB
Posts: 1228
Joined: Fri May 28, 2021 12:27 am

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

Post by MattB »

hppycamper wrote: Sat Oct 23, 2021 1:13 am
I think of Roth conversions and Roth 401K contributions as decisions worth looking into for those who may have accumulated too much in TDAs or who are interested in the "right" balance between Roth accounts and TDAs.

Regarding Roth 401K v. 401K, one opportunity period to evaluate the Roth 401K vs. 401K decision for a dual income couple may be when one spouse retires early while the other continues working for a few more years. If the marginal tax rate is the same or lower than the estimated post RMD marginal rate, Roth 401K contributions by the working spouse seems to extend the "Roth conversion" even earlier and with outside money to pay taxes.

Another opportunity period is for young people with high earning potentials to contribute to Roth 401K while doing pre-professional-career jobs/paid internships or at the beginning stage of their professional career.

Newbie here, just started learning these concepts/techniques a few months ago. Is my understanding on the right track?
Your understanding is on the right track.
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 »

MattB wrote: Sat Oct 23, 2021 1:53 am
hppycamper wrote: Sat Oct 23, 2021 1:13 am
I think of Roth conversions and Roth 401K contributions as decisions worth looking into for those who may have accumulated too much in TDAs or who are interested in the "right" balance between Roth accounts and TDAs.

Regarding Roth 401K v. 401K, one opportunity period to evaluate the Roth 401K vs. 401K decision for a dual income couple may be when one spouse retires early while the other continues working for a few more years. If the marginal tax rate is the same or lower than the estimated post RMD marginal rate, Roth 401K contributions by the working spouse seems to extend the "Roth conversion" even earlier and with outside money to pay taxes.

Another opportunity period is for young people with high earning potentials to contribute to Roth 401K while doing pre-professional-career jobs/paid internships or at the beginning stage of their professional career.

Newbie here, just started learning these concepts/techniques a few months ago. Is my understanding on the right track?
Your understanding is on the right track.
Thanks! :beer
User avatar
teen persuasion
Posts: 2327
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 »

MattB wrote: Fri Oct 22, 2021 11:53 pm
cas wrote: Fri Oct 22, 2021 4:26 pm
MattB wrote: Fri Oct 22, 2021 4:02 pm
I think my question was skipped over upthread. So, I'll repeated it: If Roth conversions always paid off (if you can hold long enough), then what would be the point of traditional 401k accounts in the first place?
Your question (routed through Roothy asking it again) got 2 responses upthread (page 3 of thread)

McQ here
FiveK here
Thank you for pointing this out. I saw it was initially passed over and never bothered to wade back into the fray.

McQ's answer is more slight of hand.

FiveK's, that: "The primary answer: conversions don't always pay out," seems to be on the ball.

Thanks again.
For us, the point of traditional 401k contributions was eligibility for federal refundable tax credits, and then more refundable credits when our state partially matched the federal credits.

We are in a low bracket (where Roth is usually recommended), but the stacked credit phaseout rates create a high marginal tax rate. We'd likely owe little to no federal tax even if we did go all Roth, because of other nonrefundable credits, but the extra refunds allow us to save even more (and those $ go to Roth IRAs, because tIRA contributions won't increase the credits as trad 401k contributions do). So we aren't ALL traditional, we have a mix of Roth IRAs and traditional 401k. As we are in a low tax bracket, we can Roth convert the traditional accounts after early retirement a bit at a time in 0 and 10% brackets - no need to go any higher.

Another reason to use traditional 401k contributions was to reduce our AGI while filing FAFSA, to be eligible for Simplified Needs Test (no asset reporting) or Auto EFC = 0.
Gaston
Posts: 1220
Joined: Wed Aug 21, 2013 7:12 pm

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

Post by Gaston »

A key assumption in nearly all Roth conversion calculations is that the US government will keep its word, and not tax Roth distributions. With the growth in government spending programs and the ever-increasing national debt, I will not be surprised if, one day, Congress retroactively changes the rules on Roths. I hope I am wrong.
“My opinions are just that - opinions.”
Post Reply