Optimizing Retirement Income...

Midpack

Give me a museum and I'll fill it. (Picasso) Give me a forum ...
Joined
Jan 21, 2008
Messages
21,318
Location
NC
...by considering taxable, IRA RMDs & when to take Soc Sec and legally minimizing taxes throughout retirement.

Have we discussed this topic already that I missed? The Taxation Of Social Security Benefits As A Marginal Tax Rate Increase? | Kitces.com

About 1/3rd of our portfolio is in TIRAs, almost completely non-deductible, and none in Roth IRAs (we weren't eligible).

With the end of the year approaching and RMD 11-13 years away, I am still wrestling with whether or not Roth conversion makes sense for us. The online calculators don't factor in taxes, or specifically how Soc Sec is taxed vs AGI.

I spent days building a spreadsheet to look at different scenarios in taking income from taxable, TIRA vs Roth Conversion and withdrawals from either, and when to take Soc Sec. To keep it manageable I assumed conservative but linear real returns, Federal taxes only, and tax rates staying the same (inflation adjusted) - even though none of that will be true. The deeper I got, the more difficult it became, it's mind-boggling for a simpleton like me at least and I'm usually reasonably capable with spreadsheet analysis. [edit: calculating taxable income is the most difficult aspect to me from taxable, 4 TIRAs, dividends, cap gains, Soc Sec, etc. and then applying the 50%/85% Soc Sec thresholds on top of it all.]

I found a spreadsheet from a Boglehead member, but it's pretty complex too, and without examining every assumption he made, I can't draw conclusions.

I even considered going to an FA, but minimizing taxes over a lifetime doesn't seem to be a focus area for any FA's I've found. They can help with taxes short term, but seemingly not long term.

Taking Soc Sec at 70 is best for us irrespective of other income sources (how most people seem to evaluate when to take Soc Sec), but RMDs call that into question for me. It appears RMDs will force us to withdraw more than we expect to need and Soc Sec on top of that will be heavily taxed - so we may be money ahead after taxes by taking Soc Sec earlier. And the fact that we're withdrawing less than SWR would suggest now, only makes our tax picture 15-20 years from now look worse. Yes, I know it's a nice problem, but I'd still like to optimize for beneficiaries. And yes I know I can't hit any of this exactly.

I have my work cut out for me in Dec. However, I'm starting to think with all the variables/unknowns the best course may simply come down to:

  • if I think tax rates will increase - take more now from IRAs to bring down future RMD income and delay Soc Sec to 70
  • if I think tax rates will stay the same (inflation adjusted) - withdraw from all sources somewhat equally such that taxable income stays the same for the long haul which may suggest taking Soc Sec earlier.
  • if I think tax rates will decrease - I simply don't.
Hopefully I'm just making this more complicated than need be...:banghead::blush::facepalm::crazy:

If you made it this far, thanks for your patience.
 
Last edited:
My spreadsheet models the tax bracket boundaries and the tax rates within them.

I figured out how they inflate the bracket boundaries along with the rounding rules and use that along with an assumed future CPI to estimate future taxes. I also model the other tax parameters' inflation and rounding.

Tax brackets are inflated using the ratio of (the average of the monthly CPI (CUUR0000SA0 U.S City Average All Items Base year 1982-1984=100) values from Sept of last year to the Aug of this year) to the same value calculated lat year. It is rounded to the nearest $50. Tax parameter rounding varies depending on what it is.

I have it set up so that the rates within a bracket can change and the number of brackets can change (and did).

I use this scheme to estimate and compare future taxes. PM if you want the code in Excel format.
 
Brett_C - wouldn't it be easier to keep everything in today's dollars and use real returns for any assumed gains?

I need to take a closer look, I think I took a shotgun approach at this a while back, and RMDs did make for a significant tax hit.

I don't want to hit the thresholds that might cause me to lose tuition deductions for my kid, so one more year and I'll start taking ROTH conversions seriously again.

-ERD50
 
Brett_C - wouldn't it be easier to keep everything in today's dollars and use real returns for any assumed gains?

I need to take a closer look, I think I took a shotgun approach at this a while back, and RMDs did make for a significant tax hit.

I don't want to hit the thresholds that might cause me to lose tuition deductions for my kid, so one more year and I'll start taking ROTH conversions seriously again.

-ERD50
Yes, but I am modelling expected expenses&income&taxes some of which have start and stop dates. I chose to model in future dollars across the board because of the expected varying inflation rates of various categories of expenses, otherwise modelling in today's dollars would make everything simpler.:flowers:
 
This paper might be of interest as it address when to take SS along with taxes on RMDs -

http://research.prudential.com/documents/rp/InnovativeSocialSecurityNov2012.pdf

We aren't doing any Roth conversions for now because we want to keep our MAGI low for college financial aid and tax credits as well as ACA subsidies. We will re-evaluate what to do again in a few years.

Corrected RMDs to Roth conversions.
 
Last edited:
I've probably spent most of my retirement working on this problem. It's not easy if everything is in play.

If you can identify the federal tax bracket that you will be exceeding during RMD's you can then try to Roth convert whenever you get a chance to within that bracket. Until it looks like RMD's will come out within that bracket. This might be determined with a fairly rough calculation of your taxes at age 70. Then if you get the chance to take extra tIRA withdrawals/Roth conversions in a lower bracket before age 70, jump on it.

It's not a big deal if you don't get it exactly perfect. But if you are well into the 25% tax bracket with RMD's at age 70 and have no taxable income now, it can make a big difference.

Any time you need to make a taxable withdrawal from a tIRA and have an equivalent amount in your taxable account, make it a Roth conversion.
 
.....Taking Soc Sec at 70 is best for us irrespective of other income sources (how most people seem to evaluate when to take Soc Sec), but RMDs call that into question for me. It appears RMDs will force us to withdraw more than we expect to need and Soc Sec on top of that will be heavily taxed - so we may be money ahead after taxes by taking Soc Sec earlier. And the fact that we're withdrawing less than SWR would suggest now, only makes our tax picture 15-20 years from now look worse. Yes, I know it's a nice problem, but I'd still like to optimize for beneficiaries. And yes I know I can't hit any of this exactly.....

But wouldn't delaying SS until 70 result in lower RMDs (compared to starting SS earlier) since you would have less income from 62 to 70 and therefore more room to do Roth conversions that will ultimately reduce your RMDs? My plan is to defer my small pension and SS as long as possible to do as much 0% capital gains/Roth conversions as possible.

Your SWR and what you need to live on shouldn't constrain your withdrawals from tax-deferred accounts since there isn't anything that requires you to spend your tax-deferred withdrawals (you can chose to save them in taxable accounts if they exceed what you want to spend).

The additional dilemma that I have is whether I should prioritize 0% capital gains over Roth conversions. I have ~2-3 year of living expenses in unrealized capital gain in taxable accounts and am finding those 0% capital gains hard to pass up but I'm not totally sure that is the best play and whether in the long run I might be better off prioritizing Roth conversions over 0% capital gains.
 
This paper might be of interest as it address when to take SS along with taxes on RMDs -

http://research.prudential.com/documents/rp/InnovativeSocialSecurityNov2012.pdf

We aren't doing any RMDs for now because we want to keep our MAGI low for college financial aid and tax credits as well as ACA subsidies. We will re-evaluate what to do again in a few years.
i just skimmed it, but mistake #4 its exactly what I am concerned about. I'll be reading it carefully this afternoon. Thank you!

But wouldn't delaying SS until 70 result in lower RMDs (compared to starting SS earlier) since you would have less income from 62 to 70 and therefore more room to do Roth conversions that will ultimately reduce your RMDs?
Yes they will, I am trying to optimize income/minimize legal taxes.
pb4uski said:
The additional dilemma that I have is whether I should prioritize 0% capital gains over Roth conversions. I have ~2-3 year of living expenses in unrealized capital gain in taxable accounts and am finding those 0% capital gains hard to pass up but I'm not totally sure that is the best play and whether in the long run I might be better off prioritizing Roth conversions over 0% capital gains.
Two of many complexities I ran into trying to develop my spreadsheet, dealing with annual dividends and STCGs plus working in LTCGs along the way AND how long to plan on the 0% div/CG rule to last. The deeper I go, the more complicated it becomes...
 
Last edited:
We aren't doing any RMDs for now because we want to keep our MAGI low for college financial aid and tax credits as well as ACA subsidies. We will re-evaluate what to do again in a few years.
How are you avoiding RMDs? The "R" means . . . are you paying the fine?

It appears RMDs will force us to withdraw more than we expect to need and Soc Sec on top of that will be heavily taxed - so we may be money ahead after taxes by taking Soc Sec earlier. And the fact that we're withdrawing less than SWR would suggest now, only makes our tax picture 15-20 years from now look worse. Yes, I know it's a nice problem, but I'd still like to optimize for beneficiaries.
Sorry, but another variable: It's worth looking at the "what if one of us dies and the other is filing Single" case. It can be a real eye-opener, especially if living costs for the survivor aren't expected to go down very much. This factor can prompt additional Roth conversions even if it might not be optimum for the MFJ situation.

If leaving a large inheritance is important then the situation might be different, but for us I try to keep in mind that our "pot" has to last us a long time that might include a lot of turbulence. Retaining some cushion there now (by paying lower taxes early on) might be worth it even if it leads to a larger tax bill later. By that time the finish line will be considerably closer. If we've ridden one of those "high and right" FIRECalc lines and are rolling in dough when we're 85, high taxes will be okay. If instead our FIRECalc line hugged the X-axis, then we won't be paying high taxes after all and will be very glad for every dime we didn't pay in taxes early on. We'll use that saved money to buy Alpo.
 
Excellent samclem, another reason I gave up on the spreadsheet after a few days.

I wonder if someone has ever modeled all this? I would think so, but I haven't found it yet. I'd pay (within reason) for a complete analysis but I can't find anyone who's begun to tackle the whole picture. At this point, I'll probably just...
 
Given that future changes to tax laws create a fog I cannot see through, IMO it's pointless considering the conversion issue down the penny. A broad brush of tax rate now vs. best guess of tax rate later is the main factor I consider. Since you mentioned beneficiaries, you might also consider estate taxes. Converting to Roth reduces the size of your taxable estate, which can reduce death taxes.
 
How are you avoiding RMDs? The "R" means . . . are you paying the fine?.

Sorry, I corrected it in the original post. That was a typo. I meant to say 401K to Roth conversions now, not RMDs.

I have taking SS early and taxes on RMDs in the retirement spreadsheet now, but when we get closer to SS age I suspect we will delay and do Roth conversions in order to avoid paying income taxes on the RMDs down the line. The trade off is lowering our portfolio in the age 62 to 70 years.

So we can maximize for portfolio balance in the pre age 70 years or minimize taxes post 70, but it is unlikely we will be able to do both.

I have modeled it in a spreadsheet. It is a trade off for us. SS early, RMDs later smooths out our net worth over the retirement years. SS later increases our net worth in later years but the net worth dips in before age 70.

If we both die before age 70, it reduces the amount of money in our estate. We liked the smoothed net worth when we did the modeling so for now we are going with that as a plan.
 
Last edited:
Sorry, but another variable: It's worth looking at the "what if one of us dies and the other is filing Single" case. It can be a real eye-opener, especially if living costs for the survivor aren't expected to go down very much. This factor can prompt additional Roth conversions even if it might not be optimum for the MFJ situation.

This will be a factor for me to go ahead and convert as much of my traditional IRA to a Roth, as a hedge for the above. While married, we'll not likely exceed the 25% tax bracket and may be able to stay in the 15% bracket some years before RMDs. As singles, we'd be in the 28% bracket. Under this circumstance, I am willing to convert to a Roth to fill up the 15% bracket and possibly the 25% bracket while married, to avoid the 28% bracket later. This is more for my wife to ensure she has more income available. If I knew I'd live longer than her, I'd do a more detailed analysis to see what would be more optimal.
 
If most of your TIRA is non-deductible (after tax), why wouldn't conversion be pretty easy to justify?
 
I am using a spreadsheet to model all this, but not for tax calculations. TurboTax on-line TaxCaster does the tax estimation: https://turbotax.intuit.com/tax-tools/calculators/taxcaster/

I have two spreadsheets open, one that shows the big picture of annual numbers until age 100 and one as a scratch-pad for looking at the effect of marginal rates each year, and TaxCaster on the browser and alt-tab between them. It goes pretty fast. Of course, everything is calculated with today's tax rates and laws. I am not trying to predict the future, just compare strategies.

If you want an eye-opener, assume single status (widow or widower) at 71, use a reasonable annual SS payout, calculate the RMD for that year and make a little table where you take out more from the TIRA in $10,000 increments, calculate the tax rate for each using TaxCaster, then calculate the incremental tax rate on the incremental income. I show something like a 45% marginal tax rate :eek: for the lower withdrawals. The marginal rate comes down at higher withdrawal rates, but the average tax paid remains high. In our case, it looks like the best strategy is to convert everything to a Roth before 70.5.

Other conclusions are obvious:
1) Try to spend as little as possible. 3.5% annual actual portfolio depletion rate before taxes is a really good place to start.
2) Things are tighter with a 6% ROI (realistic, IMHO) than with a 9.5% ROI (Wellesley for the past 26 years).

And all this is without taking into account the hit in 2035 (?) when SS can only cover 75% of its obligations, and of course major health event expenses.

It looks like a lot of beans and rice in our future.:(
 
I found a spreadsheet from a Boglehead member, but it's pretty complex too, and without examining every assumption he made, I can't draw conclusions.

Is this the spreadsheet posted by BigFoot48? He supposedly has created a spreadsheet modelling taxes considering the impact of RMD's and SS. I've not looked at it yet because it won't be a question for me until 2015. I too am delaying SS until 70 which will impact tax levels. I do have a ROTH and 2/3 of PF is after tax, but I'm still going to look at the conversion to maximize my overall tax situation.
 
The marginal rate comes down at higher withdrawal rates, but the average tax paid remains high.
This is a significant point. Larger withdrawals beyond MRDs are taxed at 25% to 28% marginal rate over a wide range and the average rate doesn't change much.
 
But wouldn't delaying SS until 70 result in lower RMDs (compared to starting SS earlier) since you would have less income from 62 to 70 and therefore more room to do Roth conversios that will ultimately reduce your RMDs? My plan is to defer my small pension and SS as long as possible to do as much 0% capital gains/Roth conversions as possible.
That plus:

Medical expenses - it could be that after 70 medical expenses become high enough to lower your taxable income.

Charitable donations as part of your RMD. It's expiring after this year, but the ability to make a charitable donation directly from an IRA and yet count as part of the RMD may be reinstated in the future.
 
If leaving a large inheritance is important then the situation might be different, but for us I try to keep in mind that our "pot" has to last us a long time that might include a lot of turbulence. Retaining some cushion there now (by paying lower taxes early on) might be worth it even if it leads to a larger tax bill later. By that time the finish line will be considerably closer. If we've ridden one of those "high and right" FIRECalc lines and are rolling in dough when we're 85, high taxes will be okay. If instead our FIRECalc line hugged the X-axis, then we won't be paying high taxes after all and will be very glad for every dime we didn't pay in taxes early on. We'll use that saved money to buy Alpo.
Yep, when it comes down to paying extra taxes now in order to avoid potential future taxes, I am always leery of taking such a step.

And the much higher future taxes outcome is because investments did very well.

My tax rate now is lower than when I was working and salaried, but it's not clearly lower than it might be in the future. Of course we have only about 10% of our investments in IRAs, so our RMDs shouldn't be a large amount of our income after 70. I've never been motivated to convert.
 
I am using a spreadsheet to model all this, but not for tax calculations. TurboTax on-line TaxCaster does the tax estimation: https://turbotax.intuit.com/tax-tools/calculators/taxcaster/

I have two spreadsheets open, one that shows the big picture of annual numbers until age 100 and one as a scratch-pad for looking at the effect of marginal rates each year, and TaxCaster on the browser and alt-tab between them. It goes pretty fast. Of course, everything is calculated with today's tax rates and laws. I am not trying to predict the future, just compare strategies.

.... In our case, it looks like the best strategy is to convert everything to a Roth before 70.5.

I have taken a similar approach, using TaxCaster to guess-timate taxes annually through age 95. I am 49 now. It is very time consuming and confusing.

From what I have observed from my work and longevity assumptions, it is best to delay SS until 70. But I also have the added complexity of a non-COLA pension available to me at 55, but progressively larger as I delay to age 65. Similar to SS, it is optimal to delay the pension as long as possible.

I also have a significant position in iBonds with maturities beginning at about age 67.

I have significant tIRA and 401(k) resources in addition to taxable.

What all of this adds up to is a tremendous amount of income coming on line beginning at age 70. Some would say it's a nice problem to have. But my federal bracket spikes to 33% at age 70 for pretty much the rest of my life. Of course I could pay a lot more in taxes if I decided to withdraw more.

Again, this is all based on a bunch of assumptions re: longevity, market returns, future tax brackets, and WRs. It is all basically just guessing, and the entire project is what some people might call "false precision".

Most of my scenarios confirm that the most prudent thing to do is Roth conversions throughout my 50s. It probably also makes sense to cash in some of my iBonds early.

Of course, I will need to take these steps at exactly the point when my health insurance premiums spike and I would want to keep my MAGI LOW in order to maximize subsidies! It is shaping up to be a real tug of war. Where is the online calculator that figures all of this out?? :confused:
 
Last edited:
If most of your TIRA is non-deductible (after tax), why wouldn't conversion be pretty easy to justify?


+1

Suppose you have an IRA worth $100k, of which $70k is non-deductible contributions. ROTH convert it at 25% tax bracket and you pay $7,500 in tax leaving you with a ROTH of $92,500.

Assuming tax rate remains the same and in x years time your money has increased 10% and you then cash it out.

$92,500 is now cash in hand of $101,750

$100,000 rises to $110,000, cash it out, pay tax on $40k and you have $100,000 cash in hand.
 
Is this the spreadsheet posted by BigFoot48? He supposedly has created a spreadsheet modelling taxes considering the impact of RMD's and SS.
Yep. He debugged it (13 iterations?) with feedback from the Boglehead community and it's probably great. But I need to pick it apart to understand what assumptions are made before I can have confidence in the results - so far I haven't had the patience.

Where is the online calculator that figures all of this out?? :confused:
That's what I'm saying. Hard to believe no one has. I'd pay for the chance to run a few scenarios through it as long as I could see all the assumptions with the results...

All the simple online calculators like Vanguard among others, tell me it's basically a push for us to convert, which would be a good thing if I could be sure it's true. BUT they don't factor in Soc Sec taxes and when Soc Sec is taken, and the results all appear to hinge almost entirely on the users future tax assumptions (higher taxes=convert, lower taxes=don't, same taxes=toss up). Problem for me is I want to assume same taxes (inflation adjusted) and see Soc Sec tax impact.
 
Last edited:
Does ESPlanner handle these? I don't own it but have looked at it here and there and it seemed pretty comprehensive the last time I looked at it.

What I have done is take my base plan without any Roth conversions in Quicken Lifetime Planner and look at my QLP projected income in the year I turn 70 (based on today's values) and compare it to today's tax table. I include pensions, SS, RMDs and 2.5% of the prior year taxable account balance (for as a proxy for dividends on the taxable portfolio) less deductions less exemptions. Based on that I would be solidly in the 25% bracket at 70.

Then I change the projection to include a negative contribution from IRAs from now to age 70 equal to the Roth conversions I could do today to bring me to the top of the 15% bracket and add an annual special expense for the taxes on that amount (with no inflation). The I recompute my projected taxable income at age 70 based on that scenario. With this Roth conversion adjustment, I am still in the 25% bracket at age 70, but substantially less so.

My RMDs are ~60% of what they would be absent the Roth conversions.

I realize it is a bit crude, but it is a starting point to get a sense for how Roth conversions would change your RMDs and taxes.
 
Does ESPlanner handle these? I don't own it but have looked at it here and there and it seemed pretty comprehensive the last time I looked at it.
Good thought, and I bought ESPlanner+ in the mid 00's but I've let mine lapse out of date. I was going to resubscribe to see Roth conversion and Soc Sec tax impact. Unfortunately last time I looked a few months ago, ESPlanner does not really handle Roth conversions. They offered up a workaround in 2009, that's cumbersome and not accurate by their own admissions, and they've promised to add that functionality for several years now, but it hasn't been done yet that I know if.

Roth IRA Conversion - The effects of taxation on this decision. | ESPlanner Inc.
 
Last edited:
I do a personal tax table each year with TurboTax. Nowadays the table bumps up IRA withdrawals in 10k increments and shows the marginal tax rates we will pay. So the columns are IRA withdrawal, SS taxed, Fed tax, State tax, Fed + State marginal rate, Roth spent (depends on IRA withdrawal), oveall tax rate.

Basically I key on the marginal rates. For us the picture is that it is about 13% at low IRA withdrawals, ramps up to 33%, and then flattens out at around 26%. When we eventually hit RMD's there is no way to avoid the 26% rate (or the 33% bump due to SS).

So any Roth conversions I made in previous years at lower then 26% were well done. Also the Roth money is what helps us now to get into the low 13% bracket until RMD's happen. ORP suggested high Roth withdrawals and I took the hint after doing my personal tables.

I don't know if this helps you Midpack, but maybe it will be something to add to your approach?
 
Back
Top Bottom