Roth conversions and IRMAA (again) (seesaw effect)

[SecondCor521]

Hi all,

I've been playing around with my RMD / QCD / conversion spreadsheets again. I am trying to figure out in some detail how I want to even out my tax bracket over time.

I am single and 52. I plan to 85; my life expectancy is probably there or a bit more if I'm reasonably lucky.

It appears to me that the first three IRMAA breakpoints occur between the 22% and 24% bracket for single filing status.

What I have discovered is that I am in the position where I get to choose whether to convert up to the top of the 22% or 24% bracket over the next several years. There is about a $90K difference between those two points.

Assume for the sake of discussion that I have the money to pay the conversion taxes in my taxable account.

1. I am seeing a seesaw effect where, if I convert to 24% instead of 22% in an earlier year (say 2027 when I'm 58), that whacks down my IRA balance to where I can have more later years in lower IRMAA brackets.

2. I also see a smaller seesaw effect where, if I convert up to the first or second IRMAA breakpoint a year or two earlier in my 70's, then I can avoid the third IRMAA breakpoint when I'm about 85.

I don't think it matters a whole lot, but any input on whether anyone "pays for" a longer period in a lower bracket later with an earlier Roth conversion "big gulp"? I'd appreciate thoughts on the tradeoffs.

The only things that have occurred to me is (a) IRMAA breakpoints don't matter before age 65 or so, (b) whether I have the money available to pay the taxes, (c) doing this seesaw now and converting to 24% early means I'm voluntarily paying more taxes now in an effort to potentially avoid some taxes later, (d) the growth that I'm projecting may not materialize.

So writing that last paragraph makes me think converting to 22% early on and then working my way in stair step fashion up the IRMAA tiers in my 70s and early 80's is the way to go.

[Dtail]

I assume your comment takes into effect the 2 years before age 65 to determine IRMAA effects.

[SecondCor521]

Quote:

Originally Posted by Dtail I assume your comment takes into effect the 2 years before age 65 to determine IRMAA effects.

My spreadsheet includes IRMAA calculations with the 2 year offset, yes.

[RunningBum]

Sure, the "big gulp" makes sense in many situations. If you max out 24% and later leave room in the 22% bracket it makes less sense.

I would keep an eye on whether tax rates return to previous rates in 2026 and do the larger conversion before then.

Apparently ACA subsidies are not a factor for you?

[SecondCor521]

Quote:

Originally Posted by RunningBum Sure, the "big gulp" makes sense in many situations. If you max out 24% and later leave room in the 22% bracket it makes less sense. I would keep an eye on whether tax rates return to previous rates in 2026 and do the larger conversion before then. Apparently ACA subsidies are not a factor for you?

They are from now through 65, which is something I remembered this afternoon

In my case, the 400% FPL level is a bit into the 22% bracket. So what I might do is still convert to the top of the 22% bracket, because that would be 22% + 8.5% = 30.5% - high, but if it keeps me out of the 32% bracket in my 80s it might be worth it. I'd have to swallow a tax hump between about 200% and 400% FPL (see second graph at https://seattlecyclone.com/aca-premi...s-2021-edition), but once I completely lost my subsidy I think I'd still have a bunch of plain 22% dollars.

I'll have to look at it year by year, of course. The other thing I've started to think about is that if I die at 75 or so, then avoiding the 32% bracket in my 80's is sort of an "N/A" situation, and then it matters what my kids' average rates will be. Which would be total guess.

And yeah, like all other threads on this, the 2026 reversion and my current thinking makes the 22% bracket over the next few years pretty attractive.

[SevenUp]

Quote:

Originally Posted by SecondCor521 I've been playing around with my RMD / QCD / conversion spreadsheets again. I am trying to figure out in some detail how I want to even out my tax bracket over time.

It's a horribly non-linear problem, but have you tried Excel's Solver to attempt to find an optimal conversion pattern?

[SecondCor521]

Quote:

Originally Posted by SevenUp It's a horribly non-linear problem, but have you tried Excel's Solver to attempt to find an optimal conversion pattern?

I had in the past, but it has been a while. From what I recall, Solver seems to think that calculating the amount of income tax due (which is part of my spreadsheet, of course) is non-linear and refuses to try. I don't understand why, but that's what I remember the problem being.

Lately, I'm not even sure what my objective function should be. Sum of all income after tax? Discounted cash flow of after tax income? Either of those two plus my remaining traditional IRA balance? Or plus the NPV of my traditional IRA balance? Derated by what tax rate?

I actually tried fairly hard to generate a reasonable objective function. From trial and error with that, it seemed to want me to delay taxation as much as possible, even if that meant the 12% bracket now and the 35% bracket when I would be 85.

In the past few days I've been toying with the idea that my objective function was broken somehow, and am just sort of playing with the spreadsheet and seeing what makes sense.

The attached is what my current thought of the hour is. It turns out that my optimal RMD/QCD approach might also be somewhat affected by my inherited RMD plan (I might disclaim part), which in turn could well be impacted by my understanding of GST and how the Build Back Better Act gets paid for.

...

The left side is the tax brackets; the right side is the IRMAA brackets. The blue line is the same in both cases and is my taxable income plan.

The first two dots are low while my kids are in college.

The next spike up to 24% is to try to keep me out of the higher brackets later (which, as you can see, doesn't quite work - there's that last dot at 32%). It is done that year because (a) it's before 2026, and (b) earlier is more impactful on slowing down compounding growth, and (c) I'd rather give up earlier ACA subsidies rather than later ones. I need to see if I have the cash on hand to pay the taxes. I suppose I could do two years at 24%, but I'm not sure that is necessary, and I guess at this point I'd rather risk the 32% bracket when I'm 85 than come up with that much in cash for the taxes now.

Then I drop down to 400% of FPL for several years.

The next bump is a guess on the 10 year SECURE Act period for an inherited IRA. I don't expect it but I have to plan on most likely case.

Then I just stair step, always converting up to the top of the bracket I'm in. I get into higher brackets as time goes by because I expect my IRA to outgrow the tax and IRMAA brackets.

[SevenUp]

Quote:

Originally Posted by SecondCor521 I had in the past, but it has been a while. From what I recall, Solver seems to think that calculating the amount of income tax due (which is part of my spreadsheet, of course) is non-linear and refuses to try. I don't understand why, but that's what I remember the problem being.

That will happen if you try to use the Simplex LP solver. Either GRG Nonlinear or Evolutionary should at least try.

Quote:

Lately, I'm not even sure what my objective function should be. Sum of all income after tax? Discounted cash flow of after tax income? Either of those two plus my remaining traditional IRA balance? Or plus the NPV of my traditional IRA balance? Derated by what tax rate? I actually tried fairly hard to generate a reasonable objective function.

That was probably effort well spent, because any decent optimization algorithm will merely maximize what you tell it to maximize, within the boundaries you give it. Might as well maximize what is important to you.

[Exchme]

Your post mentioned QCDs, but the graphs and math didn’t seem to address them? Charitable giving during your lifetime is of course the difficult thing to value in $, everything else can simply be looked at in the estate value (or the value to heirs after paying taxes on an inherited IRA).

For lack of a better method, I would value lifetime charitable giving by envisioning that you are giving to a person that is going to invest your money the way you would. That way, you will not be punishing the result for giving early.

I would put together several pieces:
Years_Alive –remaining life expectancy at start
Inflation – average inflation rate assumed
AvgReturn – the average nominal return you are expecting, before inflation correction
IndexYear – years from start, (here beginning with zero)
GiftValue – amount given in any year. Can vary from year to year. Assume it is taken from t-IRA so no taxes

Create a row or column, with an entry for each year using the FV function to take that year’s gift and turn it into an equivalent future value:

=FV(AvgReturn ,Years_Alive - IndexYear, 0, -GiftValue)

If you instead wanted an inflation corrected future value,

=FV(AvgReturn ,Years_Alive - IndexYear, 0, -FV(Inflation, IndexYear + 1, 0 , GiftValue))

Just sum those year-by-year values and add that to your estate. All you are doing is assuming that investments in people in need will do as well as for the world as appreciation in stocks and bonds gifted later (which are just the fruits of the labor of people already working). The value to you is to get the compounding off the books so you don't get driven to unnecessarily high tax brackets based on money you plan to give away.

If I was giving during my lifetime to an heir instead of a charity, I would use the same basic approach, though the heir would face taxes on investment returns each year, so I would reduce the AvgReturn slightly for that and of course the money would come from my taxable account instead of the t-IRA.

While this approach gives you a metric, the problem looks well beyond the capability of built-in Excel solvers. To really do it right requires a Linear Programming approach like i-orp to simultaneously solve the entire lifetime at once. I've seen open source LP solvers that can be added to Excel - but that's probably quite a project to get one working.

[RetireBy90]

I will convert into 24% for this year and then see what taxes do for 2022. My plan is to convert until 2025 when tax rates are set to revert and then take SS in 2026. I plan to hit the 3rd bracket for IRMAA one year (I think above $276K) then another year just below that and then back to above $176K bracket. I'll pay a couple years of higher IRMAA but my chart shows that will cost $187/Month/Person or $2,244 a year in higher bracket but save much more if tax rates go up or if I have to take big RMDs and I'm forced into higher bracket for the duration.

I guess my answer is yes, I'm jumping into higher IRMAA bracket for 1, maybe 2 years to avoid being in higher for the duration. Each situation is unique, but then folks here know that.

[SecondCor521]

Quote:

Originally Posted by Exchme Your post mentioned QCDs, but the graphs and math didn’t seem to address them? Charitable giving during your lifetime is of course the difficult thing to value in $, everything else can simply be looked at in the estate value (or the value to heirs after paying taxes on an inherited IRA).

I prefer not to go into it too much.

The graphs don't include QCDs because the graphs are tax brackets, IRMAA brackets, and taxable income. QCDs are excluded from taxable income.

It indirectly shows up in me being able to stay in relatively lower brackets in the 2040s.

I value the QCDs by how much it helps people get things like clean water, food, and medical care, and avoid things like poverty and abuse. Not really a way for me to put that into a dollar equation easily, so I just plugged in what I wanted to do in my spreadsheet and then assume that as a plan constant and adjust the other things around them.

[TheWizard]

You have an interesting "problem".
I retired at 63, not early like you.
And I'm single nowadays also. MFJ folks have additional considerations.

I don't have any experience with ACA, but what I did was try to levelize my AGI (inflation adjusted) from my 60s to my early 70s. I'm 71 now.

My Roth conversion this year will be ruffly the same as my first RMD next year. My conversions have been entirely in the 24% Federal bracket these past few years and this year's will be my last significant conversion.

My IRMAA tier has been the same for the past few years and, with inflation adjustments, might stay the same indefinitely.

But the problem is, even as you're Roth converting, your tax-deferred portfolio can continue to grow during these bull markets.

So I'm not quite sure what to recommend.
I'm sure you'll figure something out...

[Sunset]

Hope the market falls about 95%, convert everything IN KIND , and wait for it all to bounce back.

[RetireBy90]

Quote:

Originally Posted by Sunset Hope the market falls about 95%, convert everything IN KIND , and wait for it all to bounce back.

+1 that would be great from a tax perspective.

[CardsFan]

Quote:

Originally Posted by RetireBy90 +1 that would be great from a tax perspective.

Assuming you survive the heart attack you have watching it drop .

[Out-to-Lunch]

And assuming that you paid the tax bill from money that you had sitting on the sidelines...