Tax on Roth Conversions vs Tax on RMDs

[pb4uski]

don't forget as time goes on the brackets expand to allow more and more income through at lower tax rates. pretty soon 100k will be 15%.

Yes, I know... built into my model... in 2024 according to my projections.

 

[kramer]

In addition to what others have said, you will need to consider State Income Taxes and MAGI as it relates to ObamaCare subsidies.

It is unclear what other income you have after you retire early. But if it is mainly qualified dividends and capital gains from your taxable account, then you have lots of room in the lowest brackets to convert yearly.

For instance, your personal exemptions plus standard deduction amount to over $20,000 income tax free. Plus another $18,000 in the 10% tax bracket. That is $38,000 per year right there.

I am early retired in my late 40's with a large taxable account. I convert up to the top of the 10% bracket each year. The net taxes I pay on the conversion are much less than 10% because I fill up the unused portion of the 0% bracket and I also get to use my foreign tax credit which would otherwise go to complete waste since my taxes would be zero without the conversion.

I then fill up to the top of the 15% bracket with capital gains harvesting.

I don't have a pension coming. So I don't want my IRA to go to zero before RMD (since I expect all or most of the RMD to be taxed at 0% or 10%) although I am trying to reduce the size before age 70.

As I get closer to age 70 and have a better idea on what the tax laws will be then as well as my wealth level, I know I will have time to make adjustments by converting up into the 15% bracket. But I am not going to do that until I am pretty sure that I will need to.

 

[medelste]

Here's a spreadsheet I developed to optimize spending, including taxes, SS, Roth conversions and everything else I could think of. It uses iteration to optimize spending, has minimal instructions, and who knows what numbers I left in it (though they're not mine)...

Very impressive and helpful, thank you. If that's your "Simple" version, I'd hate to see your "Complex" version!

Sent from my HTC6525LVW using Early Retirement Forum mobile app

 

[sengsational]

Here's a spreadsheet I developed to optimize spending, including taxes, SS, Roth conversions and everything else I could think of. It uses iteration to optimize spending, has minimal instructions, and who knows what numbers I left in it (though they're not mine)...

I was excited to try this out, but my Excel (2002) didn't know how to do the YEARFRAC function. I started this post to ask what to do about it, but discovered the problem before posting! If you add the Analysis ToolPak (Tools > Add-ins) it works! Off to the races!

 

[Animorph]

Very impressive and helpful, thank you. If that's your "Simple" version, I'd hate to see your "Complex" version!

Sent from my HTC6525LVW using Early Retirement Forum mobile app

Glad you found it helpful. My complex version is a C language program that I wrote. It handles all my details and optimizations, but would not be as generally applicable, nor at all user friendly.

 

[bingybear]

Glad you found it helpful. My complex version is a C language program that I wrote. It handles all my details and optimizations, but would not be as generally applicable, nor at all user friendly.

I need to spend more time looking at the underlying code... and read your notes. I was shocked when a big red box popped up and said UR F@*#ed ....
I think the retirement time bomb is alive and well.

It looks like a good useful spreadsheet... just bad data. thanks

 

[sengsational]

Glad you found it helpful. My complex version is a C language program that I wrote. It handles all my details and optimizations, but would not be as generally applicable, nor at all user friendly.

I've got the model running pretty well, I think. Awesome piece of work, I must say. Thank you very much for sharing it. I keep plugging my numbers into different models. Sometimes I have a good feeling about a model, sometimes not. I feel like this model has got it right.

I think I can see what the fractional bracket fill thing does, but I'm not sure of the meaning (260% of what). When I change it, I can see that during the years where I have Roth funds, it pulls only up to a certain amount out of the tIRA and gets the rest from the Roth. That works to force filling up the low tax brackets, but not to the top of the 15%, which might be too much for ACA cliff. So I like 260%, I think, but I'm not "getting it" with respect to the table at S1...W12.

I put my ACA credits in as "Income non-taxable". And I put the value of my house now (today's dollars) also in non-taxable in the year which I plan to sell it. Is that right? I see that various things are not indexed for inflation, so I presume we don't inflate today's house value out X years to determine the 'windfall' cash that the house sale would bring.

The model is giving me a bit higher annual spending after tax than i-orp, given the same inputs. A some point, I might try to find a way to discover where the differences are, just for grins, but it's close enough. After all, it's only got to be "roughly good" for year number 2 through 'n'. That gives a framework for current year spending and tax strategies...the only time I can actually DO something is this year. Next year, I'll be running another, (probably updated and better fine-tuned) model, with whatever the latest tax and ACA laws are then.

 

[Animorph]

I've got the model running pretty well, I think. Awesome piece of work, I must say. Thank you very much for sharing it. I keep plugging my numbers into different models. Sometimes I have a good feeling about a model, sometimes not. I feel like this model has got it right.

I think I can see what the fractional bracket fill thing does, but I'm not sure of the meaning (260% of what). When I change it, I can see that during the years where I have Roth funds, it pulls only up to a certain amount out of the tIRA and gets the rest from the Roth. That works to force filling up the low tax brackets, but not to the top of the 15%, which might be too much for ACA cliff. So I like 260%, I think, but I'm not "getting it" with respect to the table at S1...W12.

I put my ACA credits in as "Income non-taxable". And I put the value of my house now (today's dollars) also in non-taxable in the year which I plan to sell it. Is that right? I see that various things are not indexed for inflation, so I presume we don't inflate today's house value out X years to determine the 'windfall' cash that the house sale would bring.

The model is giving me a bit higher annual spending after tax than i-orp, given the same inputs. A some point, I might try to find a way to discover where the differences are, just for grins, but it's close enough. After all, it's only got to be "roughly good" for year number 2 through 'n'. That gives a framework for current year spending and tax strategies...the only time I can actually DO something is this year. Next year, I'll be running another, (probably updated and better fine-tuned) model, with whatever the latest tax and ACA laws are then.

It's been a while...

The 260% means something like tax bracket #2 (the 100's digit) plus 60%. The number isn't particularly meaningful but gives a single number input that can be optimized. The spreadsheet will make sure each year's income fills the tax bracket to that level by making additional tIRA withdrawals if necessary.

Yes, no inflation, so everything stays in today's dollars.

I'm happy to be close to i-orp. There are many ways to optimize.

 

[Terryjm51]

Lol. I agree wow nice spread sheet. If that is the easy one I would hate to see the hard one

This info is a little over me. Sounds like some reading on my part.

Thanks all


Sent from my iPhone using Early Retirement Forum

 

[fishboat]

same issue..same concerns

I still need to runs the scenarios, but I have the same issues & concerns. The path of least resistance (for me) appears to be the ESPlanner software.

 

[sengsational]

It's been a while...

I know...sometimes when I come back to one of my complex spreadsheets after a long while, I wonder who was smart enough to come-up with all of this stuff...couldn't have been me

Thank for the explanation of the fill value...that's logical. The spend is 6% higher, and I'm not quite sure the inputs are perfectly aligned yet. But considering the variability of markets, this level of agreement is plenty good.

 

[Sunset]

this spreadsheet looks interesting, unfortunately I'm having an issue.
I'm using openOffice calc and it gives an error in many of the cells:

When I first open it, it has #VALUE! for many of the cells.
I set the reset cell (AF9) to 0 and there are no errors and I see numbers all over.
When I put the reset back to 1 to calculate, it then gives me errors in many cells and its: Err:523

I googled this and it supposedly is an iteration thing, I go to change the iteration setting from 1,000 to 5,100 (since the sheet has notes it does 5,001 iterations). But calc will not make the new value stick, it resets itself to 1,000 which is weird.

I'll try this on libreOffice on the off chance its ok, but wondered if anyone had other ideas ?

 

[ERD50]

this spreadsheet looks interesting, unfortunately I'm having an issue.
I'm using openOffice calc and it gives an error in many of the cells:

When I first open it, it has #VALUE! for many of the cells.
I set the reset cell (AF9) to 0 and there are no errors and I see numbers all over.
When I put the reset back to 1 to calculate, it then gives me errors in many cells and its: Err:523

I googled this and it supposedly is an iteration thing, I go to change the iteration setting from 1,000 to 5,100 (since the sheet has notes it does 5,001 iterations). But calc will not make the new value stick, it resets itself to 1,000 which is weird.

I'll try this on libreOffice on the off chance its ok, but wondered if anyone had other ideas ?

I just tried in LibreOffice, and seems the same as you report. I have not looked any further yet. Looks like 1,000 is max iterations it will accept.

-ERD50

 

[bingybear]

I just tried in LibreOffice, and seems the same as you report. I have not looked any further yet. Looks like 1,000 is max iterations it will accept.

-ERD50

I have excel... but brought up libreoffice (basically openoffice)... complaint is wrong data type

 

[bingybear]

I tried looking a what was being complained about in the formula... copied bits of it to another cell.... working through it, the calc went OK. So playing with it I found that making some changes (adding a logical calc to other cell removed some. Finally, I loaded the spreadsheet and turned autocalc off and back on. It looked like this cleared the #VALUE error. Give that a try ... this was with libreoffice

 

[ImThinkin2019]

I tried all sorts of things including the Simple Retirement Plan spreadsheet, which I want to thank Animorph for. Because I wanted pretty accurate tax models and SS payment models I ended up buying ESPlanner. It calculates taxes and SS payments for you.

If you do use this program, you will find that the calculated living standard changes over time as your Roth rollovers change. So in order to clearly understand the effect of the Roth rollover value, I built a simple spreadsheet to put in the stream of payments and reduce everything to NPV. The combination of this spreadsheet plus ESPlanner reports helps me to understand what I need to do to minimize taxation and maximize living standard/$$ left to heirs.

If you wanted to see what my models look like I would be happy to have a skype sharing session with you. This is easier than reading or trying to learn all this stuff yourself.

 

[Animorph]

For those with a bunch of errors, you might try turning off the optimizations (T2 and BL12 set to zero). That will at least significantly reduce the number of iterations required. Turn BL12 back on first, let it run. And then try T2, which should really grind away.

BL12 optimizes the yearly spending to reach the final net worth target. T2 controls the tax bracket fills.

 

[Terryjm51]

When I check tax caster it calc's my return around $5500 a year from the government ? I will pull $65k from 401k and pension 20k from social sec
Can this be right? Filling married
Doesn't seem right after all the years I have paid in


Sent from my iPhone using Early Retirement Forum

 

[sengsational]

I tried all sorts of things including the Simple Retirement Plan spreadsheet, which I want to thank Animorph for. Because I wanted pretty accurate tax models and SS payment models I ended up buying ESPlanner. It calculates taxes and SS payments for you.

If the tax calculation of ESPlanner included PPACA, I'd probably find it more useful. My tax rate for 2014 was -111%, but if I hit the cliff, I jump into positive territory (where I'd need to pay instead of receive).

 

[Happily Retired]

Ditto pb4uski coments and strategy

Well, it gets real complicated real quick. Essentially you need to construct a long term projection and then also include an annual tax calculation and then include Roth conversions to the top of your desired tax bracket (15% in my case) and see how things end up after pensions, SS and RMDs start. While each situation is different, my modeling suggests that if I do Roth conversions to the top of the 15% tax bracket from now (59) until I'm 70 that I only spend one year in the 25% tax bracket vs a decade, and my age 100 NW is 20% higher.

Note to pb4uski: Our (DW and myself) situation is identical to what you're describing.

To all: This is not complaining about paying taxes when taking tIRA distributions. Using a tax efficient strategy like this helps to make it as pleasant as possible.