Roth Blueprint

I spent quite a bit of time developing a Roth conversion plan.  Conversion to effectively modify the bottom line has several moving parts to be optimized.  Each aspect in turn has it’s own set of optimizations.  


When to conv#ert

 How much to convert

Tax consequence (single and married)


Long term payback (break even)

Medicare cliffs

Difference between assets left in a TIRA (bonds v stocks)

Eventual use of Roth in a portfolio + SS (spending model)


Self Insurance

Tax treatment diversity

Sequential portfolio diversity (non-Roth v Roth) clean out risk first.

Planning money to live on while Roth converting

Tax loss harvesting

post tax 8606 IRA money

1 When to convert

The problem of when to convert is related to the W2 income.  When you are earning any conversion will be added to your ordinary income.  If you make for example an AGI of  $189K (top of 22%, married filing jointly, under age 65, standard deduction of $24K) it  will cost $28,179 in taxes.   If you Roth convert 100K the AGI will increase to $289K (well into 24%) and your taxes will be another $24,000 or $52,179.  Every additional dollar above $189K gets 24 cents whacked out of it.  In the end you spend $24K to get $100K into the Roth.  If OTOH you are living off of cash there is no income tax from a W2 source.  So you can convert $289K and your tax whack is still $52,179, BUT when it’s done you have $289K sitting in the Roth.  You can convert 2.89 times as much money for the same tax dollar.  Tax law certainly is “progressive”.  Let’s say you want to convert $1M.  If you do 4 conversions while W2-ing it you will have made a W-2 of $756K (189×4) over 4 years,  400K net converted and a $208,176 tax bill of which 96K was due to your Roth conversion.  Not so efficient.  Convert while living on cash and in the same 4 years you will have 1,156,000 net in the Roth and the same $208,176 tax bill.  If you converted $250K x 4 years to get your 1M your tax bill would be $386,513 well into 35%.  Clearly if you want to minimize taxes, live off cash while doing it.  

My analysis of when, is when you are retired and able to live off cash with no other income and you need to be able to live off cash for 4 years plus have the tax money available during conversion, which means pre-plan how to fund the conversion.  For the above 4 year example I would go to cash at age 66, 4 years before RMD and commence to converting, to end your conversion at 70.  Once you RMD the RMD money is ordinary income and just like a W2 can dramatically raise the cost of conversion but it’s even worse since you won’t have a W2 income to compensate so your portfolio is rapidly deflating.  Where you get “living on cash money” will be covered further on down the line.

2 How much to convert

Every dollar you efficiently convert is a dollar that will never be taxed again.   If you leave money in a TIRA especially money in equities it will continue to grow and that money once RMD’d will generate a tax bill.  If you have 500K in an equity rich TIRA with an 8% return you start at 18K distribution at age 70 and by age 80 you’re distributing 38K.  Your SS is 42K + 38K for 80K taxable and a tax bill of 6K well within the 12% bracket limit.

What if you have a 2M TIRA? 

Once again you start out slow at about 73K and by age 80 it’s $153K.  You left the 12% tax bracket long ago and you are approaching 24% bracket only 10 years into RMD!   If SS is say $50K/yr your taxable SS is $42.5K (SS is taxed at .85%) so your net is $195,500, into the 24% bracket.  Tax on $195K is  $29K.  And the tax picture only gets worse from there.  So cleaning out the TIRA is a tax saving maneuver.  #1 the tax is less so the strain on the portfolio is less and #2 you have to get tax money from some where to pay.  The W2 is no longer buffering the assault.  If the market is down, you get to sell your shares low to pay the tax man.  Bad for SORR.  If you had the same situation with a 500K account SS of $42.5K + $38K is an income of 80.5K and a tax bill of $6087  Much easier to come up with! 

Let’s say you die at 80 leaving the estate to your spouse.  She will loose you’re SS and keep her own, or she will claim survivor benefits.  I estimate survivor will pay her maybe 36K/yr or 30K taxable.  At 81 the RMD will remain the same or be recalculated depending on the inheritor’s life expectancy, so let’s say it stays on the same schedule.  The next payout would be $164K, add 30K SS for 194K   The (age 81) tax bill will jump to  $39K from 29K from jsut into the 24% bracket to the 32% bracket.  Your ol lady is getting hosed!  She makes less on SS and pays a lot more in taxes because you died.  Cleaning out the TIRA into a Roth does the following:  The age 81 RMD is 41K plus the same 30K from SS for 71K  for a tax bill of $8.5K just into 22%.  So your death in this scenario keeps her 2 tax brackets below the other scenario.  Uncle Sam already has his money so he leaves her relatively alone.  It only gets worse with a bigger TIRA.  By cleaning out the TIRA as much as possible and as efficiently as possible this scenario is avoided.  

The other problem with not converting enough is growth in the TIRA post RMD.  Let’s say you have 2M and pull out 500K into a Roth leaving 1.5M 

At age 80 your RMD is 115K and SS is 42.5K for $157K and a tax bill of about $21K intermediate between a 6K bill from a 500K TIRA RMD and a 29K bill from a 2M TIRA.  Cleaning out 500K helps but because the tax is progressive the power of conversion is eroded.  

How about the whole shooting match!  You convert all 2M at the tip top of the 24% bracket!  The tip top is 340K when you include standard deductions, so 2M would be almost 6 years of conversion.  The taxes on 340K is 64K for a 6 year conversion cost of 375K.  With no RMD your tax bill on 42K SS is about $1700

so to recap 

TIRA @ RMD2M1.5M1M0.5M0MSS taxable
age 80 RMD1531157738042
tax in K going forward29211361.7
yrs conversion 05556
total tax paid at conversion 034152273384
living on cash 4 to 6 yr
conversion to keep 
bracket below 24%
Spouse tax hit 362615101.730

The race is between dying and higher tax brackets.  The less you convert the higher the future tax bill, BUT the more you convert the higher the conversion tax bill so it’s pay me now or pay me later.  You owe the money and it will be taxed at an ever accelerating rate and once you die your spouse is hosed unless you planned for that.  Once RMD hits you’re locked in so what you gonna do?  Call Gasem!  One thing I did not cover are medicare cliffs. 

Cliff 1 at 250K (married) a 3.8% surcharge is placed on your income so converting at 340K is even more expensive making a 250K ceiling more attractive

Cliff 2 If you make more they charge you more for medicare 

This does not include a supplemental (typically about 150/mo) and it is per person  so when you convert 340K and you’re both 65  you pay about 11K per year

What you want is to stay in the 12% bracket for as long as possible and then in the 22% as long as possible.  The code is progressive so being middle class is your friend.  Soak the rich is alive and well!   I was noodling with the RMD calc and if you put only low yielding assets (bonds) in the TIRA the RMD goes down The age 80 RMD @ 3% return on a 2M TIRA is 95K v 153K in a 8%  TIRA, and the RMD accelerates more slowly.  For 0.5M the RMD at 80 is only 23K  Here is a table:

TIRA @ RMD2M1.5M1M0.5M0M
age 80 RMD 3%967047230
RMD + SS138112896542
top of 12% tax bracket 104
top of 22% tax bracket189

I estimate at a 1M IRA you will be well into the 90’s before you leave 22% and at 0.5M in bonds in the TIRA at RMD it will be a couple decades till you leave 12%  The outlook improves for your widow as well.

Tax savings if you don’t convert:

tax savings between 100% to 0% 3843502321110
on RMD conversion

If you don’t convert you save those taxes and they can keep compounding BUT the tax code is progressive so more and more will be taxed.  The rate of taxation goes up while the rate of return is pretty constant so the tax saving will erode quickly.

3 Account mix

I suggest 4 types Roth, TIRA, and post tax, and a Tax loss harvest account.  Your portfolio will develop over decades and there will be occasions to tax loss harvest.  Tax loss can be mixed with cap gain for a 0% tax bill.  Also as long as you stay in the 12% bracket the cap gain tax is zero and you only pay cap gain on the amount over the 12% limit.  If you have tax loss it can be applied to the over amount so a little dab of TLH can save you money if you bother to acquire it.   My FA has all of my tax lot info in his software so it makes harvesting effortless, and lucrative.

I used my post tax account mixed with TLH to pay for my Roth conversion.  I converted about 600K of post tax stock mixed with TLH for a zero dollar tax bill.  I saved about 100K in taxes.  The 600K pays for living expense and conversion taxes.  It offers another advantage in that it lowers my overall AA during conversion and in my early retirement which offers me some SORR protection.  This is recommended.    Most people are in accumulation mode but upon retirement you enter deflation mode and need to be ready to spend some money once the W2 disappears.   As the cash gets spent the AA will rise again but I’ll be farther into retirement so SORR becomes less important.  SS is deferred till age 70 so I have no income.  My chosen conversion is 250K per year x 4yrs which avoids most of the cliffs, and will convert net 1M and  a gross 1.4M @ 4% when compounded at conversion end.  My TIRA will be about 600K all in bonds and at 3%  will RMD 22k the first year and 28K at 80.  My net tax saving will be about 130K which will stay in the post tax account compounding.  I am moving the riskiest assets first, then down the line till only bonds remain.  My disbursement model at RMD is SS + RMD plus I will sell some post tax stock as needed.  Since I only get taxed on RMD and 85% of SS, my taxes will be low for a long time.  Between my TLH and <12% income I will have no cap gains as I disburse post tax money.  I won’t touch the Roth and let that grow as insurance in case of medical disaster or as a wealth transfer vehicle.  Everybody gonna die from something and some of those causes are very expensive so having money stashed beyond living expense money is a comfort, and not only you but your wife will incur expense.  By owning 4 account types I have excellent control over my tax picture going forward.   A  lot of my “conversion money” is from interest accrued in my post tax account over the decades

4 Account imbalances 

You read a lot of boiler plate “fill up the pretax accounts” it may not be the best advice.  Those accounts are tax deferred not tax free and their size determines the tax consequences which are progressive and tilted toward soak the rich.  I’m sure there is controversy on this but I think equal tax deferred and post tax up to maybe 4M and then over in the post tax above that.  A 5M tax deferred is not a trivial liability in the grand scheme IMHO. YMMV but that’s my take. 

5 8606 Money

My IRA and my wife’s IRA were funded post tax.  I kept all the 8606 sheets.  It turns out once taxed that money changes the tax basis for the life of the TIRA.  Every year a proportional smaller amount will be taxed than what the RMD requires so if you RMD 25K and you have a 80% basis adjustment you will pay tax on only 20K.  My adjusted basis allowed me and my wife to transfer 265K into the Roth with only 245K taxable.  I’ll top that up once I have a better picture of for the taxable portion of my ordinary income from my post tax account.    All of these little tid bits add up.  265K compounds faster than 250K. 

5 Modeling 

This was all modeled in Excel in modules so  could understand the implications of each on the other and get some idea of multivariate optimization.  My portfolio and conversion strategy was further modeled in some commercial software designed to advise optimized Roth conversion strategies.  My actual modules matched the commercial strategy very well in terms of prediction.  My final optimization was different than theirs because they suggested 100% Roth conversion.  Over a very long time I agree with their strategy but it takes a while for the conversion to turn cash flow positive, for the tax progressive savings to over come the initial cost of conversion.  The initial cost can be initially viewed as a negative SORR.  A slightly less aggressive conversion has less SORR character.  The cash flow positive point is about the same.  My wife is younger than me by 7 years and is genetically predisposed to live a long life, with no real history of cancer, typically age to 90+.  My side is dead by 80 from CAD and metabolic syndrome, so some attention to that is built in.  My modeling is more extensive than 10 years age 80, but models 5 year aliquots of time with and with out married filing jointly and takes into account progressive tax codes.  I also optimized SS but won’t go into that here. 

6 Budgeting and cash flow management

I believe in budgeting as a means to judge progress but I’m not a slave to it.  I can afford my pre-retirement income, including a risk premium for health care etc aka the bennies I lost with the W2  I’ll call that Max budget.  I retired on 80% of my “max budget” as this was described as comfortable to most retirees.   After the dust settled I decided to experiment with belt tightening since you read about that as a solution to bad times and I wanted to understand what that actually felt like, not as just some bromide.  I could go down 80% of the 80% or 64% of what I can afford.  So I oscillate between 64% and 80%.  Some months are cheap months, some expensive.  If we want something like a trip to EU (been twice since pulling the trigger) I save up the differential between cheap and Max months and when I have the differential we fly.  My wife is good with that technique.  All in all I came in 11% under my 80% max budget last year.  This year looks about the same, a little more expensive but likely inflation related.  Inflation is something to consider as it will eat up the slop in the calculation, but I ain’t skaired.  

This is pretty much what I did and am doing.  I’ve invested for several decades and investment vehicles exist now that did not then so you do what you do in the environment in which you find yourself, and the environment into which you are retiring.  If you got a ten year nest egg and a 50 year horizon, that life has a very different risk profile from mine.  That life is quite leveraged, my life is not.  I can survive quite comfortably at 0% interest for decades, but it gives some insight into my thought process. It is not advice just my experience.  It’s complicated and based on probabilities and probabilities are not certain.

Calculators I used 

Tax plan calculator

Schwab RMD Calc

FV Calc


I’ve completed my first conversion this year and will convert my second early next year probably Jan and then be done till 2020.  I owe the IRS 42K by my estimate and 44K by the HR block Free filer which doesn’t take into account some writeoffs I will claim.  The official HR block software won’t be live till Dec so I will get a more precise estimate then, and will owe essentially the same tax again next year, so it’s working predictably and according to plan.  I came across a tax penalty rule that says for high earners (>150K) you need to prepay 110% of last years tax to avoid the penalty so I take that to mean if my tax is 42K this year I must pay 46.2K next year to avoid the penalty.  Of course I would then apply the extra 4.2K to next years taxes, talk about soak the rich!  They’re screwing you 2 years out on money you haven’t made yet!  All the more reason to get them out of your hair.  I also checked the credit card option to see if I could claim cash back points and with the “convenience fee” it worked out to be a wash between CC and a check.

I’m now officially sick of researching Roth conversion, but quite satisfied in the result.

6 Replies to “Roth Blueprint”

  1. Thank you for writing this Gasem. This is exactly what I need when I begin to do ROTH conversions myself.

    I figure because I plan on retiring several decades before RIMD kicks in, I should have ample opportunity to do the conversions minimizing the tax implications.

    Will definitely bookmark this for future reference.

    1. Properly implemented the plan can take decades to mature. If you retire at say 55 you would have a plan between say 55 and 65 for living expense, and you can let the TIRA and post tax accumulate and do TLH. At say 65 cash out living expense, start conversion and be done at 70. If you’re 45 it’s time to set the plan in motion so at 65 much of it’s expense can be covered by accumulated interest.

  2. Omg I do not think the internet is ready for this. But you highlite what I cringe about over other sites.

    Financial plans can get complicated very quickly. AND it is the need for revisiting and double checking the numbers. REGULARLY. You illustrate that nicely here. That is exactly why you laugh at 4 x 25 all the time. If it was really that easy, why bother even thinking about it.

    The truth is that risks await many of us all. And sometimes the numbers don’t strike zero until one is 85 plus without other options. That is the sad part. It pays to be a tad paranoid.

    1. If you were to stand on the top of a normal Gaussian curve at the mean and look down you would see the most likely outcome aka the mean but sweep your vision right and left and you see countless other possibilities less likely but still possible. Out toward the tail many of those possibilities are rather undesirable. In the tail you run out of hamburgers! The goal #1 is to stay out of the tail. Everything after that is groovy and only requires an attitude adjustment, not a dietary one. Financial plans aren’t all that hard. P-Chem was hard. Quantum physics was hard. Some of my engineering coursework was hard. The trick to financial planning is detail and follow the bouncing ball all the way to the end. Once you get to the end turn around and see what u shoulda did, then go do that. 4 x25 is far better than no plan, and a mechanical investing model like you have with controlled risk is virtually assured of some hi level of success. Tinkering with the model with made up rationalizations or failure to actually get the money into the market (follow through with the plan) is what brings down returns. Indecision overrides discipline instead of vis versa. What I laugh at is the arrogance of the certainty where that success will take you. It’s fun to dream but it takes more than wishes to do the dishes.

  3. I really appreciate your entire series of posts on Roth conversions. For most blog readers interested in the subject it’s still theoretical; you’re showing us the ingredients and process to actually bake the cake. I’ve been consulting the Schwab RMD for many years now and could easily see that the tax burden for an equity-rich tIRA escalates quickly as you age, which you demonstrated in one of your examples. While every conversion plan needs to be customized to individual circumstances I’ve picked up several tips that will improve my execution and hopefully minimize the taxes that I will ultimately have to pay.

    1. That’s exactly how I look it. The minute you enter into the tax deferred bargain, the taxes already belong to Uncle Sam, and once he comes for them at age 70 your options run out. For a small TIRA the old saw “at a lower rate when you retire” works but for a high income saver the tax rate is maintained if not progressed. Once you have a handle on all the dials it’s not too hard to tune it for max efficiency

Leave a Reply

Your email address will not be published.