Straightforward tax-related math problem I can't figure out

[SecondCor521]

Hi all,

I have a straightforward tax-related math problem in my RMD/conversion/IRMAA spreadsheet that I can't seem to wrap my brain around.

When I am trying to optimize my Roth conversions, I want to convert up to, but not beyond, a marginal rate that I will face later. But I know that at some point I won't be alive later, and it'll be my kids' problem to optimize their tax rates. So my goal is to calculate a weighted average marginal tax rate taking my life expectancy into account.

I have a year-by-year projection of my marginal tax rate between now and age 90. So for example, I might be in the 12% bracket now, the 22% bracket when I am 70, and a 36% bracket when I am 90.

I also have a year-by-year probability that I will be alive (based on SSA tables). So for example, I have a 100% chance of being alive today, about an 80% chance of being alive when I am 70, and about a 20% chance of being alive at age 90.

In both cases, I actually have about 40 rows, not just 3, because I'm 53 now and plan each year through age 90.

What I have done so far is just multiply the rate by the probability that I will be alive for each year, then take the moving average of those results. But that seems to result in a percentage that never makes it above about 16%, which seems intuitively to be too low for my situation. (In the example below, I just use the three data points; in my real full table with 40-ish rows, there are a lot of rows in the 22% bracket which raises the moving average column up to 16%.)

Age Tax Life% Product Average

53 12% 100% 12% 12%

70 22% 80% 17.6% 14.8%

90 36% 20% 7.2% 12.27%

Anyone understand what I'm asking about and what the proper math is?

 

[RunningBum]

I just don't think it makes sense to multiply the % chance you'll be alive by your tax bracket. Those two things are too different. It's like trying to multiply apples and dryer sheets. At least that's what it feels like to me.

I think what I would do is to pick some age you only have maybe a 25% chance of reaching, and try to optimize your conversions to level out taxes and anything else affected by added conversion income, and your heirs tax rate.

Or, you might pick some age that you could possibly hit like 100, and optimize conversions to that age. If you die earlier and leave some unconverted, that is your heirs' issue to deal with. This is what I do.

 

[gauss]

If you average in years where there is low probability of you being alive, could this be the bias that you intuitively see in your effective tax rate?

I suspect if you ran the analysis out to age 120 years, you will see your "effective rate" come down much further illustrating my point.

Perhaps using tax paid (in $) instead of tax % in your formulation you may have better results.

-gauss

 

[dak2018]

I have a year-by-year projection of my marginal tax rate between now and age 90. So for example, I might be in the 12% bracket now, the 22% bracket when I am 70, and a 36% bracket when I am 90.

I also have a year-by-year probability that I will be alive (based on SSA tables). So for example, I have a 100% chance of being alive today, about an 80% chance of being alive when I am 70, and about a 20% chance of being alive at age 90.

I think I sort of see what you are trying to do, but have a comment:

A 40 year planning horizon that depends heavily on tax policy is probably not all that useful. Unfortunately, you are not dealing with a set of physical laws but rather dealing with constantly changing rules. Your results for this calculation are essentially guaranteed to be incorrect, and probably massively so.

Can you explain the utility of (probability of being alive X marginal tax rate), as that is something that I am struggling to understand the utility of in making conversion decisions.

 

[RunningBum]

Another reason it doesn't make sense: I'm not sure why your tax rate would be 36% at age 90, but let's say it will be, while your other rates are 12% and 22%. You wouldn't convert any at 36%, right? So why would that even be factored in?

 

[SecondCor521]

Thanks all for the replies.

I just don't think it makes sense to multiply the % chance you'll be alive by your tax bracket. Those two things are too different. It's like trying to multiply apples and dryer sheets. At least that's what it feels like to me.

I think what I would do is to pick some age you only have maybe a 25% chance of reaching, and try to optimize your conversions to level out taxes and anything else affected by added conversion income, and your heirs tax rate.

Or, you might pick some age that you could possibly hit like 100, and optimize conversions to that age. If you die earlier and leave some unconverted, that is your heirs' issue to deal with. This is what I do.

Right. My current approach is to base my decisions on my 50% life expectancy, which is currently age 82. I can look at my spreadsheet and see what my projected federal marginal rate is at that age, which happens to be about 31%.

That approach is easy, but it completely ignores age 83 onward, which I have about a 50% chance of living at least some of those years.

The more complicated approach in my OP is trying to incorporate all of those years into the math.

If you average in years where there is low probability of you being alive, could this be the bias that you intuitively see in your effective tax rate?

I suspect if you ran the analysis out to age 120 years, you will see your "effective rate" come down much further illustrating my point.

Perhaps using tax paid (in $) instead of tax % in your formulation you may have better results.

-gauss

Well, yeah.

In the OP, there's only a 20% chance I'll be alive at age 90, and so that results in a low number in the fourth column, which brings down the running average in the last column.

Which can be roughly translated as "Yeah, I might be in the 36% bracket, but there's an 80% chance I'll be dead, so I probably shouldn't care too much about that."

Your suggestion is interesting. I have effective rate (tax paid / taxable income); maybe I should use that instead.

I think I sort of see what you are trying to do, but have a comment:

A 40 year planning horizon that depends heavily on tax policy is probably not all that useful. Unfortunately, you are not dealing with a set of physical laws but rather dealing with constantly changing rules. Your results for this calculation are essentially guaranteed to be incorrect, and probably massively so.

Can you explain the utility of (probability of being alive X marginal tax rate), as that is something that I am struggling to understand the utility of in making conversion decisions.

I appreciate the comment. I agree things will change. My MO is to make plans and decisions based on current law and my best assumptions for things like rates of return and inflation and life expectancy. Then I regularly iterate and update my plans as things like tax law and my account balances and IRMAA thresholds change.

With this approach, I will probably be wrong as the future unfolds, but it's the best I can do without my crystal ball.

As for the explanation:

I do Roth conversions and pay taxes now in hopes that my marginal rate now is less than it will be when I am older and RMDs and SS force me into higher brackets. Google the term "tax torpedo" or read Ed Slott or any number of Roth conversion threads on this site to learn more about the general issue if you're not familiar.

At some unknown point in the future, I will die. Probably not this year, and only a small chance the year after that, but that probability increases over time to a virtual certainty by age 115.

When I die, I will no longer be paying income taxes at my marginal rate on Roth conversions; my three kids will inherit my IRA and they'll pay income taxes over 10 years as they drain the remainder at their marginal rates.

So let's say I'm converting based on my age 82 tax rate of 31% and I Roth convert up to 31% this year.

Let's say I live longer than that, to age 90. Well at age 90 my tax rate would be 36%, so I didn't convert enough this year, because it would have been better for me to convert more, up to 36%.

Or I might not live as long - let's say I die at age 55. Well then I converted at 31%, which is probably more than what my three kids would have to pay as they drain the remainder over the following 10 years. So I converted too much.

The "proper" balance between too little and too much is to take into consideration what my tax rates would be (as best as I can plan for them, which for me is assuming status quo and historical averages), and weight them (somehow) by the chance that I'll be dead. I almost certainly won't be dead next year, so I should almost certainly care about my tax rate next year. But I'll almost certainly be dead at age 115, so I almost certainly don't care about my tax rate then.

There is a continuum between now and age 115, and the associated probability I'll be alive, which I can easily calculate using the SSA life expectancy table.

...

I should probably clarify that my current plan has the feature that my marginal rate climbs over time. The reason for this is because I (a) assume a growth rate on my traditional IRA which exceeds the inflation adjustments in the tax brackets and IRMAA thresholds, (b) because I'm relatively young and have a plan that runs to age 90, compounding has a big effect, (c) I don't spend much anyway, and (d) in general I'm one who prefers to delay taxes now even knowing I might have to pay more later. Because of all of this, I have deliberately chosen not to completely levelize my marginal rate in my plan yet - a few years now in the 22% bracket reduces my higher rates in my 70s/80s/90s by quite a bit.

 

[SecondCor521]

Another reason it doesn't make sense: I'm not sure why your tax rate would be 36% at age 90, but let's say it will be, while your other rates are 12% and 22%. You wouldn't convert any at 36%, right? So why would that even be factored in?

TLDR: First world problems. I might not want to, but compounding, RMDs, and SS would make me do so.

I sort of explained this at the end of my previous post, but my current plan is sort of a stair-step plan with the following features:

1. It incorporates my SS and RMD as required income.

2. It generally "fills in" and converts to the top of the next tax bracket or IRMAA threshold.

3. It assumes an IRA that grows at 10% but inflates most brackets and thresholds at 3%.

4. It incorporates the IRS divisors for RMDs.

5. It inflates IRMAA premiums at 7.7%

As a result of all of the above, when I am 90, I have to take about 8% of a large traditional IRA as an RMD and almost $100K in taxable SS, which would put me in the 24% OI bracket. I would also have IRMAA in the fourth tier which adds the equivalent of another 12% tax.

...

I really don't like the 7.7% increase in IRMAA premiums aspect of my spreadsheet, but my rule is to pick the most likely historical average I can find, and 7.7% was the number I came up with. This sort of goes to the previous poster's point: I really don't think I'll be paying $92K in IRMAA surcharges in 2059, but that's what the math seems to say.

 

[all4j]

Like some others, I'm not sure the "product" column is meaningful. Without seeing the actual 40+ rows, can speculate...for example, 1st year of 12% x 100% will be included in every following average; whereas if the bracket didn't move up for 17 years (age 70) & then only 80% or less... hard for me to see that pulling the average up. I don't get the "weighted" aspect maybe.

I do have a few questions though. I did something along these lines before, but without so much rigor. My objective from yours was different. I was really trying to minimize overall (mine & heirs) tax spend & maximize inheritance. So, bear that in mind.

1) Am I right that you are sorta assuming 7% real growth? That is, increasing tax bracket by 3% inflation, but 10% ira growth?

2) Seems as if filling up to, but not beyond, a marginal bracket is really your limiting factor. If doing that each year, how do you say at age 80-something "I should have done more earlier"? If you didn't have that constraint, there are a variety of possibilities, but no guarantees.

 

[SecondCor521]

(I'm attaching a graph that shows my stairstep plan. This one is AGI based and shows the IRMAA and ACA thresholds. Where the AGI line is not at a threshold, it's because it's at a tax bracket threshold or an AOTC AGI limit.)

Like some others, I'm not sure the "product" column is meaningful. Without seeing the actual 40+ rows, can speculate...for example, 1st year of 12% x 100% will be included in every following average; whereas if the bracket didn't move up for 17 years (age 70) & then only 80% or less... hard for me to see that pulling the average up. I don't get the "weighted" aspect maybe.

I do have a few questions though. I did something along these lines before, but without so much rigor. My objective from yours was different. I was really trying to minimize overall (mine & heirs) tax spend & maximize inheritance. So, bear that in mind.

1) Am I right that you are sorta assuming 7% real growth? That is, increasing tax bracket by 3% inflation, but 10% ira growth?

2) Seems as if filling up to, but not beyond, a marginal bracket is really your limiting factor. If doing that each year, how do you say at age 80-something "I should have done more earlier"? If you didn't have that constraint, there are a variety of possibilities, but no guarantees.

Right.

So the goal of my efforts is to figure out what marginal rate I should Roth convert up to now, based on my projected future marginal tax rates and projected future mortality curve.

I'm willing to stipulate that the math I presented in the OP might be wrong. What I'm hoping for is that someone understands what I'm trying to do and how to do the math properly. It can't be hard - conceptually it's a weighted average - but the math seems to escape me.

Answers:

1. Pretty much, yes. My spreadsheet incorporates tax brackets, 400% FPL, IRMAA cutoffs, IRMAA premiums, SS growth, and IRA growth. IRA growth is 10%, IRMAA premiums are 7.7%, the rest are mostly CPI of 3%.

2. I'm not sure I understand your question/point here.

I also guess I didn't explain this very carefully. My spreadsheet allows me to set a target conversion based on either a tax bracket, 400% FPL, or an IRMAA threshold each year. I can put in, for each year, whatever I want to aim for as a target.

I currently generally choose, because I want to delay taxation, to only fill up to the top of the bracket that my "required income" (SS, RMD, inherited IRA) puts me in for any given year. Because my IRA is growing at 7% real, though, the compounding means that with my current plan, the "required income" pushes me into higher and higher brackets over time - 12% -> 22% -> IRMAA1 -> IRMAA2 -> etc.

I currently have my spreadsheet set up for a single "pulse" year where, in 2025, I convert to the top of the 22% bracket. A relatively larger early conversion at a relatively early age (56), this results in getting me out of tax brackets in the 30-40% range in my 70s/80s/90s. This also might work out well with my kids' college tax credits and the TCJA rate expiration 1/1/2026.

I could do more pulse years, but have chosen not to for various reasons I could go into if it matters.

I could even out or levelize my plan more if I wanted to by Roth converting more earlier, but mostly have stuck to my stair step idea. If I levelized, I might be able to, for example, stay entirely in the 22% bracket and avoid IRMAA surcharges. But again, I circle back to the fact that the assumptions that would have to turn out correct to make that the proper play: (a) I have to live long enough, (b) my IRA has to grow at 7% real during that time, (c) IRMAA surcharges have to go up at 7.7% nominal during that time, etc.

In all cases, whatever targets I put in for my plan, the subsequent years all mostly recalculate automagically. So the graph below shows what happens if (a) I follow my plan and (b) my assumptions about the future are accurate. Even following my plan with all the conversions I already plan to do between now and 82, the spreadsheet shows I'll be at a 31% marginal rate at age 82. So this year, I'll convert up until my rate exceeds that 31% rate.

Lather, rinse, repeat next year.

What should happen over time, as actual inflation and SS and IRMAA and FPL, IRA performance, TCJA tax rates, inherited IRA timing and amounts - as all that varies - that data all gets fed back into the model. Maybe next year my age 82 tax rate will be 32%, or 28% or whatever. I'll convert next year up to 32% or 28% then.

I'm sure as I go along and if/when I reach 82, I'll discover that I've sometimes converted only to the 28% level and other times up to a 33% level or whatever. And I won't regret that or beat myself up; I'll be fine with it. I'll just recognize that the inputs to my plan changed over time and I don't have a perfect crystal ball.

 

[SecondCor521]

P.S. - What happens in practice is that the age-82 marginal tax rate is calculated at 31% or whatever, and then I hit some threshold income level where my marginal rate jumps. Last year this was a FAFSA-related income level where my marginal rate went from 24% to 41%. So that was a clear cutoff for me.

(This also points out that for my purposes, I include fed + IRMAA + state + FAFSA in my analysis. But for the point of this thread, that's an extraneous detail.)

 

[all4j]

(I'm attaching a graph that shows my stairstep plan. This one is AGI based and shows the IRMAA and ACA thresholds. Where the AGI line is not at a threshold, it's because it's at a tax bracket threshold or an AOTC AGI limit.)



Right.

So the goal of my efforts is to figure out what marginal rate I should Roth convert up to now, based on my projected future marginal tax rates and projected future mortality curve.

I'm willing to stipulate that the math I presented in the OP might be wrong. What I'm hoping for is that someone understands what I'm trying to do and how to do the math properly. It can't be hard - conceptually it's a weighted average - but the math seems to escape me.

Answers:

1. Pretty much, yes. My spreadsheet incorporates tax brackets, 400% FPL, IRMAA cutoffs, IRMAA premiums, SS growth, and IRA growth. IRA growth is 10%, IRMAA premiums are 7.7%, the rest are mostly CPI of 3%.

2. I'm not sure I understand your question/point here.

I also guess I didn't explain this very carefully. My spreadsheet allows me to set a target conversion based on either a tax bracket, 400% FPL, or an IRMAA threshold each year. I can put in, for each year, whatever I want to aim for as a target.

I currently generally choose, because I want to delay taxation, to only fill up to the top of the bracket that my "required income" (SS, RMD, inherited IRA) puts me in for any given year. Because my IRA is growing at 7% real, though, the compounding means that with my current plan, the "required income" pushes me into higher and higher brackets over time - 12% -> 22% -> IRMAA1 -> IRMAA2 -> etc.

I currently have my spreadsheet set up for a single "pulse" year where, in 2025, I convert to the top of the 22% bracket. A relatively larger early conversion at a relatively early age (56), this results in getting me out of tax brackets in the 30-40% range in my 70s/80s/90s. This also might work out well with my kids' college tax credits and the TCJA rate expiration 1/1/2026.

I could do more pulse years, but have chosen not to for various reasons I could go into if it matters.

I could even out or levelize my plan more if I wanted to by Roth converting more earlier, but mostly have stuck to my stair step idea. If I levelized, I might be able to, for example, stay entirely in the 22% bracket and avoid IRMAA surcharges. But again, I circle back to the fact that the assumptions that would have to turn out correct to make that the proper play: (a) I have to live long enough, (b) my IRA has to grow at 7% real during that time, (c) IRMAA surcharges have to go up at 7.7% nominal during that time, etc.

In all cases, whatever targets I put in for my plan, the subsequent years all mostly recalculate automagically. So the graph below shows what happens if (a) I follow my plan and (b) my assumptions about the future are accurate. Even following my plan with all the conversions I already plan to do between now and 82, the spreadsheet shows I'll be at a 31% marginal rate at age 82. So this year, I'll convert up until my rate exceeds that 31% rate.

Lather, rinse, repeat next year.

What should happen over time, as actual inflation and SS and IRMAA and FPL, IRA performance, TCJA tax rates, inherited IRA timing and amounts - as all that varies - that data all gets fed back into the model. Maybe next year my age 82 tax rate will be 32%, or 28% or whatever. I'll convert next year up to 32% or 28% then.

I'm sure as I go along and if/when I reach 82, I'll discover that I've sometimes converted only to the 28% level and other times up to a 33% level or whatever. And I won't regret that or beat myself up; I'll be fine with it. I'll just recognize that the inputs to my plan changed over time and I don't have a perfect crystal ball.

Some scattered thoughts. I'm still not grasping what equates to "optimize" & not sure the "product" column really gets to anything meaningful. But, that may be just me & is really irrelevant. I bolded a couple of lines that stood out. For example, if avoiding going above 22% bracket & avoiding Irmaa isn't "optimal", then what would be? It is a moving target, so perhaps can't quantify today, but otherwise any plan is as good as another.

Most don't understand the significance of assumptions in models such as this. In your case, I think the 10% return & 3% inflation are quite optimistic & thus most will drive the model to have you convert sooner than actuals might. Perhaps worth trying to see how different it might be if you tweaked those; would it make meaningful shifts?

I'm somewhat curious as to source of funds to pay taxes on converted amounts -- & how are you modeling the impacts of that? For example, if using non-ira sources, what if they weren't used to pay tax?

Lastly, if anything, I'd change the average to look at upcoming years & not previous. When you are age 60, how can it be useful at all to factor in your age 53 info in deciding how much to convert?

Glad you won't be 2nd guessing yourself & you'll be ok.

 

[SecondCor521]

Some scattered thoughts. I'm still not grasping what equates to "optimize" & not sure the "product" column really gets to anything meaningful. But, that may be just me & is really irrelevant. I bolded a couple of lines that stood out. For example, if avoiding going above 22% bracket & avoiding Irmaa isn't "optimal", then what would be? It is a moving target, so perhaps can't quantify today, but otherwise any plan is as good as another.

Most don't understand the significance of assumptions in models such as this. In your case, I think the 10% return & 3% inflation are quite optimistic & thus most will drive the model to have you convert sooner than actuals might. Perhaps worth trying to see how different it might be if you tweaked those; would it make meaningful shifts?

I'm somewhat curious as to source of funds to pay taxes on converted amounts -- & how are you modeling the impacts of that? For example, if using non-ira sources, what if they weren't used to pay tax?

Lastly, if anything, I'd change the average to look at upcoming years & not previous. When you are age 60, how can it be useful at all to factor in your age 53 info in deciding how much to convert?

Glad you won't be 2nd guessing yourself & you'll be ok.

The optimal plan won't be known in advance. The argument for converting to 22% now to avoid IRMAA later is based on me actually being alive during my IRMAA years to avoid it.

But I can try to do optimal based on my best guesses of the future. That's what I'm trying to do.

I appreciate the critique that I'm being too optimistic. I've heard such comments for years, maybe decades, during which my portfolio has gone up 10% and inflation has averaged 3%. I've also read the arguments against optimism and generally disagree with them.

I agree that assumptions are critical. A sensitivity analysis might be useful - it least might convince me that predicting my tax rate in 2051 to two decimal places is excessive precision with little accuracy.

I have not had to pay taxes on my Roth conversions yet, really. They are taxable income, but my tax situation is such that I owe no federal taxes and little to no state taxes. But when I do, I would pay them from my taxable account. For the purposes of this spreadsheet (which generally focuses on forced income and the taxes on that forced income - RMDs, SS, and an inherited IRA at some point), I don't model the taxable account at all. The reason is that the taxation of my SS and RMDs and inherited IRA are going to happen and be due at a minimum regardless of what happens in my taxable account.

I do not look at past years in my spreadsheet. It is always current year to age 90, so as the years go by I will delete a row each year. What I was trying to say is that when I get to age 82 or 90 or whatever (if I do), I could look back and at that future point realize that my plan was not optimal in retrospect, but that I wouldn't do that since I can't change the past. It wouldn't be optimal in retrospect because over the years my guesses of the future would have obviously been off. But I am hopeful that by making my best guesses each year and updating my inputs each year and iterating, that I'll at least get fairly close and as close as I can get given no actual crystal ball.

 

[SecondCor521]

While I'm enjoying the conversation, this thread has gone off on some tangents.

To put the question as precisely as I can:

You are 53. Assume you believe at the following ages you will have the following federal marginal tax rate and chance of being alive:

Age Tax Life%

53 12% 100%

70 22% 80%

90 36% 20%

Knowing nothing else, to what federal marginal rate would you believe it would be appropriate to Roth convert this year?

 

[NgineER]

Considering the TCJA sunsetting in 2025 if not extended and that taxes will be due at some point, either paid by you or your kids when inheriting your IRA, I'd go for the 22% bracket now for the next couple of years. I'd do the same myself if I was in your shoes.

 

[Dash man]

I don’t think figuring mortality rates helps in any way. Isn’t the goal to maximize the funds you keep and minimize taxes you pay. Tax rates and portfolio growth rates are already just assumptions we use for planning, but can be way off. They will not go up and down in a straight line no matter how accurate we are over time in our guesses. Adding mortality is just throwing an unnecessary wrench into your plan. I suggest planning to age 90 or 95, because that’s when previous Roth conversions will benefit you, and actually much earlier.

 

[cathy63]

The optimal plan won't be known in advance. The argument for converting to 22% now to avoid IRMAA later is based on me actually being alive during my IRMAA years to avoid it...

Why are you trying so hard to avoid IRMAA? It's not just you, I see lots of people worrying about potential future IRMAA costs and I always feel like there's something I'm missing when it comes up.

The max IRMAA penalty (without Part D) this year is 12 x (578.3-170.1) = $4898.40 for single people who have income over $.5M.

If you are at the top of the 12% bracket and decide to convert to the top of the 22% bracket this year, you'll pay .22 x (89076-41776) = $10,406.00 in taxes.

This always seems to me like taking on a larger tax burden now in order to avoid what's essentially a fairly small burden later. There are other reasons why it can make sense to pay more tax now, but avoiding IRMAA seems more like a nice side benefit rather than a driving reason to do it.

(In the meantime, our current ACA plan for two people is over $17K before subsidies, so even at the very highest IRMAA rates, which I don't expect to pay, Medicare is a bargain by comparison.)

 

[RunningBum]

While I'm enjoying the conversation, this thread has gone off on some tangents.

To put the question as precisely as I can:

You are 53. Assume you believe at the following ages you will have the following federal marginal tax rate and chance of being alive:

Age Tax Life%

53 12% 100%

70 22% 80%

90 36% 20%

Knowing nothing else, to what federal marginal rate would you believe it would be appropriate to Roth convert this year?

I won't even try. At a minimum I'd need to know your rate at age 72 when RMDs would start, and what the size of your tIRA is, and how close you are to the top of those brackets. I would also need to know if any of those rates already include RMDs, and if that's the rate assuming no conversions at all.

It may make sense to go to the top of 24% if you're going to be at least that high at 72 with RMDs. But there's the whole balancing thing of conversions now reducing RMDs later so you kind of have to map that out with a spreadsheet. People want a simple answer of how much to convert but it's not a simple question.

If I had a large tIRA balance, I'd probably convert to 24% now, figuring that 22% might go to 24%, and even if it doesn't, converting at 24% now and paying 22% later isn't too bad. Also, even if you think making it to 90 is unlikely, I would try not to have any RMDs on top of that 36% rate. Or is that your likely rate with projected RMDs? See, I'm just making a guess without really knowing what your numbers include. It's hard enough to make my own calculations so I'm back to not guessing for someone else if I really don't have a clear picture of the numbers. And getting a clear picture on someone else's situation is not something I really want to make the time or effort for.

The good news is, you get your biggest bang for the buck converting at obvious rates. As it gets harder to decide on perhaps 22% or 24%, there's probably not a huge difference either way.

 

[DawgMan]

Roth conversion strategies at this level frankly make my head spin. That said, I would like to optimize my conversion strategy like many of you here, but due to so many unknowns, I default to the KISS method...

Step 1 - Assume I Disregard Roth Conversions
- Based on what I know today/assume in the future, can I fund my planned spend until the projected dirt nap?
- If/when one spouse dies, same question?

Step 2 - Minimize Lifetime Taxes/Maximize Legacy
- Based on what I know today/assume in the future, what should my Roth conversions be....
a) before age 65 (arguably 63 before IRMMA is calculated
b) between 65 (or 63) - 70 when I plan to collect SS
c) before RMDs
- Each year, look at all income (i.e. dividends, interest, capital gains) and then Roth convert to specific tax bracket
- Rinse and repeat annually.

The way I look at it, if I can answer yes to the questions in Step 1 then both DW and I will be fine. Sure, a single spouse will pay more taxes as well will my kids... but so what?

That said, I will move on to Step 2, but won't sweat it if I look back at age 70 and say "I really wish I would have converted another $50K when I was 65!"

But if you all come up with the perfect spreadsheet, would love to see it!

 

[levindb]

Some random thoughts...

1. A 10% ROI on tIRA may be a little aggressive as you age. I would expect you to reduce from 100% equities as you age to maybe 60/40 or lower. Your ROI should be adjusted as well.

2. To reduce the tax to heirs, start giving them $16k/year right now. No tax to you and no tax to them...in fact, it starts reducing your future tax burden when RMDs start.

3. Look for alternate ways to get more deductions through donations to charities or acquiring rental properties. There is more to managing your taxes than just Roth conversions.

 

[MarieIG]

I was going to do equal Roth conversions for 22, 23, 24 and 25; but may bump up this year.

Barring unforeseen circumstances, I don't see our income going down in the next few years, rather turning on SS and RMDs should bring it up. Conversions are an attempt at some tax smoothing.

I don't know that I will drain my traditional IRA before RMDs, but to the extent that there is any of it left, I would like to use it for direct charitable donations.

 

[SecondCor521]

Thanks for the additional replies.

@cathy63, I'm not particularly trying to avoid IRMAA; I think I happened to use it as an attempt at a simplified example in a reply. What I should have written was "The argument for converting to 22% now to avoid [higher taxes] later is based on me actually being alive during [those later] years to avoid [them]."

What I'm trying to do is minimize my lifetime taxes, which for me includes fed marginal, state marginal, IRMAA, and FAFSA. IRMAA happens to seem to be a big deal in my spreadsheet due to several factors which cascade and compound over the next 40 years or so: My IRA compounding at 7% real produces large RMDs. That plus my pretty good SS, ends up pushing me into the various IRMAA tiers. And my assumption that IRMAA premiums rise at 7.7% annually (I don't think they will, but that's the number I found) means that my IRMAA surcharge in my 90s is equivalent to an additional nearly 12% marginal rate. (We of course can disagree on whether my assumptions are good, and whether my spreadsheet math is correct.)

It also just so happens that with my stair step approach, most of my tax rates in my 70s/80s/90s happen to be stair stepping through the IRMAA brackets (which for me, as a single, appear to fall between the 22% and 24% brackets).

@RB, thanks for the reply. I understand how it's hard to get inside another person's plan and strategy and assumptions. Simply put, I'll be in the 22% bracket from about my mid-60s to my mid-70's, and then in the 24% bracket beyond that, and that rate includes me doing Roth conversions every year according to my plan. I start hitting IRMAA tiers at age 74 and the surcharge gets high starting around age 82.

@DawgMan, the key question I'm trying to address is "to what specific bracket?" at the end of your Step 2. I already know I have more than enough for myself. I just try to optimize on this stuff because I've already figured everything else out ;-P

@levindb, for reasons that I've discussed elsewhere, I'm currently at 98% stocks. The only reason that number would go down would be if I start spending a whole lot more and I had a much shorter life expectancy.

Regarding giving to my kids, I've already started that for the reasons you point out as well as several others.

Regarding deductions, I already do those as well, although not the ones you mention. I left them out of this thread because it was something that I thought was not directly relevant. In particular, I do have a QCD column in my spreadsheet that I can play with; thanks for the reminder on that.

...

Overall, again, I find myself tilting at windmills to try to optimize something that is difficult to optimize and from a position where I really don't need to (I'm at somewhere around a 0.6% WR). Then I try to let it go. <sigh>

Again, thanks all.

 

[SnowballCamper]

This problem is what Optimal Retirement Planner - Home and Essential Parameter Form was excellent at solving...because the tax code is a set of linear constraints, an assumed straight line growth model is a linear constraint, and the problem is to maximize lifetime after tax income, subject to the above constraints.

Multiplying a tax rate % by a probability of being alive yields a meaningless number.

There really isn't any substitute to solving a linear programming problem... this is why it is the subject of college courses.

 

[SecondAttempt]

I understand what you are trying to do but I think you are overanalyzing this. Trying to make tax decisions over multiple decades is hopeless. Just go look at what tax rates were in 1982. Totally different from today.

I'm not trying to be critical of you. I also have a massive and very detailed spreadsheet going way past my likely demise. But I use somewhat vague "average tax rates" (marginal and total, state and federal) without trying to predict actuals in any specific year. I can make tactical decisions over the rolling next couple of years when I have actual information.