Roth Conversions - To Do or Not?

[SurfaceBlow]

This is my first post on Bogleheads but have been reading the forums and blogs for some time. Great information, advice and knowledgeable people. I am not sure if Roth Conversions are appropriate for us after reading and reviewing the forums, articles and spreadsheets from McQuarrie, Mr. Mustache, and Kites. i have read (and re-read trying to understand it all) and tried using the spreadsheets but still not sure if I am interpreting it correctly. We have retirement income in the mid tax brackets with little chance of going in lower brackets. Our background is:

Age 68 and 66.5, both retired for about 2 yrs, and file MFJ.
Started SS this year in spring, about 70K per year. Most likely will always be taxed on 85%.
Both have Pensions, about 72K per year. Her's is COLA.
Have variable income approximately 70K per year, +/- 75%. 70% of gross income is taxable.
Have tIRA and 403b (will be converting soon to tIRA) with total about 1 million.
Very small converted Roth of about 7K. Put all retirement monies into 401K or 403b when working.
CASH about 550K and CD's about 125K. Deciding on relocating so have kept large CASH amount.
Will likely remain in 22% tax bracket but could slide into 24% based on variable income. Low 24% bracket with RMD's.
Medicare typically standard or tier 1, but could enter tier 2 with RMD's and high variable income. Have medicare advantage through previous work.

Have tied using the available spreadsheets indicated above, along with modifying for personal conditions, as I said not sure if I am using or interpreting them correctly. Looked at doing 100K conversions per year for 4 yrs before RMD's start. My results indicate TDA's and post taxed RMD's in taxable accounts will grow larger than if Roth Conversions were done and not enough time to recover from the conversion taxes even in mid 90's or later.

Would like some advice and recommendations on if I am on the right track - on not doing Roth Conversion. Thanks for your responses in advance.

[Carl53]

You could payback your SS and wait until a later date for SS to buy some more space for conversion.

[FiveK]

...not enough time to recover from the conversion taxes even in mid 90's or later.

Time is irrelevant or at best an indirect consideration. What has led you to believe that time is important?

Because you have enough to pay the conversion tax from cash on hand, "tie goes to the Roth" if your conversion marginal tax rate now is the same as it will be in the future. See "Traditional plus taxable" vs. Roth for more on that.

To the extent you expect to
- use qualified charitable contributions and bequeath traditional accounts to charitable organizations, or
- incur large medical expenses not covered by insurance, or
- have your traditional account investments suffer large losses, or
- find that Congress has reduced tax rates in your income range
your future marginal tax rate might decrease and thus small or no Roth conversions could be correct.

To the extent you expect to
- have a surviving spouse subject to single filer brackets and rates, or
- find that Congress has increased tax rates in your MFJ range, or
- have your traditional account investments suffer large gains,
your future marginal tax rate might increase and thus large Roth conversions could be correct.

You mentioned "using the available spreadsheets indicated above." Do you get results anything like the ones in this post, or Worth pushing through the Social Security hump and/or IRMAA cliffs?, or Roth Conversion with Social Security and Medicare IRMAA? How are you interpreting what you see in yours?

[WhiteMaxima]

Delay SS till 70 (a guaranteed 8% return). Take tIRA distribution plus Roth conversion to 12% tax bracket (2026 tax will increase). Now you already started SS, can you undo it?

[Carl53]

Delay SS till 70 (a guaranteed 8% return). Take tIRA distribution plus Roth conversion to 12% tax bracket (2026 tax will increase). Now you already started SS, can you undo it?

Delay SS until 70 is not a guaranteed 8% return but an 8% per year delay increase in benefit (8% increase calculated from FRA benefit and not compounded).

You can either withdraw your application and pay back benefits which I think would provide the greatest increase in conversion opportunities, or suspend your benefit which would allow increased conversion opportunity next year.
https://www.fool.com/retirement/2022/10 ... how-to-us/

[CloseEnough]

If you were working with a clean sheet of paper, i.e., were not already drawing SS, you could do the analysis and perhaps conclude that Roth conversions would more likely result in tax savings. And, that could factor into your decision when to begin SS.

As FiveK above points out, the Roth conversion has many factors and assumptions over many years in the future that might result in tax savings, or could lead you to conclude Roth is not likely to save much. Given the complexity, and the uncertainty, if it were me in this situation, the added hassle of changing your SS benefits would not be worth it (to me) to try to launch a strategy at this point that doesn't have a more definite clear benefit. If you were to conclude that based on your circumstances taking further into account "what you expect" as listed by FiveK, the Roth will very likely result in savings, I might go through the hassle. I just think there are so many variables and unknowns over such a long future time period that would lead me to not go down the Roth conversion road - good luck with your decision.

[HeavyChevy]

This is my first post on Bogleheads but have been reading the forums and blogs for some time. Great information, advice and knowledgeable people. I am not sure if Roth Conversions are appropriate for us after reading and reviewing the forums, articles and spreadsheets from McQuarrie, Mr. Mustache, and Kites. i have read (and re-read trying to understand it all) and tried using the spreadsheets but still not sure if I am interpreting it correctly. We have retirement income in the mid tax brackets with little chance of going in lower brackets. Our background is:

Age 68 and 66.5, both retired for about 2 yrs, and file MFJ.
Started SS this year in spring, about 70K per year. Most likely will always be taxed on 85%.
Both have Pensions, about 72K per year. Her's is COLA.
Have variable income approximately 70K per year, +/- 75%. 70% of gross income is taxable.
Have tIRA and 403b (will be converting soon to tIRA) with total about 1 million.
Very small converted Roth of about 7K. Put all retirement monies into 401K or 403b when working.
CASH about 550K and CD's about 125K. Deciding on relocating so have kept large CASH amount.
Will likely remain in 22% tax bracket but could slide into 24% based on variable income. Low 24% bracket with RMD's.
Medicare typically standard or tier 1, but could enter tier 2 with RMD's and high variable income. Have medicare advantage through previous work.

Have tied using the available spreadsheets indicated above, along with modifying for personal conditions, as I said not sure if I am using or interpreting them correctly. Looked at doing 100K conversions per year for 4 yrs before RMD's start. My results indicate TDA's and post taxed RMD's in taxable accounts will grow larger than if Roth Conversions were done and not enough time to recover from the conversion taxes even in mid 90's or later.

Would like some advice and recommendations on if I am on the right track - on not doing Roth Conversion. Thanks for your responses in advance.

I have found conversions to be of marginal value with high uncertainty for my own situation. (have converted ~10% of total portfolio over the last 3 years)

If I was in your situation, I would not do conversions.

(Burning up your most valuable asset, available already-taxed income, to foster an immediate taxable event (conversion) is commonly promoted on this forum without the caveat that any benefits likely come 30 years later)

[WoodSpinner]

...not enough time to recover from the conversion taxes even in mid 90's or later.

Time is irrelevant or at best an indirect consideration. What has led you to believe that time is important?

Because you have enough to pay the conversion tax from cash on hand, "tie goes to the Roth" if your conversion marginal tax rate now is the same as it will be in the future. See "Traditional plus taxable" vs. Roth for more on that.

To the extent you expect to
- use qualified charitable contributions and bequeath traditional accounts to charitable organizations, or
- incur large medical expenses not covered by insurance, or
- have your traditional account investments suffer large losses, or
- find that Congress has reduced tax rates in your income range
your future marginal tax rate might decrease and thus small or no Roth conversions could be correct.

To the extent you expect to
- have a surviving spouse subject to single filer brackets and rates, or
- find that Congress has increased tax rates in your MFJ range, or
- have your traditional account investments suffer large gains,
your future marginal tax rate might increase and thus large Roth conversions could be correct.

You mentioned "using the available spreadsheets indicated above." Do you get results anything like the ones in this post, or Worth pushing through the Social Security hump and/or IRMAA cliffs?, or Roth Conversion with Social Security and Medicare IRMAA? How are you interpreting what you see in yours?

FiveK,

If you are executing Conversions with the hope of Saving on Taxes then Time is important. You can easily see this by comparing your Tax Burden with NO conversions to own WITH Conversions. If your breakeven TAX wise is past your expected lifespan then it’s unlikely that you will accomplish your goal.

In addition, Time becomes an important factor when looking at the Tax Adjusted Portfolio balances to understand the potential value of Conversions. The longer the interval between Conversion and Valuation the larger the difference in Tax Adjusted Portfolio Balances between No Conversions and Conversions.

Just a different perspective….

WoodSpinner

[Uniballer]

Your situation (pensions + already taking SS) is one of the least likely to result in obvious benefits from Roth conversions, unless your goal is to benefit your heirs. However, depending on your medical status and so on, it is probably worth looking at whether Roth conversions would benefit one of you if your spouse passes away, and whether that is a big enough reason to start conversions now.

If you have already run the numbers, and it looks like you are better off not converting, then that is probably the case. It might be worth running the numbers again in the future if anything changes (tax rate changes, passing of your spouse, SS or pension changes, etc).

[FiveK]

...not enough time to recover from the conversion taxes even in mid 90's or later.

Time is irrelevant or at best an indirect consideration. What has led you to believe that time is important?

FiveK,

If you are executing Conversions with the hope of Saving on Taxes then Time is important. You can easily see this by comparing your Tax Burden with NO conversions to own WITH Conversions. If your breakeven TAX wise is past your expected lifespan then it’s unlikely that you will accomplish your goal.

In addition, Time becomes an important factor when looking at the Tax Adjusted Portfolio balances to understand the potential value of Conversions. The longer the interval between Conversion and Valuation the larger the difference in Tax Adjusted Portfolio Balances between No Conversions and Conversions.

Just a different perspective….

WoodSpinner

"Saving on Taxes" for most is not nearly as important as "having more spendable after-tax money." Those are identical if one looks only at a single year, but for different years the Calculations for what is best* rely on tax rates, not tax amounts.

*Assuming that "best" means having more money to spend, regardless of how much tax has been paid.

It will be interesting to see what the OP has seen/heard on this topic.

[SurfaceBlow]

Thank you everyone for your responses, suggestions, and recommendations. It appears I have so reading and homework to do, along with some more spreadsheet work, which will require a few days before responding back. But it is great to have an opportunity to discuss with others.

Some quick responses to some comments:
* Decided to take SS a couple yrs early because "Open Social Security" indicated only a small increase at 70 over taking our SS and spousal increase.
* Doing a do-over with SS may not help in Roth Conversion either because of our variable income. This year without SS would still be in mid 22% range and with SS (partial year) would be in the high 22%. Even with a full SS year a high 22% or low 24% tax bracket may occur. Biggest issue would be possible higher IRMAA charges and looking into that.
* Our Medicare Advantage is through my wife's previous employer and so far it has been the same or better than our employer sponsored healthcare.
* Concerning the spreadsheets and my attempts at modifying them include: 1) Calculating the actual federal taxes each year based on tax tables instead of assuming the marginal tax rate. Using the inflation rate to increase the tax tables each year and have gone with estimating the increased tax rates tables for 2026. 2) Tax conversion done over 4 yrs to reduce the taxes incurred, trying to stay in lower tax bracket. 3) Instead of just the $100K conversion used in the available spreadsheets, with no previous TDA or taxable account, I have included my present values into the spreadsheet to see how TDA reductions/gains plus taxable accounts gains may compare with and without tax conversions. 4) Still maintaining similar conditions (inflation, interest/dividends, etc.) for both non and with tax conversions. 5) Will try to do to separate spreadsheets, one without conversion and with conversion. Trying to do all in one spreadsheet gets confusing and possibly more error prone.

Forgot to mention: Saving on taxes would be great, but main interest is to make sure monies are available through lifetime but mostly to provide a good portfolio for children and heirs.

[FiveK]

...1) Calculating the actual federal taxes each year based on tax tables instead of assuming the marginal tax rate.

If using the personal finance toolbox spreadsheet, cell Calculations!R80 controls whether the chart below and to the left of it shows "smooth" or "hyper-accurate" marginal rates.

Using the inflation rate to increase the tax tables each year and have gone with estimating the increased tax rates tables for 2026.

Using real (instead of nominal) returns for investment growth allows use of current tax brackets. For what will happen in 2026, your guess is as good as anyone's.

2) Tax conversion done over 4 yrs to reduce the taxes incurred, trying to stay in lower tax bracket.

Seems reasonable.

3) Instead of just the $100K conversion used in the available spreadsheets, with no previous TDA or taxable account, I have included my present values into the spreadsheet to see how TDA reductions/gains plus taxable accounts gains may compare with and without tax conversions.

Maybe we aren't thinking of the same spreadsheet(s)? Not aware of a limitation to use only a $100K conversion amount...?

[Jack FFR1846]

IRMAA is the big set of cliffs you are aiming to stay away from. This means total income needs to be under $194k this year. Going a dollar over means $64 more for part B and I have no idea on advantage plans, but part D is $12 more. The next rung of IRMAA is $250k and going over, your part B doubles to $328 and part D to thirty something.

Income includes social security, interest, dividends on taxable, pensions. I'll tell you what we did. I retired 2 1/2 months ago and used the opensocialsecurity calculator to determine that the most SS we could collect would be if I took it at 70 and DW took it at 62, which coincided with my retirement. We have since understood IRMAA and put in to cancel her social security (they've taken a couple months so far) so we'll pay her payments all back. Why? Because we, like you have a lot in traditional IRAs. For us, to drain those, we'd have to convert over $300k a year, which ain't happening because of IRMAA. So I add up:
interest
dividends on taxable
1099
W2 earnings

You would add pensions and social security. If you stop social security, your income goes down and perhaps you can convert. Good luck. At RMD time, you'll be paying big time.

[Wiggums]

Still not sure why anyone converts traditional iras . Pay the tax when you withdraw to spend the money instead of when the conversion is made.

My brother is single and in the defer forever camp.

We retired early. Plenty of time to convert some before medicare, taking SS or RMDs. We convert per our estate plan to avoid NIIT tax, higher Medicare cost, and higher social security taxed to name just a few considerations. When one of is passes, the surviving spouse will file as single pushing then into a higher tax bracket

[FiveK]

Still not sure why anyone converts traditional iras . Pay the tax when you withdraw to spend the money instead of when the conversion is made.

Paying 12% now instead of 22.2% later would be one reason.

[privateID]

Using the inflation rate to increase the tax tables each year and have gone with estimating the increased tax rates tables for 2026.

Using real (instead of nominal) returns for investment growth allows use of current tax brackets. For what will happen in 2026, your guess is as good as anyone's.

My issue with doing it this way is that since the formula for SS taxation is not indexed for inflation, it would need to be discounted every year to be accurate. I do it both ways in my home-grown spread sheet.

[FiveK]

Using the inflation rate to increase the tax tables each year and have gone with estimating the increased tax rates tables for 2026.

Using real (instead of nominal) returns for investment growth allows use of current tax brackets. For what will happen in 2026, your guess is as good as anyone's.

My issue with doing it this way is that since the formula for SS taxation is not indexed for inflation, it would need to be discounted every year to be accurate. I do it both ways in my home-grown spread sheet.

Absolutely correct on the SS taxation and other non-inflated tax law numbers (e.g., NIIT threshold). The error introduced by not accounting for those, however, tends to be small in comparison with error band around other assumptions (e.g., investment returns, other tax law changes, etc.).

[privateID]

Using the inflation rate to increase the tax tables each year and have gone with estimating the increased tax rates tables for 2026.

Using real (instead of nominal) returns for investment growth allows use of current tax brackets. For what will happen in 2026, your guess is as good as anyone's.

My issue with doing it this way is that since the formula for SS taxation is not indexed for inflation, it would need to be discounted every year to be accurate. I do it both ways in my home-grown spread sheet.

Absolutely correct on the SS taxation and other non-inflated tax law numbers (e.g., NIIT threshold). The error introduced by not accounting for those, however, tends to be small in comparison with error band around other assumptions (e.g., investment returns, other tax law changes, etc.).

In my spreadsheet, in about 20 years, I show it being off by approx an absolute 20% (ie., instead of 70% of my SS taxed, closer to 50%). Doesn't yield a huge different ending value, but enough to make me think at 70% not to worry so much about the humps because any SS changes will probably bump me to the 85% max (or maybe even blast past the hump otherwise), but at 50% maybe not. Just some food for thought. I like doing it both ways because of issues like this.

[LadyGeek]

I removed a contentious interchange regarding Medicare Advantage. The discussion was derailed.

Please stay on-topic, which is Roth conversions.

[986racer]

I've done a very similar analysis for my parents a few years back and they were in a similar situation as you... Already claiming SS and about to start the RMD process..

The TLDR version is that it is still beneficial to do the Roth conversions

The long answer is that there are a lot of assumptions that goes into any of this analysis and so you need to have a spreadsheet and a way to model a bunch of different scenarios. Here is stuff that you should be modelling
1. Death of spouse. The sooner one spouse passes, the sooner you move into the "filing single" brackets.
2. Growth of various accounts
3. Reasons to need money in T-IRA accounts (e.g., get deductions to pay for medical expenses, QCD, etc)
4. Change in tax rates

What I found is that when I did the analysis with my parents at age 70, the projections showed that they were obviously ahead at around age 95. In between those years, it becomes less and less obvious. My definition of "obviously ahead" was that the projections showed the same net worth between doing the two strategies (doing conversions vs not) but the strategy of doing conversions had all of the money in Roth and therefore no more taxes were being paid.

Basically, if you look at a chart of how the strategies play out, you see the conversion strategy has a lower net worth in the beginning (due to making tax payments now and IRMAA costs) but then makes it up when larger RMDs start hitting the non-conversion strategy, and even more so when a spouse passes.

[FiveK]

Using the inflation rate to increase the tax tables each year and have gone with estimating the increased tax rates tables for 2026.

Using real (instead of nominal) returns for investment growth allows use of current tax brackets. For what will happen in 2026, your guess is as good as anyone's.

My issue with doing it this way is that since the formula for SS taxation is not indexed for inflation, it would need to be discounted every year to be accurate. I do it both ways in my home-grown spread sheet.

Absolutely correct on the SS taxation and other non-inflated tax law numbers (e.g., NIIT threshold). The error introduced by not accounting for those, however, tends to be small in comparison with error band around other assumptions (e.g., investment returns, other tax law changes, etc.).

In my spreadsheet, in about 20 years, I show it being off by approx an absolute 20% (ie., instead of 70% of my SS taxed, closer to 50%). Doesn't yield a huge different ending value, but enough to make me think at 70% not to worry so much about the humps because any SS changes will probably bump me to the 85% max (or maybe even blast past the hump otherwise), but at 50% maybe not. Just some food for thought. I like doing it both ways because of issues like this.

As the saying goes, All models are wrong, but some are useful. One might guess that SS benefit taxation 20 years from now will be quite different than today, but who knows what (if any) difference there will be? Neither of the "both ways" seem "wrong enough" to avoid.

[tibbitts]

Basically, if you look at a chart of how the strategies play out, you see the conversion strategy has a lower net worth in the beginning (due to making tax payments now and IRMAA costs) but then makes it up when larger RMDs start hitting the non-conversion strategy, and even more so when a spouse passes.

But you have to adjust the current net worth for the amount of taxes you anticipate (guess) you or someone will have to pay eventally, so... isn't it close to a wash, depending of course on your assumptions?

[986racer]

Basically, if you look at a chart of how the strategies play out, you see the conversion strategy has a lower net worth in the beginning (due to making tax payments now and IRMAA costs) but then makes it up when larger RMDs start hitting the non-conversion strategy, and even more so when a spouse passes.

But you have to adjust the current net worth for the amount of taxes you anticipate (guess) you or someone will have to pay eventally, so... isn't it close to a wash, depending of course on your assumptions?

At the end, the Roth conversion strategy is clearly better because it has the same net worth but all the money had been taxed already. In the middle, it's very difficult to determine which strategy is really ahead because of the adjustment you would need to make. It was difficult determining what would be the correct adjustment.